Russia Going To GNU/Linux Late Rather Than Never

“Google, Apple Inc. and Microsoft Corp., collectively worth more than Russia’s gross domestic product, have all entered German Klimenko’s crosshairs since he was named Putin’s first Internet adviser six weeks ago.
 
In a 90-minute interview peppered with expletives, Klimenko said forcing Google and Apple to pay more taxes and banning Microsoft Windows from government computers are necessary measures best explained in terms of barnyard economics and marital infidelity”
 
See Putin’s New Internet Czar Wants Apple and Google to Pay More Taxes
Back in 2010, Putin put into (slow)motion a move to GNU/Linux. There were several projects but nothing concrete and system-wide. Finally, in 2016, thanks to the price of oil, sanctions and global politics, the time is ripe. The new “Internet Czar” has this on his todo list with a high priority. The Russian government has had it with US corporations having control of their IT. FLOSS is the way to go. If they have the code, they can control their own IT. That’s the right way to do it, even if “it” is sometimes the wrong thing like invading Ukraine or Syria.

About Robert Pogson

I am a retired teacher in Canada. I taught in the subject areas where I have worked for almost forty years: maths, physics, chemistry and computers. I love hunting, fishing, picking berries and mushrooms, too.
This entry was posted in technology and tagged , , , , , , , , . Bookmark the permalink.

192 Responses to Russia Going To GNU/Linux Late Rather Than Never

  1. DrLoser says:

    A competent GNU/Linux system administrator has a dozen options for keeping systems up to date.

    And your competence as a GNU/Linux administrator is amply demonstrated by your inability to come to terms with modern configuration options such as systemd, Robert.

    Never mind. Rather than a “not very competent” Pogson, let us stipulate a “well paid, professionally competent, GNU/Linux administrator.”

    Why should such a person need more than, at most, two options for keeping systems up to date?

    In the real world, Robert, one does not simply play with computers. One has to use them for purposes that have financial consequences.

    Pissing around with a dozen or so playthings doesn’t really work.

  2. DrLoser says:

    Does anybody want to explain to Fifi what happens when you open Pandora’s Box, incidentally?

    I assume you are all (other than the lamentable Dog-Brain) educated well enough to understand the allusion.

  3. luvr wrote, “a competent administrative user who manages multiple Ubuntu desktop computers will make sure that he won’t have to periodically log in on each of them, just to install updates. If he cannot figure out how to do it, then he’s not competent.”

    That’s one definition of competent. I think in the world of ordinary people using PCs it’s not relevant. Ordinary people barely think of updating software for any purpose let alone security. A competent GNU/Linux system administrator has a dozen options for keeping systems up to date. I like LTSP because I could update a few servers and do the whole system. Then, you could have local scripts running periodically to check and update from repositories or you could have your own local repository or cache. You could also generate whole disc-images and distribute them via SSH or whatever. Or, you could choose to do nothing. They are all valid options for many cases. The typical consumer wanting to browse FaceBook or check the weather really doesn’t know he has those options, doesn’t care, and probably won’t do any of them. It’s like the original consumer VCRs. To set the clock, many consumers needed a teenage relative to get the job done. I remember. I was that teenager for my family. I’d bet there are a few percent maximum of consumers who will even configure updates if they were not done by default.

    Added to this is the certain knowledge that if automatic updates are used eventually one will do some harm even if it’s to hang the update process. It happens. I worked at one place where the updates involved were malware, a worm that was making the rounds. There the top IT guys decided to pull the plug on the Internet making automatic updates impossible or difficult. So, even a competent administrator can choose to deny updates. It’s an imperfect world. Live with it.

  4. luvr says:

    kurkosdr wrote, over and over again, “Meanwhile, Ubuntu still requires a sudo password to apply security updates (what if the privileged user logs in once every 6 months?), doesn’t apply security updates automatically”

    You do realise it’s a piece of cake to enable unattended updates, don’t you? In fact, Ubuntu Server enables unattended security updates by default (but offers you the opportunity to automatically install all updates, or even to turn the feature off, upon installation). Ubuntu Desktop, too, has all the infrastructure in place to enable unattended updates, but only activates automatic refresh of the package lists. Thus, a competent administrative user who manages multiple Ubuntu desktop computers will make sure that he won’t have to periodically log in on each of them, just to install updates. If he cannot figure out how to do it, then he’s not competent.

  5. kurkosdr says:

    Just like there is no Windows/Malware/Anger connection?

    Since Windows has auto-update turned on by default since the turn of the millennium, got UAC in 2007, has closed the USB auto run loophole and the unsigned drivers loophole, can you tell me how exactly Windows is more vulnerable to malware? And please don’t mention tales from the era of Windows XP, that thing is computer-museum stuff in 2016.

    I want to hear detailed penetration testing scenarios, not “unix-heritage diddly daddly doo” stuff.

    Meanwhile, Ubuntu still requires a sudo password to apply security updates (what if the privileged user logs in once every 6 months?), doesn’t apply security updates automatically, and I bet the Linux Mint guys didn’t run Microsoft Windows on their servers.

    I am still waiting for the penetration testing scenarios that make Windows more vulnerable to malware than Desktop Linux btw…

  6. kurkosdr says:

    May we all, at least, come to the blatantly obvious conclusion that there is no such thing as “Gnu/Linux/Android?”

    Also, there isn’t such a thing as GNU/Linux, except for specialized use-cases where no-GUI Linux Distros are used.

    X/GNU/Linux is a more appropriate term. But maybe we should separate between KDE/X/GNU/Linux and GNOME/X/GNU/Linux. Or just call the thing Desktop Linux. But… does Jolla count as Desktop Linux?

    See? This is what happens when your “OS family” is a mess of swappable bits.

    At least Android/Linux is Android/Linux.

  7. dougman says:

    Just like there is no Windows/Malware/Anger connection?

  8. DrLoser says:

    May we all, at least, come to the blatantly obvious conclusion that there is no such thing as “Gnu/Linux/Android?”

    It seems a trivial observation to make. We can all carry on from there.

  9. DrLoser says:

    Fifi Has A Teeny Problem With Reality
    Go a head search DeafSpy. The first person to write that block of text here is DrLoser.

    A lot of people here have conveniently short memories. I am not cheating. I am quoting. For some bizarre reason, the idiot oiaohm actually insisted on this interpretation.

    Now let us turn to chapter 2 of his cite. What do we find in the first paragraph?

    A first advantage is the optimization of code due to the removal of all comments from the header files. This reduces the size of all those headers and helps with a speedy process in a storage-constricted space.

    What an unbelievable idiot you are, Fifi.

  10. DrLoser says:

    DrLoser really not point complaining now you opened this Pandora box when you decide to incorrect attack paper.

    This being the same paper (your cite) wherein really not point complaining now you opened this Pandora box when you decide to incorrect not admit first para chapter two verbatim him say yes me talk-talk good ting good real word many ting quote big man not evil real be de accurate ting Fifi little girlie deny real ting deny deny?

    Yup. You’re a pathetic lying little loser, Princess.

  11. DrLoser says:

    Since Android is a Linux Distribution why would it treatment be different.

    Words mean nothing at all to you, do they, Fifi?

  12. DrLoser says:

    DrLoser really not point complaining now you opened this Pandora box when you decide to incorrect attack paper.

    Now then, Princess. Don’t go getting all tizzy and stuff. One word at a time. Baby steps, baby steps. Subject, verb, object if verb not intransitive, possible punctuation, relevant sub-clause.

    Baby steps, Princess. Baby steps.

    Once you can convey your intent coherently in some sort of approximation of the English language, we may resume our interesting conversation regarding the manifest deficiencies of gcc.

    Which, as I say, is quite interesting. Because you brought them up, complete with a number of cites. I did not. You want to know why I didn’t?

    Because, Princess, they are utterly irrelevant to the question of why Google chose Bionic as the C runtime library for Android, rather than glibc.

    Chunter on, tiny little dag.

  13. DrLoser says:

    That came out of the point you DrLoser pulled out of a cite because you could not see why stripping headers was an optimization.

    No it did not, oiaohm. Precisely the opposite. Let me now explain the simple trick that makes anything at all between a lexed LEFT-COMMENT and a lexed RIGHT-COMMENT completely invisible to the AST of a proper parser.

    As a parser, you know what you do? You just throw all those intermediate tokens away.

    It’s highly entertaining that, in certain circumstances involving the pre-processor, the GNU compiler apparently fails to do that.

    But, yeah, mate, that’s what a sensible compiler would do. Given the horrid undersigned confusion between the gcc compiler and the various gcc optimisers … may I interest you in the Dragon Book?

    It’s not up to date in theory, but then neither are you. As of 1960, in your case.

    Ever wonder why I vary between calling you oiaohm and Fifi and Princess?

    Think about that. Think very hard, oiaohm.

    Stripping headers is never, under any circumstances, an optimization.

    And, bonus free gift:

    There is no such thing as Gnu/Linux/Android.

  14. oiaohm says:

    The stuff about the gcc optimizer was spectacularly off-topic, particularly since I have pointed out that I was using Bison and Flex as a pedagogical example. However, I’ll admit that it’s all very interesting. I’m sure the subject will crop up again in future.
    That came out of the point you DrLoser pulled out of a cite because you could not see why stripping headers was an optimization. In the android environment with gcc with this issue it is.

    Not knowing gcc flaws means you did not understand what was a optimization that provided a gain so you attacked an author of a paper you have not said sorry to yet DrLoser.

    And you still haven’t figured out how to get the front end (optimizer and all) to ignore comments in headers, have you? Little problem. You are going to guess preprocessor and be badly wrong after the preprocessor what file each part has come from was noted. In fact the front end optimizer goes and opens the source file back up in some cases. So straight past the preprocessor. The only way is strip it out the header file itself. Precomplier of headers does not avoid it either as a check of a precompliered header can still trigger Genericizer to go check the source file.

    I presume you have never once attended a lecture in elementary compiler theory.
    I have and those lecture will lead you into complete incorrect guess what the gcc Genericizer is going todo. Tokens at that point look strange it goes and refers to source again. So it not only optimizing it validating. Normal lecture says once you pass the preprocessor you will not be opening the source files up again. This is not gcc. Gcc is you might not open up the source files again but with the correct triggers the source files will be opened again and reprocessed for different reasons.

    This is your problem DrLoser in this case you are apply elementary complier theory not how gcc is implemented or behaves. Name where in elementary complier theory is there a validation step hidden in the first optimizer to check what preprocessor generated valid code. Code that was added because there was a odd bug where cpp was doing stupid things and it remained there ever since creating a very odd behavior.

    That is central to my question. For some reason you insist on going off-topic, which is fine, just so long as you admit that you are not answering my question.
    I have answered you point a few different ways. Maybe you don’t like it. glibc loader will work on android its just not a straight forwards nice process. So glibc dynamic will work on android.

    Libhybris is a subloader to load bonic parts with glibc core. Is being exploited by a few projects to run android applications on normal Linux core. So replacing bonic by rebuilding the complete android code base is truly possible and this is the same process you have todo normally when you replace a libc with another type. Like glibc to musl on a Linux distribution.

    DrLoser really not point complaining now you opened this Pandora box when you decide to incorrect attack paper. Then you apply wrong theory. Changing the core libc type in a normal Linux Distribution is rebuild the complete thing. Since Android is a Linux Distribution why would it treatment be different. If you don’t rebuild the complete thing you use a method that places the other libc loader on top. Normal Linux Distribution process.

  15. DrLoser says:

    You know what? (And I assume this will sound a little freaky.) I’m going to have to tip my hat to oiaohm on this one. He may have a point:

    The GNU libc also currently contains macros to optimize calls to some string functions with constant arguments and those that can be implemented by processor specific instructions. These transformations are better performed in GCC, both to reduce the overhead of macro expansion and to take advantage of the functions attributes, for example to avoid a second call to a pure function altogether. The use of these macros tend to cause huge blowup in the size of preprocessed source if nested; for example, each nested call to strcpy expands the source 20-fold, with four nested calls having an expansion ten megabytes in size. GCC then consumes a huge amount of memory compiling such expressions. The remaining optimizations need to be implemented in GCC and then disabled in glibc, with benefits to other systems as well, and the potential to use information GCC has about alignment.

    My emphasis, just to point out the bit that’s important to the current discussion.

    As a matter of fact, this doesn’t invalidate my claim that stripping header information should have absolutely no impact on optimization whatsoever. It still shouldn’t. For any other compiler I can think of, other than gcc, the question doesn’t even arise, because no other compiler I can think of performs this insane and unnecessary trick.

    But even with gcc, my as yet unstated simple mechanism for the parser (optimiser, whatever bit of junk hangs around in various random layers of this most excellent example of freetard design) still applies.

    It’s just that it applies to the preprocessor rather than to the parser.

    I’m indebted to oiaohm. Without his persistent off-topic googling, it would never have occurred to me that gcc could be quite this wretched, straight out of the box.

  16. DrLoser says:

    Static linked glibc applications do run on Android.

    That was never my question. I imagine that static linked just-about-anything will run on Android.

    Dynamic linked glibc is tricker since glibc is not compatible with the normal Android bionic loader.

    That is central to my question. For some reason you insist on going off-topic, which is fine, just so long as you admit that you are not answering my question.

    The stuff about the gcc optimizer was spectacularly off-topic, particularly since I have pointed out that I was using Bison and Flex as a pedagogical example. However, I’ll admit that it’s all very interesting. I’m sure the subject will crop up again in future.

    And you still haven’t figured out how to get the front end (optimizer and all) to ignore comments in headers, have you?

    I presume you have never once attended a lecture in elementary compiler theory.

  17. oiaohm says:

    DrLoser gcc front end was not really designed. The backend was not either until gimple was implemented. Someone basically still need to go back and

    The front end optimizer of gcc is called of all things “genericizer” Its allows to look pass the parser and snoop threw the source to decide what optimizations to turn on or off in latter passes. This is about connecting what was tokened to generic gimple structure for latter processing as well. So there is a particular point where gcc can kick it toe and screw up badly. The genericizer is where language dependent optimizations are to be performed as well. Please note parser is finished before gcc spins up the genericizer.

    There are three distinct locations for optimizers in gcc. 1 in the front end. 1 in the mid end and 1 in the backend. It is the front end one that causes some really creative strangeness.

    Yes gcc naming also is trick central.
    Optimizer –this is the back end optimizer name performs platform targeted optimizations–
    Tree SSA Optimizer –this the name of the mid end optimizer this performs language neutral and mostly arch neutral optimizations–
    Genericizer — is the name of the front end optimiser and this performs language Dependant optimizations and turn on and off particular optimizations at particular points in code in latter processing and how it works is sensitive to ram at times due to the fact at that point you can still snoop through the source.–.

    2b) The user can compile glibc on an Android device and, at the very least, build applications on that Android device that do use glibc but do not use Bionic.
    Static linked glibc applications do run on Android.

    Dynamic linked glibc is tricker since glibc is not compatible with the normal Android bionic loader. Static linked binary to glibc does not reach out and use Bionic. glibc does not mandate being built as a dynamic binary. So if user is aiming to build glibc as a static library everything is simple other than license. With license using musl another common GNU/Linux libc that is MIT license is legally safer in this case. Of course that legally safer option exists still does not stop developers from being lazy and incorrectly using glibc.

    2a) The user can simply “drop in” a compiled version of glibc in place of Bionic.
    This point is kinda wrong. For static linking yes glibc can be dropped in place of Bionic to be used with android for drivers support features.

    Dynamic linking you run into the same problem as you run into with musl and glibc and any other dynamic libc on Linux. Dynamic loader is part of the libc not the Linux kernel. Using a Dynamic loader under Linux involves linking particular code into your executable. The stub. How the stub talks to the loader is allowed to be unique per libc solution. Loader connection was never was standardized. From kernel to loader is fully defined from loader to executable is not. Majority of Linux binary incompatibility starts and ends with the loader what is in fact not part of the kernel.

    Result you cannot just drop in replace musl with glibc or glibc with musl in normal Linux systems or any other Dynamic loaders from other libc on linux that are not forks. To this will always require a custom subloader and most cases the result is stuff doing that when you can just run both side by side with each other..

    Linux kernel design does not forbid you from using more than 1 loader and Android kernel is not patched forbidding using alternative loaders.

    There are examples of completely rebuilding android where bionic has been removed and extra subloader being used and that subloader is a modification to the bionic loader to allow the embedded in executable code of bionic crtbegin_dynamic.o and crtend_android.o to work.

    Linux kernel design does not forbid running loader directly either. Since a loader is only a normal static binary on Linux.

    And to work around issues.
    http://code.metager.de/source/xref/chromium/tools/android/run_pie/run_pie.c
    By the way its nothing exactly strange to have to use a subloader on android to cover up bionic loaders failures. So totally disabling applications from using their own loader under Android is impossible because if they do it will break too many applications.

    If android application ships with a glibc loader it can use glibc dynamic. If it does not ship with glibc loader it can use glibc selectively with the bonic loader. Issue with the bonic loader it takes host installed .so files over those in applications so where glibc requires the same library has host todo something special all hell breaks loss. Yes this could be a nasty problem in future for android applications shipping with libraries if google adds a library and there is a name match because then the library in the application will be over ridden unless the application protect itself with either its own loader or subloader.

    glibc and musl have not spent the time to make a subloader not when their standard loaders 90+ percent work on android. The LSB wrapper code still works quite well to make application load glibc or musl loader from a path related to where it executed.

    Please note you cannot write a pure native code application user application under android either because you have to use the java like language to create the output window. Changes are seeing all that head in the direction of LGPL.

    DrLoser I guess you have never used a multi libc Linux to understand how the beasts work. The split between Libc is the loader more often than not. This is where your arguement goes south really. How side by side libc are done on GNU/Linux is a little odd.

  18. DrLoser says:

    Really we don’t need to. There is stupidity of gcc. Front end section of gcc is not just bison and flex it also contains optimizer. Because someone thought this was a great Frontend code Optimizer is free to go and search though the raw source files again searching for extra instructions that the parser did not include.

    This, for once, is very interesting, and quite possibly accurate. Note that it implies a catastrophic deficiency in the high-level design of gcc, if true. I was merely using Bison and Flex as friendly examples — in general, mature compilers are quite strict about the boundaries between lexing, parsing, static optimization and per-target optimization.

    Whilst doinking around a few years ago trying to chain a Gentoo build together, I spent a fair bit of time examining the gcc code. (Hoorah for Freedom!) It was my impression at the time that the gcc lexer and parser are relatively clean and unaffected by optimization predicates — either implicit (say, inlining of functions) or explicit (generally, referring to command line switches).

    Having read oiaohm’s assertions above, it occurs to me that I was being far too generous to the designers of gcc. I was, of course, looking at the back-end options such as “arch” and so on, which imply that you set up the various targeting options such that the “pseudo-assembler” emitted by the Abstract Syntax Tree produced by Yacc and Bison can be applied via what is, in any normal compiler, the third stage — to whit, the Optimizer.

    Fifi’s comment lends me pause. Thinking about it from a higher level, it’s quite possible that it really is the steaming pile of shit that he implies.

    I have no evidence either way (I could never quite get the “arch” bits to fit my use case), but I must admit that it is a profitable line of exploration.

    Oh, and btw. In no possible way can it invalidate my scenario, as presented. Do I seriously need to explain to anybody how the Parser can “escape” the danger of being fed endless tokenised rubbish by the Lexer?

    Anybody but Fifi, of course. But then, Fifi is a lost cause.

  19. DrLoser says:

    Me: … can build an executable using gcc and glibc on an actual Android device.
    oiaohm: Opps I missed that bit. You don’t need glibc on an android device to build gcc. Once you have gcc built you can built glibc if you wish.

    Not relevant.

    Try again.

    I made my requirements quite plain in the first place. I have continued to make them plain. I shall repeat them until you give me an answer that satisfies them.

    1) gcc shall be a fully-functional compilation tool on an Android device. (I believe it is. This is not at issue.)
    2) glibc shall be available as a replacement library for Bionic on an Android device. Two possibilities occur (others may be possible):
    2a) The user can simply “drop in” a compiled version of glibc in place of Bionic.
    2b) The user can compile glibc on an Android device and, at the very least, build applications on that Android device that do use glibc but do not use Bionic.

    I don’t see how either case is possible. (Paving over the device with an alternative distro is possible, but that is not the question.)

    Nobody has yet cared even to attempt to explain how either case is possible.

  20. DrLoser says:

    Really I already told you that GNU is GNU is Not Unix. So talking about Posix to make a point that GNU cannot be Android is pointless.

    I never cease to be amazed by the shallow breadth of your ignorance, Princess. Apparently you cannot even master simple predicate logic.

    I did not say that GNU cannot be Android. I said that Android is not Gnu/Linux. You are reversing the proposition, which can only be done in first-order logic if you are proposing an equivalence relationship. I am not proposing an equivalence relationship.

    The proposition that “GNU is defined as ‘GNU is Not Unix'” is indeed a partial equivalence relationship. (A trollop in northern NSW is also not Unix. It depends upon how you bind your variables.) However, it is totally irrelevant to the discussion.

    I was not talking about Posix to make a point that GNU cannot be Android. Or, to make your proposition a little more coherent, I shall rephrase it as: I was not talking about Posix to make a point that Android cannot be GNU.

    (In passing, “Android cannot be GNU” is not a categorical imperative. Android is not GNU right now, certainly. I offer no assurances for the future. I realise this is a little besides the point — if you can’t even understand first-order logic, you are unlikely to understand Kantian metaphysics — but I thought I’d point it out, simply for completeness.)

    My proposition refers to my conclusion that Android isn’t even Linux, let alone Gnu. My evidence for this proposition shall be presented shortly, and certain questions about Posix shall be offered as relevant parts of that evidence.

    I thought it was worth testing your intellectual mettle by allowing to you explain Posix in the context of Android, Princess, but I missed out a crucial prior requirement.

    You don’t have any intellectual mettle at all, do you? In fact, I doubt you even have the intellect of a hamster.

  21. DrLoser says:

    Not really thinking the guy at that time is doing a extra qualification. Standing professors are not forbid in undertaking extra study as long as it external at a different UNI to where they are. So its knowing who is where.

    Apparently you are narrowing down your unsubstantiated idiotic claim to refer to Mathieu Devos. (And it still does not matter. A Professor spouting nonsense is just a guy with fur on his academic gown spouting nonsense.)

    Let’s see a cite for his PhD then. The best evidence I can find is from his Google+ profile:

    Education: Ghent University (industriëel ingenieur ICT, 2010 – present)

    It might be that Herr Davos is hiding his professorial light under a bushel, Fifi. It’s what professors do all the time. Hardly a single one of them will ever remind you that they are, in fact, a professor.

    Alternatively, it might just be that you are talking gibberish in a foolish attempt to delay the inevitable moment when you will have to admit that you cannot natively compile an executable on an Android device using glibc.

    There is no such thing as Gnu/Linux/Android. Remember? That was my one and only point to begin with.

    Kindly admit that obvious fact, and we can all get on with our lives.

  22. oiaohm says:

    Wizard Emeritus
    Personally I await your example of somebody actually using glibc in a non developer fully functional product that they offer via the Google store. And remember it has to work without rooting or techno game playing.
    http://kevinboone.net/kbox2_how_it_works.html
    Just look around on the google play store for stuff using kbox2 or equal as these solutions are using glibc based applications yes they will have normally at least 1 bionic libc application with them.
    https://play.google.com/store/apps/details?id=com.myboyfriendisageek.airterm&hl=en
    Like above. Works on non rooted phones and yes it is using glibc applications, Its also using libfakechroot.so to over map to applications see paths place as per FHS. Hey lets not have to modify the source code for normal Linux or bsd to use it on android heck in some cases don’t even bother altering the binary just play with the loader. User-space deception works well enough to get normal console dynamic binaries for linux using glibc working on Android devices as long as you have an native android app providing a console to display output.

    Get the catch. glibc applications on android will be background applications at this stage because writing to screen requires using the bionic libc. APK package does not say you have to use 1 executable. So having 1 bionc libc application for OS handling and everything else glibc binaries works.

    Does glibc security risk effect Android hell yes. Will it always be in face as much as applications providing terminal using kbox no. In fact just reading the words kbox you would not have considered for one min that airterm is in fact a glibc using program that you have just installed on your phone/tablet.

    From my point of view the security risk from glibc on Android is greater than the security risk form glibc on other normal Linux Distributions due to the fact users know they are using glibc or not on normal Linux Distributions and Android users may not have a clue they are so are likely to ignore any security advice about glibc issues. Yes using glibc on Android is all about saving the all important developer time by not reinventing wheels.

    The reality I said glibc apps are in google play for android this is a fact. I never said you did not need to know what you were looking at to even spot them. The reality is without knowing what you are looking most applications using glibc you ask users they would say they are not using any glibc applications. Ask them about the one of the solution names that allows glibc on android and the response changes yes I am using an application that uses that. Even if they say no there are developers out their with there own internal solutions.

    It also worse than you can dream. Some applications download parts after they are installed. So glibc free in the APK from google play after installed and configured glibc containing. Welcome to Android stealth dependencies. Not only stealth in google play store also stealth outside it as well.

    http://stackoverflow.com/questions/10798357/want-to-compile-native-android-binary-i-can-run-in-terminal-on-the-phone
    Also it surprising what works with gcc when you try. Like this is OS X gcc being told to use Android Libs producing of all things a working Android binary. If you shove gcc around the right ways you can use gcc on Linux to product a OS X binary and vice verse as well. Gcc is no where near platform locked in code production. If binary type is ELF and Arch type is right its all a question of how many overrides to make gcc spit out the functional binary for any OS using ELF. One gcc with matching binutils building Linux, BSD, OS X, Android and Solaris is more than possible.

  23. Wizard Emeritus says:

    “They probably already exist because of the NDK:“”

    Anyone can hack anything, Robert Pogson. What counts is what is used by those developers who wish to target the Android platform. Hacking together a toolchain to allow for the compilation of glibc on top of Android is at best technical esoterica ,

    nothing more.

  24. Wizard Emeritus wrote, “I await your example of somebody actually using glibc in a non developer fully functional product that they offer via the Google store.”

    They probably already exist because of the NDK:“The Android NDK is a toolset that lets you implement parts of your app using native-code languages such as C and C++. For certain types of apps, this can help you reuse existing code libraries written in those languages.”

    With so much computing power now on smartphones, this could be very useful for compute-intensive applications.

  25. Wizard Emeritus says:

    “Anyone who knows gcc knows this. So basically DrLoser a question that shows you had no clue on this topic at all.”

    Nothing is impossible for the person who does not have to do it himself.

    Personally I await your example of somebody actually using glibc in a non developer fully functional product that they offer via the Google store. And remember it has to work without rooting or techno game playing.

  26. oiaohm says:

    can build an executable using gcc and glibc on an actual Android device.
    Opps I missed that bit. You don’t need glibc on an android device to build gcc. Once you have gcc built you can built glibc if you wish.

    Yes the native linking against instructions to build static work. Its not like gcc or binutils have to be built using glibc to make applications use it. glibc is just a set of libraries with a dynamic loader. You can build musl libc on andorid using a gcc using bonic libc just as straight forwards as building musl libc on a gcc using glibc.

    Basically the libc of the gcc complier makes no difference if you can build the libc or build applications using it. The complier and the libc are not strongly linked.

    Anyone who knows gcc knows this. So basically DrLoser a question that shows you had no clue on this topic at all.

  27. dougman says:

    Re: Is that just like your faith in bitcoin?

    That’s like saying faith in math is a joke.

    Bitcoin is a commodity, so I find your lack of faith disturbing. You are just mad that you did not buy any when it was low a few years back.

    http://www.davispolk.com/sites/default/files/2015_09_28_CFTC_Brings_First_Bitcoin_Enforcement_Action.pdf

  28. Wizard Emeritus says:

    “Lol, the REACTOS thing was a joke, obviously you missed it.”

    Is that just like your faith in bitcoin?

    “I contributed enough to myself so that I could semi-retire, not that it matters to you.”

    What only “semi” … Too bad.

  29. dougman says:

    Lol, the REACTOS thing was a joke, obviously you missed it.

    Re: And what have you contributed little man?

    Whoa!..who told you I was a midget? *kicks* you in the shins. Joking aside, I contributed enough to myself so that I could semi-retire, not that it matters to you.

  30. oiaohm says:

    If you’re referring to a guy who was in the first stages of a Masters in 2013-2014 (the dates of your cite. I mention this because you do not read your own cites. You are presumably unaware of the date of publication), then he’s climbed the greasy pole of academia remarkably quickly, hasn’t he?
    Not really thinking the guy at that time is doing a extra qualification. Standing professors are not forbid in undertaking extra study as long as it external at a different UNI to where they are. So its knowing who is where.

    =-Correct comments don’t come part of final code but they can cause final binary not to be optimized properly due to consuming ram.–

    Ah, an interesting limitation in that part of the back-end of a compiler that deals with optimization.

    Curiously enough, the lecturers at Cambridge never mentioned this one. Nor have I seen it mentioned in the literature over the last thirty years.
    Of course note most lectures don’t cover complier exact issues. Gcc testsuite warns about it in the form these tests may fail to product optimized code in case of out of ram. Other than that its undocumented. Only people like me who have build gcc and read the testsuite usage instructions are the ones who mostly know it. Yes google developers making android did read the gcc testsuite.

    Let’s anthropomorphize a sample candidate just for the purpose, shall we? Imagine you are the front end of a C compiler. Essentially you are a parser backed with a lexer.

    Really we don’t need to. There is stupidity of gcc. Front end section of gcc is not just bison and flex it also contains optimizer. Because someone thought this was a great Frontend code Optimizer is free to go and search though the raw source files again searching for extra instructions that the parser did not include. Of course this code is free to not run if memory is low because the theory was latter steps in gcc hopefully should pick up these optimizations(guess what they don’t). Gcc is not the most well written bit of code. Basically the front end of gcc is not cleanly separated as any book to write complier tells you it should be.

    Yes at times you do stupid things to match up to how the real world compliers perform. Stripping comments out of headers before running gcc over them is one of those stupid optimizations because the complier contains a odd design selection so it works. Most other compliers other than gcc the alteration would have been pointless.

    And no emulators, please. You’re going to have to provide a cite that describes how somebody — given familiarity with gcc and if you like other parts of the Gnu tool chain — can build an executable using gcc and glibc on an actual Android device.
    https://play.google.com/store/apps/details?id=com.n0n3m4.gcc4droid&hl=en
    Email that guy and ask for the source code then I can find out who you are. I did include a link where you could get the information todo it just you are being a stubborn and asking me for a cite to cover what was already covered.

    Deafspy already asked that question DrIdiot and I had already given a cite covering it.

    There is no such thing as Gnu/Linux/Android.
    Really I already told you that GNU is GNU is Not Unix. So talking about Posix to make a point that GNU cannot be Android is pointless.

    Exactly what says GNU is glibc. Glibc is only a subpart of GNU list of possible parts.

    The bare min to have title GNU is using Gcc complier. So yes there is such thing as GNU/Linux/Android it just show how vague the GNU title is. Since android uses the gcc complier you have to check the GNU mailing list for security flaws for gcc built code. As I said Android out box is not GNU/Linux/Android. But after you have added gcc and binutils to build the core drivers on the device then its GNU/Linux/Android.

    Yes people use GNU/Linux to describe installs using musl libc with gcc complier. So glibc is not a required define factor to the GNU title DrIdiot. To be correct some of the musl libc Linux systems that people call GNU/Linux the only GNU things they contain are gcc and binutils. So that defines the bare min for the GNU title.

  31. Wizard Emeritus says:

    ” Living a cubicle life, contributing nothing really at all is a rather dismal and bleak life.”

    And what have you contributed little man?

  32. Wizard Emeritus says:

    “Better late then never!”
    ReactOS? This ReactOS?

    http://liliputing.com/2014/11/reactos-inches-closer-becoming-true-windows-2000-clone.html

    And I quote…

    “Despite decades in development, ReactOS is still considered alpha software that’s recommended for testing and evaluation purposes. You probably won’t want to replace your primary operating system with ReactOS anytime soon.”

    You never cease to amuse Dougie.

  33. dougman says:

    REACTOS Recent Related News
    • 2016-02-17: OS Release: 0.4.0

    Better late then never!

  34. dougman says:

    Re: None of you have a clue what I’m talking about, do you?

    See, the funny thing is, in your case. No one actually cares. Sad, but true. Living a cubicle life, contributing nothing really at all is a rather dismal and bleak life.

    Dr. Tosser is a perfect example of Basil Fawlty, in that he famously is known for losing his rag and beating his car with a tree branch.

    WACK…WACK…WACK…out you demons of stupidity!

  35. DrLoser says:

    You know, Robert, the more I look into this — and thanks for the article, because I had never even thought of looking into it before — the more obvious it becomes.

    There is no such thing as Gnu/Linux/Android.

  36. DrLoser says:

    Never mind. Let’s not let Dog-Brain distract us from the interesting question of whether Bionic/Android has any particular relationship to Gnu/Linux.

    We have discussed the lock-down of Android not letting a FLOSS user substitute glibc. This is of imperative importance to all of us, I feel. The discussion goes on.

    At a slightly (though not very much) deeper level, does anybody here on this blog have an opinion on how Android handles the Posix standards?

    These standards are important too, you know. After all, they are open standards.

    And indeed, for those of us who have spent twenty years or more in a multithreaded environment, the Posix standards practically define the way a portable *nix program is supposed to work.

    Any takers on this rather vital bit of FLOSS?

    (I will graciously assume that you have all downloaded, read, investigated, improved and redistributed Posix C code. After all, this is a big part of what makes us free.)

    Isn’t it?

    None of you have a clue what I’m talking about, do you?

  37. DrLoser says:

    You really are a worthless imbecile, aren’t you, Dog-Brain?

  38. DrLoser says:

    While you’re at it, Princess, consider the front-end of a compiler. Let’s anthropomorphize a sample candidate just for the purpose, shall we? Imagine you are the front end of a C compiler. Essentially you are a parser backed with a lexer.

    Bison (for it is Free): This parsing stuff is a breeze, innit? All I have to do is to accept a stream of tokens and run them past a bunch of rules. I feel so l33t!
    Flex (for it, too, is Free): Not so fast, little parsing ninny! I can really throw a spanner into the works! I’m in control of the token stream, you know!
    Bison: You are my inferior. You will do as I say. You have no knowledge of the Big Wide World of memory and optimization and slowly parboiled frogs!
    Flex: Not true. I cannot influence them directly, but I can cause you to do so.
    Bison: You, little red-leather-skirted nincompoop*? I doubt it. You can’t even spell “compiler” correctly, can you? What do you know of RAM?
    Flex: Not much, but I can get you to chew it up like crazy! All’s I need to do is to throw a gazillion little tokens inside LEFT-COMMENT and RIGHT-COMMENT, and you’ll choke! You’ll eat up all the RAM and leave nothing left for the Optimizer!
    Bison: Gee, you got me there. I’m up shit creek without a paddle, aren’t I?

    Well, that’s the way that Fifi seems to think that the conversation would go. There’s only one teeny, tiny, little flaw in this argument.

    First one to spot it Wins The Intertubes.

    *I’m doing Flex a rhetorical disservice here by comparing it to Fifi. Unlike Fifi, Flex is very useful and also knows what it’s talking about.

  39. DrLoser says:

    Correct comments don’t come part of final code but they can cause final binary not to be optimized properly due to consuming ram.

    Ah, an interesting limitation in that part of the back-end of a compiler that deals with optimization.

    Curiously enough, the lecturers at Cambridge never mentioned this one. Nor have I seen it mentioned in the literature over the last thirty years.

    Please help me, little Googlicious one! A cite, a cite, my optimizer for a cite!

  40. DrLoser says:

    There are many examples in the google play store and the instructions to build gcc using bionic go back to Android 1.0 to build drivers platform matched the hardware as best as possible.

    Cite one. No need to read it — we’ll do that for you. Just pick a keyword or two. Don’t even bother spelling them correctly — both Google and Bing handle excruciating mis-spellings very well indeed.

    And no emulators, please. You’re going to have to provide a cite that describes how somebody — given familiarity with gcc and if you like other parts of the Gnu tool chain — can build an executable using gcc and glibc on an actual Android device.

    No cite? Then it’s fair to say that an Android device is basically locked down by Google. And it’s also fair to say that:

    There is no such thing as Gnu/Linux/Android.

  41. DrLoser says:

    So you will keep on insulting a person who is a current standing professor.

    Since when did you acquire a PhD, Fifi? Back of a bubble-gum packet, possibly?

    Because you’re the only one we’re “insulting” — or more accurately calling out as a complete and utter ignorant fraud.

    If you’re referring to a guy who was in the first stages of a Masters in 2013-2014 (the dates of your cite. I mention this because you do not read your own cites. You are presumably unaware of the date of publication), then he’s climbed the greasy pole of academia remarkably quickly, hasn’t he?

    Oh, and one more thing. I reserve the right to question, disagree with, or indeed roundly insult anybody and everybody, contingent only upon the views they express. Arguments from authority do not wash, unless you’re a star-struck little girl in a red leather dress standing under a lamp-post in an outback town.

  42. oiaohm says:

    Deaf Spy that is how platform tweaking for android performance is done.
    https://play.google.com/store/apps/details?id=com.n0n3m4.gcc4droid&hl=en
    Now, now, Fifi. You claim that you can use GCC to compile source under Android. Not in an emulator like KBOX, Fifi. Natively.
    Yes Gcc and binutils runs natively on Android using bionic libc.

    There are many examples in the google play store and the instructions to build gcc using bionic go back to Android 1.0 to build drivers platform matched the hardware as best as possible. Yes the first instructions todo this were from Google. Today it open to a lot more people just installing application from google play. So this is not a claim it is reality of Android. It is technically possible to rebuild every part of Android on Android.

    The fact you have to question me about this show you just followed DrLoser idiot in boots and all.

    What’s up, Fifi? Do your googling skills fail you? You are clueless again, aren’t you?
    Microsoft in CE and the like never offered on tightly restricted embedded platform compliers with the ability to run diagnostics to work out what asm options are faster while building the code. Device OS’s Microsoft has not historically done this. Cut back OS with Microsoft equaled cross compile only. Google did not follow this path.

    Microsoft compliers don’t in fact include auto platform diagnostics that gcc has. Why Intel has patented it and Microsoft compliers cannot legally contain it. It was Intel that gave it to gcc. So idiot Deaf spy I know my point. Microsoft simple does not have the tech todo it. Microsoft cannot teach you todo something they don’t have the tech todo.

  43. Deaf Spy says:

    Now, now, Fifi. You claim that you can use GCC to compile source under Android. Not in an emulator like KBOX, Fifi. Natively. Care to prove this? Because, you know, I am inclined to say you pull this stuff out of where the sun don’t shine.

    So its only stupid to you because you an idiot who does not know how gcc works… Microsoft does not do binary optimized exactly for a particular devices hardware configuration.

    What’s up, Fifi? Do your googling skills fail you? You are clueless again, aren’t you?

  44. oiaohm says:

    Deaf Spy
    Haven’t you heard of virtual memory, you illiterate, ignorant urchin?
    Other than most Android devices not having virtual memory and no means to set it without rooting the device. Applying virtual machine will alter the produced code by gcc in -mtune=native mode optimization mode. So this is not a option you big idiot.

    I am not sure this is possible at all. But, since I am not into the internals of Android, I will graciously assume it is. Then, this is the most stupid thing to do.
    Gcc contains code to do optimization to the produced binary based on the CPU/platform it is running on (-mtune=native -march=native triggers it). So its only stupid to you because you an idiot who does not know how gcc works.

    Someone should learn from Microsoft how to properly do this stuff.
    Of course you cannot learn this from Microsoft. Microsoft does not do binary optimized exactly for a particular devices hardware configuration.

    Sorry Deaf Spy your the idiot here. Virtual memory is not an option. You did not understand the complier to understand what was required to make optimized to target binaries for Android. Every one of your counter idea is bogus.

  45. Deaf Spy says:

    Fifi, you are a masochist, dear, aren’t you? You enjoy being humiliated, don’t you? Wasn’t all the trashing you received already not enough?

    You didn’t get writeln() and now you are up to your next utter idiocity.

    Correct comments don’t come part of final code but they can cause final binary not to be optimized properly due to consuming ram

    Haven’t you heard of virtual memory, you illiterate, ignorant urchin?

    Of course not but Android developers could be rebuilding applications on the device it self with limited ram.

    I am not sure this is possible at all. But, since I am not into the internals of Android, I will graciously assume it is. Then, this is the most stupid thing to do.

    Someone should learn from Microsoft how to properly do this stuff. I will leave it to your astounding googling skills to find out what I am talking about.

  46. oiaohm says:

    Comments do not become part of the compiled code
    Deaf Spy LOL. You don’t know Gcc when it runs short on ram it will skip optimization steps. Correct comments don’t come part of final code but they can cause final binary not to be optimized properly due to consuming ram. Comments are not a 100 percent no-op factor. So comments + gcc can cause some nasty effects in resulting binary.

    https://play.google.com/store/apps/details?id=com.aide.ui&hl=en
    Android compiles C++ libraries from source when cold-starting applications
    Of course not but Android developers could be rebuilding applications on the device it self with limited ram. Android was designed that Android could be used to rebuild Android. This was mentioned in CAVEATS and since ides that directly run on Android have appeared.

    Basically your –Very simple– is attempting to apply common sense to something that does not in fact obey the rules the way you think.

  47. Deaf Spy says:

    Deaf Spy explain how that does not work.

    Very simple, Fifi. Comments do not become part of the compiled code. I hope you will not tell me that Android compiles C++ libraries from source when cold-starting applications 🙂

  48. oiaohm says:

    Deaf Spy does not change the point that what is stated there is a valid optimization.

    Just you want to poke fun at someone paper without understanding what it says.

    Deaf Spy different documents of android have disappeared because google does re export it git repositories after deleting the git repositories outside google.

    An 11-page observatory paper is anything but solid, Fifi. In academic societies, where your chances to even be allowed to look through a basement window on a rainy street”, this will be laughed upon and discarded as a joke. This is not a paper. This is a homework.
    That something is homework does not make the statements in it incorrect.

    Notice that DrLoser is no longer pushing this point.

    “A first advantage is the optimization of code due to the removal of all comments from the header files. This reduces the size of all those headers and helps with a speedy process in a storage-constricted space.”

    Deaf Spy explain how that does not work. In fact you cannot because the block of text you are picking on like it or not is correct and you are an idiot for nit picking on it. Storage-constricted spaces include building on like a raspberry pi with like 256 megs of ram. Stripping the header files in that case will allow applications to build that would have failed due to running out of ram.

    So like it or not only a idiot would attempt to argue that its not a valid advantage.

  49. Deaf Spy says:

    Mathieu Devos made solid source paper like it or not.

    Bwaha-ha-ha-ha!

    An 11-page observatory paper is anything but solid, Fifi. In academic societies, where your chances to even be allowed to look through a basement window on a rainy street”, this will be laughed upon and discarded as a joke. This is not a paper. This is a homework.

    Disappearing documents… Do you also see disappearing pink elephants, Fifi? Or appearing white flying pigs?

  50. Deaf Spy says:

    Again, especially for Fifi:

    “A first advantage is the optimization of code due to the removal of all comments from the header files. This reduces the size of all those headers and helps with a speedy process in a storage-constricted space.”

    Does it hurt, illiterate, simple, ulcerous and lice-ridden little one?

  51. oiaohm says:

    Deaf Spy in fact there was zero faults in the cited document.

    DrLoser made out a fault that does not exist. Citing a document for evidence of a document that has gone missing does not mean I had to care about the content that is evidence of a missing document.

    No good, Fifi, no good. You cited out a document. I don’t care what part of it. I quote something exactly from the document. A missing document.

    Also there use to be a CAVEATS file in the Bionic source that stated the reasons.
    http://irati.eu/wp-content/uploads/2012/07/bionic_report.pdf
    Was there as recently as 2012. Disappeared since.

    Read this very carefully. Deaf Spy at does this say I am using the contents of that cite. No I am using it as evidence to a missing file.

    In a university attempting to nit pick a source that is only evidence of a document that has gone missing.

    You cited from a document, which is utter crap.
    So you will keep on insulting a person who is a current standing professor. The document is not crap. Never was crap. Just it was not important to my arguement.

    Mathieu Devos made solid source paper like it or not.

    It about time you say sorry Deaf Spy because you are insulting a document that does not in fact have flaws. Displays some strangeness in Google logic listing all the documented reasons.

  52. Deaf Spy says:

    I most likely skipped the first paragraph due to …Also that was not what I had cited out the document.

    No good, Fifi, no good. You cited out a document. I don’t care what part of it. You cited from a document, which is utter crap. Dear, if any of my students hand me over such a homework, I will make sure they don’t even get to the exam. A thesis for Masters? No chance whatsoever.

  53. oiaohm says:

    If you have copy of libc/CAVEATS. The header comment stripping was done to improve performance of Gcc under Windows. It does in fact make a difference in older gcc running on windows altered the build time by 15 percent by CAVEATS claim. Almost no advantage on OS X and Linux building. Yes it one of those strange things about Windows memory management meeting gcc and odd things happening. So the claim does match up to something google developers did state not something I class as important. This is why it was kinda important to read CAVEATS before presuming anything in that write up was wrong.

  54. oiaohm says:

    Deaf Spy and DrLoser both owe Mathieu Devos an apology. For being idiots.

    Ok I claimed something was not Mathieu Devos was not his work when it was. On page 6 I most likely skipped the first paragraph due to scrolling down on computer screen and saw the other 3 points that matched exactly what Google had released. And it was not like trimming header files as a optimization note required any special attention in my memory. Also that was not what I had cited out the document.

  55. oiaohm says:

    Android out box is not Gnu/Linux/Android is correct. Android after installing a few applications maybe.

    What a very, very, stupid concept. You could say the same thing about Windows, Princess. And it would be equally meaningless.

    Another idiot attack by DrLoser. From a security point of view defining what sources mix up a solution allow you to know what security reports to be watching. Guess what / defines are for. Mostly that. This is why not all Linux Distribution are GNU/Linux because not all them contain GNU core parts.

    As I mentioned, the Android application layer isn’t Unix in any real sense.
    Please note he says Application Layer. This has a meaning that is critical to be aware of. Application Layers on Linux are mostly not Kernel API. Java is a application layer.

    Now this is from “2008-11-10” lot has changed.
    http://elinux.org/Android_Mainlining_Project
    So other than binder and a other few odds and ends you don’t need to call the Android kernel is close to stock. So any linux libc will work as long as long as does not demand fixed paths or you can turn off demand for fix paths.

    Linux is like Windows in the fact it does support running more than I subsystem at the same time.

  56. oiaohm says:

    Chapter 2
    Advantages of bionic
    A first advantage is the optimization of code due to the removal of all comments from the header files. This reduces the size of all those headers and helps with a speedy process in a storage-constricted space.

    Now that is correct quoted.

    DeafSpy and DrLoser cannot not read. Reducing size of header files in fact does help in code building. Its a dirty trick to build an application when you don’t have enough ram. The advantages the the author writes about are not a mistake. But when someone has to choose that code advantage.

    I should not have been asked to defend something that is correct and is just stating a fact. The reality that at max there are only 4 advantages and that is if you include the header reduction. Security is not one of them because the early android libc is using older unmaintained code than glibc or any other in use libc at the time.

    Given that you don’t have the courtesy of reading your own (irrelevant, in this case) cites, why should we offer you the same courtesy?
    I gave DeafSpy a chance to read the document and change mind and not follow you idiot DrLoser.

    So you are both determed to insult Mathieu Devos.

    I don’t care which part of it you had in mind. This specific paragraph is a perfect illustration of the quality of your sources, and the quality of your claims.
    Sorry the quality of my source is not at fault here. Quality of DrLosers nit picking and your idiotiness and following it is. DeafSpy.

  57. DrLoser says:

    I note that John Foxx, founder member of the very influential Modern Romantic Floss band “Ultravox,” possesses a PhD, Fifi.

    Perhaps you could consult his Wisdom on your future cites?

    God knows, you’ve been doing far worse, recently.

  58. DrLoser says:

    Look closer is from a person with a PHD.

    Deaf Spy: And who that unfortunate fellow might be?
    Me: I suspect that unfortunate fellow is Fifi, Deaf Spy. It’s a reasonable guess. He was indeed the unfortunate fellow who supplied a cite that proclaims at the top that it is a “Bionic vs Glibc report: Master thesis,” which pretty much obviates the remote possibility that it is Mathieu Devos.

    I’m also guessing that Fifi is using the abbreviation to mean something else altogether. Now, given that the honoree is Fifi herself, we can ask ourselves a very simple question. Does she mean:
    1) Pretty Handsome Dude or
    2) Paranoid Hater, Delusional.
    Place Bets Now!

  59. DrLoser says:

    Android out box is not Gnu/Linux/Android is correct. Android after installing a few applications maybe.

    What a very, very, stupid concept. You could say the same thing about Windows, Princess. And it would be equally meaningless.

    It’s been a while, I admit, but Matthew Garrett’s thoughts on the matter are still worth quoting. Read the lot. It’s instructive. And not to cherry-pick, but I think we can all agree on his following assertion:

    As I mentioned, the Android application layer isn’t Unix in any real sense.

    Or perhaps we can’t. In any case, just in case those of you (all of you) with a total lack of a sense of humour read the following:

    To a certain extent, my hopes were fulfilled. We got a git server in California.

    Please do not take this as a literal “fulfilment of Four Freedom hopes.”

    Because it isn’t. It’s bitter sarcasm.

  60. DrLoser says:

    Perhaps I should put the quote into context for our suddenly word-blind little friend.

    Chapter 2
    Advantages of bionic
    A first advantage is the optimization of code due to the removal of all comments from the header files. This reduces the size of all those headers and helps with a speedy process in a storage-constricted space.

    Do you have a problem reading as far as page 6 (or Chapter 2) of your cites, Fifi? It wouldn’t surprise me.

    Given that you don’t have the courtesy of reading your own (irrelevant, in this case) cites, why should we offer you the same courtesy?

  61. DeafSpy says:

    The cite I used does not in fact contain this.

    Oh yes, Fifi, it does, it does. Chapter 2, first paragraph.

    You can twist and turn, Fifi, but it is you who brought this link to the discussion. I don’t care which part of it you had in mind. This specific paragraph is a perfect illustration of the quality of your sources, and the quality of your claims.

    Be a man, Fifi. Be a men. Confess. Confess about writeln(), too. Ask forgiveness and you shall be forgiven.

  62. oiaohm says:

    Deaf Spy get better of course.

    The cite I used does not in fact contain this.
    A first advantage is the optimization of code due to the removal of all comments from the header files. This reduces the size of all those headers and helps with a speedy process in a storage-constricted space.
    http://irati.eu/wp-content/uploads/2012/07/bionic_report.pdf
    Go a head search DeafSpy. The first person to write that block of text here is DrLoser.

    It not in the https://en.wikipedia.org/wiki/Bionic_%28software%29 wikipedia cite that DrLoser used either.

    What happened here is DrLoser has made crap up out of mid air as normal and moron idiot DeafSpy has fallen for it hook line and sinker.

  63. oiaohm says:

    Deaf Spy part of the reason for so called “schizophrenic claims” is you taking claims written by DrLoser and others and assign them to me. About time you learn to go after the right persons.

  64. oiaohm says:

    Deaf Spy really idiot question.
    http://irati.eu/wp-content/uploads/2012/07/bionic_report.pdf
    Number 1 was in fact referring to libc/CAVEATS with that cite that had gone missing. It was DrLoser who cut it out of context out the PDF. And made a claim based on it.

    You quoted this crap as proof to your schizophrenic claims. Now, be a man and stand behind them. Explain how cutting out comments from a source code can result in an improved storage requirements for the runtime.
    No I cited a document I did not quote from that document at all. Since I did not quote from the document I don’t have stand behind what it says at all.

    The person who quote the it is DrLoser. So please do the correct thing and challenge the correct person or get lost.

  65. Deaf Spy says:

    Sorry, Fifi, you must learn to bear the burden of your own claims and proofs you bring to the discussion. There is no excuse to quote this hilarious first-term homework, masqueraded as master’s thesis. Three people to produce 11 pages of static observations… Dear, dear.

    Back to you, Fifi. You quoted this crap as proof to your schizophrenic claims. Now, be a man and stand behind them. Explain how cutting out comments from a source code can result in an improved storage requirements for the runtime.

    Look closer is from a person with a PHD.

    And who that unfortunate fellow might be?

  66. oiaohm says:

    Deaf Spy in fact I did not write that line at all.
    http://irati.eu/wp-content/uploads/2012/07/bionic_report.pdf

    Look closer is from a person with a PHD.
    Now, this is something. This is the Titan of Fifi’s Stupidify.
    Really its the level of your stupidity that when Drloser said fifi cite you read it as something I wrote and attacked me.

    I have got up DrLoser repeatedly for not correctly citing stuff. If he had of cited the correct documenting in instead of being a smart ass this would not happened.

    If Deaf Spy was not being such a troll attempting to attack where ever I write this would not have happened either.

    So really both of you have just prove yourself as twits to Robert and about time both of you stop it.

  67. Deaf Spy says:

    A first advantage is the optimization of code due to the removal of all comments from the header files.

    Now, this is something. This is the Titan of Fifi’s Stupidify.

    I do wish Robert and Luvr can read this.

  68. Deaf Spy says:

    Hey, Robert, pay close attention here:

    There is no such thing as Gnu/Linux/Android.

  69. oiaohm says:

    libc/CAVEATS
    DrLoser That file documents why bionic libc is not a full libc.

    Small size: Bionic was much smaller than the GNU C Library; more importantly its memory requirements were (and remain) much lower.
    DrLoser this is the second point in the wikipedia that you over look. Small memory foot print is important. CAVEATS go through detailing cutting long list of remove functionality to save on memory foot print. And just to be wrong most this is functionality that musl has implemented with a lower memory foot print than bionic.

    Speed: Bionic was designed for CPUs at relatively low clock frequencies.
    Third point I had forgot about was lower cpu usage. Again Bonic bionic libc is heavier than musl on cpu usage as well.

    Since that was my sole proposition, now proven beyond doubt even on your own cite, Fifi — I have no reason at all to respond to the rest of your pointless off-topic crap.
    DrLoser this was not your claim. Your claim was Android was not effected by glibc flaws that is not in fact the case if you look at what is in android app store and what developers have done.

    Yes getting dynamic linking to glibc to work under android also means binding to closed source is possible

    DrLoser stop trying to win by doing 1/3 of the proper quote.

    Out of the proper quote on the wikipedia
    I think the unsafe bit here might have played more of a role in Google’s choice for Android.
    Has no grounds. Fixing the unsafe bits was not the reason why Google started bionic libc it was after thought since they were building a libc anyhow.

    There is no such thing as Gnu/Linux/Android.
    Sorry this is not true either. The fact Gnu parts are found on top of Linux/Android at times that term is wrong.

    Android out box is not Gnu/Linux/Android is correct. Android after installing a few applications maybe.

    http://www.gnu.org/licenses/gcc-exception-3.1-faq.en.html
    The most widespread standard C library for the Linux kernel is the GNU C Library (glibc), which is subject to the GNU Lesser General Public License (LGPL), also a copyleft license. In contrast to the GPL, the LGPL explicitly allows for dynamic linking but it does not allow static linking of proprietary software.
    This is a claim Google made for making bionic libc but its not true. glibc allows static linking against proprietary software as long as you still provide those down stream with access to the glibc source.

    DrLoser basically please stop using incorrect history. Google had 3 reasons for their libc in Android. There is no guess work to it. License, Memory foot print and Speed. Everything else is suger coating. License is also part miss interpretation of license.

    If google had not miss understood what LGPL allowed google might have gone uClibc. Again more functional than Bionic and lower memory foot print than Bionic and lower number of bugs than Bionic. musl starts in 2011 done better than uClibc, dietlibc, Bionic and glibc because the developer of musl thinks all three are junk.

  70. DrLoser says:

    Interesting little lemma, though.

    Not only do those items factor into this decision but also faults in glibc such as buffer overflows, unsafe handling of strings and memory functions.

    Buffer overflows, and unsafe handling of strings and memory functions? A mere trifle.

    But let’s assume that you are God’s Gift To FLOSS. You have three ways of tackling these issues in this frankly utterly essential GNU component:

    1) Do what Debian does. Admit that you are worthless in this domain and cannot significantly improve what is handed down by your Lords and Masters at Red Hat.
    2) Do what Google should have done, according to FLOSS principles.
    2a) Run it.
    2b) Study it.
    2c) Eat your own toe fungus. (Everybody needs a break from hard work.)
    2d) Improve it.
    2e) Redistribute it.
    3) Chuck it away as a useless broken and restrictive bit of Gnu/Linux rubbish, and build your own.

    Well, even though Google might briefly have considered being God’s Gift To FLOSS, it appears that they chose option (3) rather than option (2). Bit of a shame that, because clearly Google have the resources and the finances to handle improvement and distribution of glibc, if only they had the will.

    Goodness knows why. Maybe Californians aren’t as keen on eating their own toe fungus as a hermit in a broom cupboard in MIT.

    This is not your father’s 1960s, is it?

  71. DrLoser says:

    I just found this little gem in Fifi’s cite:

    A first advantage is the optimization of code due to the removal of all comments from the header files. This reduces the size of all those headers and helps with a speedy process in a storage-constricted space.

    Well, that sounds pretty darned authoritative and convincing to me, Princess.

    Bwahahahahahahahahaha!

  72. DrLoser says:

    It also brings us back to the VMWare thing.

    Is it really permissible to mix and match bits of BSD with bits of Linux (whilst withholding the source for large amounts of the result)?

    Perhaps it is. Perhaps Google’s lawyers have carefully crafted some devious excuse or other.

    In which case they are basically in the same boat as VMWare, I would think. With the exception that the bits of Linux supposedly abstracted by VMWare are quite trivial, whereas the bits of BSD bolted on to a pseudo-Linux Android base seem, to me, to be quite central to the working of the system.

  73. DrLoser says:

    Though, to be fair, BSD is not so much “not Gnu” as “not Linux.”

    Perhaps you would see that as an improvement, Fifi?

    What with Google fighting hard to drop all those inconvenient FLOSS mandatory requirements and all.

    I commend the BSD license to any serious practitioner of IT.

  74. DrLoser says:

    Serous-ally Half a meg of memory usage that glibe triggers is insane.

    Evidently not half as insane as you are, Fifi. You cite an eleven-page Master’s Thesis. (Apparently standards are slipping in Ghent. Only eleven pages?) I have no idea why you cite this. It does not appear to be written by the author(s) of Bionic. And, searching this paltry two page PDF for the word “memory,” we find:

    1) “Due to limited storage and cpu speeds an adaptation was required before google could implement the library into android.”

    Oh yeah? No further information provided. Half a meg is half a meg. And what CPU speeds might have to do with the matter is entirely beyond me. And look!

    1a) Parts for this library were found in both glibc and in the BSD library. Not only do those items factor into this decision but also faults in glibc such as buffer overflows, unsafe handling of strings and memory functions.

    I think the unsafe bit here might have played more of a role in Google’s choice for Android.

    2) “[STL] initiation of templates increases memory usage.” Obviously not relevant.

    One other point of interest from the PDF does spring out, however:

    Bionic is an adaptation of the BSD1 standard C library. It is used on the android platform and replaces glibc.

    This is emphasised further by the Wikipedia link for the thing:

    The original publicly stated goals for Bionic were the following:

    * BSD-licensed: Google wanted to isolate Android applications from the effect of copyleft licenses to create a proprietary user-space and application ecosystem.

    Like I said in the first place, Fifi.:

    There is no such thing as Gnu/Linux/Android.

    Bionic confirms this very obvious fact. Your link does nothing to deny it. And Wikipedia’s sources — impeccable as always, I would assume — put the final nail in the coffin.

    There is no such thing as Gnu/Linux/Android.

    Since that was my sole proposition, now proven beyond doubt even on your own cite, Fifi — I have no reason at all to respond to the rest of your pointless off-topic crap.

    Hallelujah!

  75. oiaohm says:

    I have no particular clue why Google chose to dump glibc on Android, Fifi, but I very much doubt that a reduced footprint in memory was the deciding factor.
    DrLoser basically shut up you are idiot on this topic. The developer of Bionic libc at early Android Conference started the reasons. Also there use to be a CAVEATS file in the Bionic source that stated the reasons.
    http://irati.eu/wp-content/uploads/2012/07/bionic_report.pdf
    Was there as recently as 2012. Disappeared since.
    1) License.
    2) Memory.
    CAVEATS described why particular things were done.
    Serous-ally Half a meg of memory usage that glibe triggers is insane.

    Google has in fact told us why with Android they have not gone glibc. The reasons are all technical. DrLoser you really should not be posting about topics you know nothing about then calling me fifi.

    Can you tell the difference between a megabyte and a gigabyte? Yes I can and you are such a idiot you don’t understand how far out the ball part 1/2 meg just to start up a libc with a hello world program is.

    Also DrLoser you are such a idiot that you are not asking the right questions exactly what glibc is doing with that 1/2 meg. If you were you would have some interesting cases.

    Hello world full memory usage at worst
    musl 20kb .
    Visual Studio lib C 100Kb
    Bionic libc 200kb.
    Glibc 400-700kb depending on version.
    Since glibc version 2.3.4, the glibc wrapper function for getpid() caches PIDs, so as to avoid additional system calls when a process calls getpid() repeatedly.
    Welcome to one of many glibc surprises for consuming ram. Causing way heavier memory usage.

    And can you install it over Bionic without, to repeat, paving over the Android OS? I can’t imagine it would be very easy. Do please provide the steps to do so.
    Also idiot answer. You don’t need to install over current libc to use another libc next to it. If you use libhybrid you technically drop glibc in instead of bionic libc but this is not what causes the security issue.

    Method 1 for Glibc to end up on android device: Static Link.
    http://headlessandroid.blogspot.com.au/2012/02/building-your-favorite-glibc-app-for.html
    Statically link the binary on a standard Linux system as per normal for arm. Then it runs under android as per normal. Now glibc is welded into the binary that is now shipped out by google play store along with all the other libraries that were welded in.

    DrLoser the reality here is you never tried you presumed it was hard when its easy other than the heavy file size cost if developer chooses static link. Yes it a lot simpler to a static link than port the same code to bionic libc.

    Method 2: Dynamic link that is a little harder problem does not require chroot at this stage either.
    http://forum.xda-developers.com/showthread.php?t=1299962
    Reading stuff this you would kinda think so.

    People forget you can direct execute ld-linux.so itself with the executable you wish to run following it and more fun the -z nodeflib flag to build program. -z nodeflib means don’t check /lib /usr/lib heck don’t even bother check /etc/ld.so.conf and ld.so.cache add on more flag -rpath,’$ORIGIN/lib. What you have now built is a standard I don’t care about host dynamic binaries files Linux program. Place glibc and all require libraries for application /lib relative to executable and almost done.

    Some games on Normal Linux use a shell script to call there own ld-linux.so on their executable. Of course they are coded this way. Of course developers that were doing this to release stuff on Linux normal have done same thing on android.

    So as long as your android app you are shipping to customer use the correct ld-linux.so directly to run the glibc application on android everything is fine no system wide modification required. Yes more space effective this way if you are using multi applications inside your android application needing glibc.

    Both static linked and dynamic linked to glibc exist in the google play store Android Applications.

    The fact your core APK file is allowed to be 100MB these days you can be a bit sloppy and get away with it.

    Of course personally I would be recommending musl and its loader if you can. Smaller and no license issues.

    Whether some poor moron has decided to build upon this piece of crap and compromising himself when there are pretty better runtimes available for Windows is purely his own responsibility.
    DeafSpy security is about attempt to make sure people don’t get the wrong idea. If glibc is pushed as only a Linux issue. The person you called a Moron is a moron because in believes your stupidity that glibc is not Windows problem. The fact something exists on a platform does not matter if the OS maker ships it or not means you should worry about it from a security point of view.

    Same arguement was made about openssl that windows was fine because it had it own TLS library only turned out about 70 percent of applications using TLS were not using it.

  76. DrLoser says:

    For a long time in compact memory space glibc has been avoided. Firing up glibc costs you half a meg of memory. Yes google went and made a new libc with cut back features on Android to attempt to be memory light.

    Memory light? I have no particular clue why Google chose to dump glibc on Android, Fifi, but I very much doubt that a reduced footprint in memory was the deciding factor. Can you tell the difference between a megabyte and a gigabyte?

    Apparently not.

    Bwahahahahahahahaha!

  77. DrLoser says:

    I did not say pave over. There is such thing as glibc for android without paving over or running a chroot.

    And can you install it over Bionic without, to repeat, paving over the Android OS? I can’t imagine it would be very easy. Do please provide the steps to do so.

    And even assuming that you are idiot enough to do so (not much of a stretch in your case, Fifi, I admit), why would you want to?

    And even if you wanted to, the fact remains that Google have not chosen to go with glibc on Android. Which means that there is no such thing as “Gnu/Linux/Android.”

    You’re trying to evade my simple claim here, Fifi. It won’t work. I’m pointing out reality. You are just pithering around with every irrelevant link you can dig up.

  78. DeafSpy says:

    Fifi, glibc is not part of Windows, nor it is any officially supported API, period.

    Whether some poor moron has decided to build upon this piece of crap and compromising himself when there are pretty better runtimes available for Windows is purely his own responsibility.

    Fifi, did you learn how writeln() in Pascal works, you illiterate joker?

  79. oiaohm says:

    http://gnuwin32.sourceforge.net/packages/glibc.htm
    DeafSpy there are some vile versions of glibc you find shipped for windows. You find them embedded in applications that have been ported from Unix or Linux. Sorry DeafSpy another total crap. Basically glibc issues are like openssl on a smaller scale on Windows and OS X. People looking after Windows and OS X systems should audit the applications they are using to make sure they are not using stuff like gnuwin32 very old and major-ally security flawed glibc. You are not going to find 2.2.5 glibc unpatched on any maintained Linux Distribution. But you could find glibc 2.2.5 on Windows 10 or OS X.

    Did I say it was? However, should you choose to run a Linux distro of practically any kind (is there still one out there that relies on a fork of glibc? There might be one. I don’t know), then you are choosing to run glibc.
    This is incorrect. There are linux distributions that don’t depend on glibc at all not even a fork.

    Musl libc does not contain a single line of code from glibc. Is an option when using of the more edge distributions to use pure Musl libc.

    The fact remains, Fifi. That Google chooses not to rely on glibc — a central part of Gnu, I would think — is pretty solid evidence that ANGLE.*
    Sorry chromebooks come to mind ChromeOS made by Google do use glibc when they could have used Musl libc.

    For a long time in compact memory space glibc has been avoided. Firing up glibc costs you half a meg of memory. Yes google went and made a new libc with cut back features on Android to attempt to be memory light. Stupid enough Bionic libc uses more memory than Musl while providing less functionality.

    For under researched idiots like Drloser. There are 5 libc you find around Linux.
    Bionic libc of course is android only. That leaves 4.

    klibc is specialist for initrd and other places were compact and not feature complete libc is good enough. That leaves 3.

    newlib is a cross platform libc so feature limited. Now that leaves 2.

    Glibc and musl are the two left both aiming to be full libc and both having distributions.

    So yes it would be valid to open debate if more distributions should change default to Musl over Glibc. Interesting enough Musl is MIT license, quite compact and says stuff putting version symbols in .so files in a lot of cases fixes Forwards compatibility issues where new program fails on old system.

    It is correct that eglibc is no more but mostly that once pressure from Musl came the guys running glibc and eglibc thought they better sort out their deferences.

    And those choosing to pave over Android and replace it with a Linux distro will also be at risk of glibc issues. This should hardly come as a surprise. Those choosing to swallow their mobile phone or tablet will similarly have made a potentially injurious choice.
    Really heavy memory usage Linux Distrobution on Android device most likely not good. Light memory using Linux Distrobutions will mostly be all musl not glibc.

    I did not say pave over. There is such thing as glibc for android without paving over or running a chroot. Just like there is glibc for windows and OS X.

    DrLoser spreading disinformation that glibc is the only choice really is prevent people going looking at distributions that use musl at core.

    A not point in Linux history has there ever been the case that 100 percent of distributions are using glibc there is always a percentage using alternatives.

  80. DrLoser says:

    I regularly build my kernel and a few applications where I want the latest improvements.

    Which improvements would those be, Robert? Something tangible, such as systemd, or some wafty little bit of unproven insubstantiability that barely exists outside your own mind?

    From memory I seem to recall that by “regularly” you mean “at least once a week.”

    Do you really see, say, an incremental improvement of 1% every week, or are you just fiddling around in the vain pretence that anything you do will make the slightest difference?

    I’m seeing the latter of those two propositions, myself. Because the former would suggest that the Linux desktop is improving at an exponential rate that would make Gordon Moore and his obsolete law gasp in amazement.

  81. DrLoser says:

    Wasting time trying to score irrelevant points again, are we, Fifi?

    So glibc is not a Linux only bug.

    Did I say it was? However, should you choose to run a Linux distro of practically any kind (is there still one out there that relies on a fork of glibc? There might be one. I don’t know), then you are choosing to run glibc.

    Well, it’s sort of forced on you, really, but as we all know, every Gnu/Linux user religiously scrutinises every last line of code, as prescribed by the Four Freedoms, so I think “choosing” is an acceptable word here.

    On the other hand, if you own an Android device, you are not (ceteris paribus) “choosing” to run glibc. Why not? Because Kindly Uncle Google has “chosen” not to run it on your behalf.

    Those getting applications from third party Android app stores will be at risk of glibc issues.

    And those choosing to pave over Android and replace it with a Linux distro will also be at risk of glibc issues. This should hardly come as a surprise. Those choosing to swallow their mobile phone or tablet will similarly have made a potentially injurious choice.

    The fact remains, Fifi. That Google chooses not to rely on glibc — a central part of Gnu, I would think — is pretty solid evidence that ANGLE.*

    (“Android is Not Gnu/Linux, Evidently.” Not quite recursive, but still a valid acronym. And in fact the only point I was trying to make, Fifi.)

  82. DeafSpy says:

    Hey there, little incompetent and illiterate lying fraud!

    Due to glibc not being centrally managed on OS X and Windows those are at a higher risk.

    To start with, glibc is no part of Windows whatsoever. Therefore, Fifi, you speak total crap, just as usual.

    However, since I am in a good mood, I will indulge you that you mean some runtime like, hm, let’s say, some runtime like Visual C++ runtime. Then you are wrong again, because it is simply not a part of the OS. Next, you can mean a runtime like .NET, which can be considered part of the OS (though it is not). But, .NET is subjected to updates via Windows Update.

    Fail, Fifi. Fail as usual.

    Go back and learn about writeln(), you fraud!

  83. oiaohm says:

    My point was that it could not possibly have affected Android.

    Because Google/Android does not use the notably unreliable glibc.

    DrLoser to be correct Android is more likely to be effected by current glibc issue long term than a normal Linux Distribution. Because Android applications can have a embedded glibc. So can OS X and Windows. So glibc is not a Linux only bug.

    Yes glibc is not default Android, OS X and Windows but being free of glibc issues is not true. Due to glibc not being centrally managed on OS X and Windows those are at a higher risk. Android app store can scan for applications containing alternative libc solutions. Those getting applications from third party Android app stores will be at risk of glibc issues.

    The core parts need to be very heavily audited because everyone can be at risk when they have faults.

  84. Wizard Emeritus says:

    “I regularly build my kernel and a few applications where I want the latest improvements,”

    Using which option Robert Pogson?

  85. DrLoser, barking up the wrong tree, wrote, “you haven’t, have you?”

    I regularly build my kernel and a few applications where I want the latest improvements,

  86. DrLoser says:

    Never let it be told of me that I cannot explain the obvious to dimwits. Fortunately, this site has an overabundance of dimwits.

    So, allow me to specify how a Gnu/Linux/Small Smart Thingie should accommodate an application based on FLOSS principles. Here I go, and I welcome any corrections, no matter how fatuous … I will be rating them on a scale of 1-10 …

    1) Build the app, put it in the Store, and so on. This one loses on FLOSS.
    2) Use the equivalent of apt-get, preferably with the “-devel” option. Ideal for people like Robert, who cannot program in C and quite frankly have no idea whatsoever how to navigate the bizarre and outdated GNU build chain*.
    3) Download the raw code and configure/make/make install the thing.

    You can’t do a single thing there, can you? Hypocrites!

    * I don’t blame you, Pog. Every time I look at this monstrous 1980s hack-job, I feel precisely the same way. The only difference is, I have tried to navigate it as per the principles of GNU/FLOSS.

    And you haven’t, have you?

  87. DrLoser says:

    Oh, I almost forgot. To quote myself:

    Feel free to proclaim that Google (an entirely closed ecosystem) is preferable to Microsoft (an entirely closed ecosystem).

    And both of these, practically equivalent, entirely closed ecosystems have walled gardens between the consumer and the available applications.

    Isn’t Corporate Convergence wonderful?

  88. DrLoser says:

    I can accept the occasional glitch a few times a year over multiple glitches every month like TOOS.

    Care to quantify this acceptance criterion, Robert?

    Your choice of pain points.

    Your choice of what, for the sake of this discussion, I shall refer to as “Mean Time Between Glitches?”

    Numbers are good. Performance targets are good.

    I’m very glad we agree on this fundamental principle, Robert. Now: let’s hear your basic requirements for pain points and MTBG*.

    * Mean Time Between Glitches.

  89. DrLoser says:

    Quite on the contrary, I said it was trivial (because it is).

    No, you did not, Luvr.

    You don’t read or analyse my posts, but I read and analyse yours.

    You never once said that dual-booting Linux on an extant Windows system was “trivial.”

    Quite the opposite, in fact.

    Go ahead, give us the cite.

  90. DrLoser says:

    It doesn’t affect GNU/Linux either.

    My point was not whether it affected your GNU/Linux distro, Robert.

    My point was that it could not possibly have affected Android.

    Because Google/Android does not use the notably unreliable glibc.

    Ergo, it is about time you stopped this fanboi nonsense about Android being “GNU/Linux.” Because it isn’t.

    At the very best, Google pick and choose between the bits of GNU and Linux they need to cobble up the underpinnings of the Android VM. Apparently glibc is not one of those pieces.

    Oh, and did I mention that you can’t FLOSS/download the actual OS?

    You can’t, you know.

    Feel free to proclaim that Google (an entirely closed ecosystem) is preferable to Microsoft (an entirely closed ecosystem). But … Get real. Do not do it on the basis of Freedom.

    Surely Microsoft Hate should be enough to satisfy you?

  91. DrLoser wrote, “mostly, it doesn’t affect Android because Android does not use glibc.”

    It doesn’t affect GNU/Linux either. We patched everything the day before. So much for the GNU/Linux bashing. I can accept the occasional glitch a few times a year over multiple glitches every month like TOOS.

  92. luvr says:

    DrLoser wrote, “I love how you bleated about how impossible it is to dual-boot Linux on an extant Windows system”

    I didn’t bleat, and if I did, then it wasn’t about how supposedly impossible it is to dual-boot Linux on an extant Windows system. Quite on the contrary, I said it was trivial (because it is).

    I can repeat this as often as necessary. I don’t suppose it is necessary for you to accept that I have said it. There’s no reason for you to scroll down and check, is there?

  93. DrLoser says:

    Good news on the latest glibc security hole by the way — it doesn’t affect Android. Hoorah! The future for Small Smart Thingies from Google is assured!

    Of course, mostly, it doesn’t affect Android because Android does not use glibc.

    So much for this Gnu/Linux/Android nonsense.

  94. DrLoser says:

    So, DrLoser, to sum up what you said: Dual-booting Windows and Linux is trivial, but it’s too hard for Microsoft.

    No, Luvr, to repeat my earlier summary of what I said: Microsoft sees no benefit in provisioning such a possibility, and yet several possible issues (non-technical) in so provisioning.

    I can repeat this as often as necessary. I don’t suppose it is necessary for you to accept that I have said it. There’s no reason for you to scroll down and check, is there?

    I love how you bleated about how impossible it is to dual-boot Linux on an extant Windows system — it’s all Microsoft’s fault! — right up to the point where I explained, step-by-step, how to do it. At which point you brought up UEFI for some reason — it’s all Microsoft’s fault! … and, who knows, perhaps oiaohm is right and there are current problems with UEFI. I wouldn’t know. It wasn’t the point at issue.

    It’s never, ever, going to be your fault for slip-shod technical ignorance, is it? It’s always Microsoft’s fault!

    What a lovely cozy little padded cell you must live in.

  95. oiaohm says:

    DrLoser even with EFI you have issues of Windows clearing settings owning to other OS systems.
    http://askubuntu.com/questions/235567/windows-8-removes-grub-as-default-boot-manager
    From own experience this is true even up to Windows 8.1 I have not tried windows 10 maybe Microsoft has at long last stopped changing default loader to Windows(highly unlikely).

    Those upgrading from windows 7 and 8.1 to 10 with EFI systems
    http://askubuntu.com/questions/655011/windows-10-upgrade-kills-grub-and-boot-repair-doesnt-help
    Have also suffered I magically reset the default OS EFI var to point to Windows. So I can fairly much bet Windows 10 first boot loader update will be status Microsoft normal we don’t class any other OS able to have default so we will reset the EFI var for default OS to ours.

    Of course it gets more fun EFI. Grub and the other open source EFI loaders in fact support chainloading other EFI loaders. At least with MBR windows you could add grub to the NTLDR as a chainloading option and have it always appear one way or the other. Guess with EFI windows does not support chainloading yet it still wants to take over default. So in a lot of ways this is worse than dealing with NTLDR over writing MBR.

    DrLoser the VBR is in fact written twice by Windows prior to UEFI. Into MBR and the start of the boot partition.

    Over all not checking if some other OS is on the system before doing a disruptive change has not changed at all with the introduction of EFI with Microsoft.

    The single unit stuff is kinda bull. Its also funny when you install a system with Windows 2000, XP and 7 side by side. Microsoft is smart enough to detect hey the MBR is newer so don’t over write that so the older OS have to chainload. So its not that Microsoft cannot do multi OS. Microsoft only does multi OS if the OS’s you are using are Microsoft. So if it was a single unit doing what I describe should have still caused failure.

    Like it or not DrLoser Microsoft is the guilty party making life hard.

    Also there have been cases of EFI implementations hard coded only to load the Microsoft EFI loader. So going UEFI can be impossible on some Motherboards for alternative OS solutions.

    Sorry DrLoser give boot loader advice when you have some experience dealing with it.

    The day Microsoft can do a loader update or OS install without disrupting other OS’s they they deserve pat on back until then they deserve all the brick bats they get over it.

  96. luvr says:

    So, DrLoser, to sum up what you said: Dual-booting Windows and Linux is trivial, but it’s too hard for Microsoft.

  97. DrLoser says:

    And just to clear up any possible confusion, yes, when I speak of NTLDR, I am assuming a combination of Microsoft’s handling of the bootstrapping code in a VBR and NTLDR. From Microsoft’s perspective they come as a single unit. The reasoning I applied to NTLDR “on its own” was a simplification, because the actual mechanics between any Master Boot Record and the resultant boot loader are irrelevant to the argument.

    (And when I speak of NTLDR I am referring to the historical artefact as of all NT systems up to XP, and am therefore presently referring to the equivalent mechanisms in Vista and later. Once again, the actual mechanics are irrelevant to the argument.)

    Anyway, the original whine was about the near-impossibility of dual-booting a Linux system with a pre-existing Windows system, as I recall. The near-impossibility of which I believe I have, basically, proven to be complete and utter bullshit.

  98. DrLoser says:

    Bullshit. Booting Windows in a multiboot environment governed by a GRUB (or LILO, or whatever) Master Boot Loader is trivial. You make it sound like NTLDR can no longer be used if GRUB sits on the MBR, which is nonsense (as you are well aware) … [and so on and so on]

    I did no such thing, Luvr. I merely observed that, from Microsoft’s perspective, abandoning NTLDR for Grub has precisely no benefit and yet offers several inherent disadvantages.

    Obviously one can multiboot through Grub. I have done so. Refer to my earlier post. Goodness me, with attention to detail this cursory, it’s no wonder you’re so bloody useless at establishing a dual-boot system on an extant Windows platform.

    And even if I did, MBR would be irrelevant, and I would have to deal with the far greater mess that is that UEFI crap.

    One more thing for which you are proud to own up to total technical ignorance. Gee, whining is so much fun, isn’t it? Far more fun than actually accomplishing anything.

    It’s a triffically difficult and convoluted subject, but I can help. Let me know if you need any more hand-holding beyond a standard Debian install, won’t you?

  99. Wizard Emeritus wrote, “In spite of your views, the legality of software patents is a fact elsewhere.”

    In fact, very few countries have been lead down the rabbit-hole of software patents:
    ” United States of America
    Australia
    Canada
    Japan”

    See Software Patents

    Copyright is the proper way to protect rights in software.

  100. Wizard Emeritus says:

    “In any case, the issue is moot anyway these days, since I no longer care about Windows. And even if I did, MBR would be irrelevant, and I would have to deal with the far greater mess that is that UEFI crap.”

    For someone who doesn’t care, you sure have a lot to say, Mr. L.

    As far as UEFI is concerned, The only mess is caused by your own desire to use an also-ran operating system as your desktop. Were your to use a commercial Distribution like Red Hat Enterprise Linux to run server based applications, as I did until recently, you would find UEFI worked just fine.

  101. Wizard Emeritus says:

    “One of the ways that M$ has been blocking OEMs from providing better operating systems has been the abuse of patents, threatening to sue unless certain actions are taken. ”

    So it is abuse to insist that others pay up for your innovations, eh Mr. Fish. In spite of your views, the legality of software patents is a fact elsewhere.

    The fact that the Russian kleptocracy doesn’t honor software patents is irrelevant. OEM’s are not going to provide “better” desktop Operating systems unless their customers demand it, and the customers in Russia like elsewhere want to run desktop apps on top of windows.

    As a result it is probably more likely that Russians who need to run applications on windows will make due with what they have or, unfortunately pirate what they need.

  102. luvr says:

    DrLoser wrote, “The immediate implication of this is that Microsoft would not be able to support dual-booting Linux even if it wanted to, at least not without considerable work.”

    Bullshit. Booting Windows in a multiboot environment governed by a GRUB (or LILO, or whatever) Master Boot Loader is trivial. You make it sound like NTLDR can no longer be used if GRUB sits on the MBR, which is nonsense (as you are well aware): Computer boots, GRUB early boot code is loaded from the MBR, GRUB early boot code runs, GRUB early boot code loads GRUB core, GRUB core runs. User may want to boot Windows (if that is one of the Operating Systems installed on the computer). If so, then boot sector from Windows system partition gets loaded and run. Windows will then boot in whatever fashion it always does, which will involve NTLDR as always.

    Microsoft has to do nothing, nada, to make this possible. All that would be required from them, if they were interested in playing nice for a change, would be to provide an option to leave an existing functional MBR intact. They wouldn’t even have to go through the trouble of reconfiguring the Master Boot Loader for me; I will do that myself, thank you.

    Not that it’s a disaster when Windows overwrites the MBR; I do know how to repair it, and it’s not a huge issue, just an annoyance. Even so, I still hate it.

    In any case, the issue is moot anyway these days, since I no longer care about Windows. And even if I did, MBR would be irrelevant, and I would have to deal with the far greater mess that is that UEFI crap.

    I cannot believe that a “self-admitted useless incompetent” like me would have to explain any of this to an expert like you. I’m sure that I didn’t tell you anything that you didn’t already know.

  103. Modular sunfish says:

    One of the ways that M$ has been blocking OEMs from providing better operating systems has been the abuse of patents, threatening to sue unless certain actions are taken. Although software patents are not valid in Russia, see 1350.5.5. So that pressure is off them from that, at least in principle.

  104. kurkosdr says:

    @Pog

    So you do monitor this thread. Then you wouldn’t mind clarifying your position about the original topic, right?

    Allow me to resubmit some questions, with a misguided expectation you will not ignore them

    1) How do you define a legitimate (non-illegitimate) regime?

    2) Is the Saudi Arabia regime legitimate?

    3) If the Saudi Arabia regime starts dropping bombs on rebels trying to overthrow them inside their territory (if that happens sometime), will they still be a legitimate regime?

    4) If the Saudi Arabia regime starts dropping barrel bombs on rebels trying to overthrow them inside their territory (if that happens sometime), will they still be a legitimate regime?

    You have to answer all those questions, otherwise the default definition for your “legitimate regime” term will be: “a regime that is cozy with the US, not with Russia”

  105. DrLoser wrote, “Corporate customers want dual booting. I’m sure they’ve demanded it of M$.”

    I can’t imagine any corporation wanting dual-booting widely. Corporations may waste resources but there’s no way to use both OS simultaneously so one or the other is a waste. If they needed to use both, they would use virtual machines or remote access, not dual-booting. I once installed a whole lab with dual-booting. TOOS was never used again by teachers or students. It didn’t meet their needs as well as GNU/Linux. TOOS was forever giving them trouble. GNU/Linux gave neither me nor them any trouble in that installation.

  106. DrLoser says:

    Microsoft is honest: We assume Windows will be the OS of the harddrive.

    The reality is slightly different, as I pointed out. Microsoft uses NTLDR. Not Grub.

    The immediate implication of this is that Microsoft would not be able to support dual-booting Linux even if it wanted to, at least not without considerable work. So let’s examine the possible reactions to this obvious fact (which somehow escapes our renowned expert in such things, Luvr).

    Reaction #1: Typical M$ — proprietary lock-in. Why couldn’t they just use Grub like everybody else?
    a) I am kindly going to assume that Grub is not crap, even though I know from personal experience that it is, very much, crap.
    b) Why should Microsoft “lock themselves in” to a moving target? Lilo, Grub, Grub2, who knew what was coming next? And if “BootyGnulicious” cocks up and eats your “M$ distro,” well then, who is the customer going to blame? “BootyGnulicious?” I don’t think so. Most people have barely heard of Linux, and 99% of those who have are as ignorant as Luvr about the workings of the boot-loader.
    c) Conversely, wouldn’t it be easier for the Million Eyes to reverse-engineer NTLDR? It’s another case of the Linux zealots wanting the 80% (or 60% or take your pick) to convert wholesale to their way of doing things, isn’t it?

    I have news for you, folks. As Joel Spolsky pointed out, it doesn’t matter what business you are in. IT, bill presentment, whatever. That is never going to happen. Consumers do not think that way.

    Reaction #2: Well, of course, there’s a technical issue here. But it’s easy to fix. Corporate customers want dual booting. I’m sure they’ve demanded it of M$. M$ is unresponsive and illegal and furthermore a corporate bully.

    a) You are a paranoid idiot.
    b) There is no point (b).

    Reaction #3: Well, we’re technically competent folk. We can make the best of a bad deal. Dr Loser has shown us how to do so, in a very small number of extremely straightforward steps. Thank you, Dr Loser! Armed with this useful set of instructions, I shall go forth and set up dual-boot Linux systems!

    a) You’re welcome. It was my pleasure.
    b) But you won’t, will you? You’d rather squeal like a stuck pig than deal with reality, wouldn’t you?
    c) You’d be hopeless at following instructions in the first place.

    I think that about covers it. Don’t you?

  107. Wizard Emeritus says:

    “The e-mails in the US DOJ v M$ exhibits show M$ regularly got insider information from competitors. What do you think motivated those insiders to divulge? 😉”

    I think that you assumptions are not proof of anything. And the fact that nothing came of what was in the correspondence confirms my assessment.

  108. Deaf Spy says:

    Off-topic: Is it just me, or Firefox is the new IE7?

    Not really, because Firefox is dying away. The new IE8 however, will be Safari.

  109. DrLoser wrote, “It’s relatively straightforward, though it takes a bit of time. First, create your RAMdisk. Then, load GNU-like tools to that RAMdisk.”

    It was even easier on the old IBM 1620. One typed a line of numerals on the keyboard and stuff happened. The machine slurped up stuff from whatever device and started at square one.

  110. Wizard Emeritus wrote, “there is still no indication that microsoft used bribery as a policy.”

    The e-mails in the US DOJ v M$ exhibits show M$ regularly got insider information from competitors. What do you think motivated those insiders to divulge? 😉

  111. kurkosdr says:

    BTW, since this thread is about to stall in a while, allow me to resubmit some questions to Pog, with a misguided expectation he will not ignore them

    1) How do you define a legitimate (non-illegitimate) regime?

    2) Is the Saudi Arabia regime legitimate?

    3) If the Saudi Arabia regime starts dropping bombs on rebels trying to overthrow them inside their territory (if that happens sometime), will they still be a legitimate regime?

    4) If the Saudi Arabia regime starts dropping barrel bombs on rebels trying to overthrow them inside their territory (if that happens sometime), will they still be a legitimate regime?

    You have to answer all those questions, otherwise the default definition for your “legitimate regime” term will be: “a regime that is cozy with the US, not with Russia”

  112. Wizard Emeritus says:

    “Huh? You really expect a “self-admitted useless incompetent” like me to know all this?!?!? 😉 😉 😉”

    Not really. After all the owner of this blog doesn’t either, so it wouldn’t be fair to expect his posters to be any better.

  113. luvr says:

    DrLoser wrote, “But, of course, you knew that all along. Didn’t you, genius?”

    Huh? You really expect a “self-admitted useless incompetent” like me to know all this?!?!? 😉 😉 😉

  114. kurkosdr says:

    Desktop Linux sucks even in dual-boot scenarios = Desktop Linux sucks even in non dual-boot scenarios

  115. kurkosdr says:

    So one bug in one distro means GNU/Linux sucks? By the same logic TOOS blows chunks.

    Lol are you serious? Ubuntu is the most friendly distro to proprietary drivers out there. Others don’t even bother including the firmware, as an effort to punish the user for not hand-picking Libre hardware or whatever. Or under the mistaken belief that WiFi vendors will just open the sources which contain proprietary algorithms they paid lots of money to really smart people to come up with. Or that all vendors will adopt the Atheros silliness of running their drivers in a mini-cpu on the card.

    And it’s not one bug. The Driver Manager thingie failed because the kerner module loader wasn’t installed and couldn’t load the module. That’s not a bug. That’s a “I don’t care”. Most Distros don’t have a tool to load proprietary drivers anyway. It’s Terminal-fu all the way.

    But, before I do that, Luvr, let us go into some detail about “knowing what you are doing.”
    Which is my main gripe. The marketing of Desktop Linux distros focuses around how easy it is to dual boot with Windows, without providing any real tools to do manage such a dual-boot system. Microsoft is honest: We assume Windows will be the OS of the harddrive.

    Anyway, Desktop Linux sucks even in dual-boot scenarios (see my post below).

    —–

    Off-topic: Is it just me, or Firefox is the new IE7? I was browsing the internet with my Aspire One using Chrome, and -to my deepest disbelief- the internet felt faster on the lilliputian Atom N270 than it did to the Core i5 my main laptop has -which had only Firefox installed. The Core i5 is 13x faster to the Atom according to cpubenchmark.net I mean… wow! I knew Firefox was slower, but I didn’t expect things to have degraded that much. Also, Chrome doesn’t block scrolling when some script runs on some other tab, or even in the same tab. Probably has something to do with the fact Google has a financial interest to make sure pages run fast, while Firefox is like a government employee, receiving a fixed paycheck from Google regardless. Anyway, when the last cling-overs like me leave the place, the paycheck will stop, and Mozilla won’t be able to misuse the money to chase Firefox OS, and now IoT, fairies. I will only use Firefox when I need VideoDownload Helper from now on…

  116. DrLoser says:

    Well now. Having entirely demolished Luvr’s claims to technical competence, perhaps we can move on to more important matters?

    I am quite content to be called a “troll,” because that term is basically value-free. I do, however, object to being called a “shill,” because that term suggests some form of payment. Possibly even an illicit payment.

    Would any of you fine upstanding gentlemen out there care to speculate on the cumulative remuneration that I, Wizard Emeritus, and Deaf Spy receive from Microsoft or (perhaps) from their Secretive Goons?

    Feel free to speculate. I wouldn’t worry about Kurkos — he’s been interviewed, but he turned out to be completely worthless.

  117. Wizard Emeritus says:

    “Indeed. I did consider offering the obvious solution, Wiz, but I felt I should go through the gnarly alternatives first.”

    Me Too. I could also have talked about the microsoft boot manager with its boot to VHD capability. But that would have added unnecessary complication as well.

  118. DrLoser says:

    Oh, and Luvr?

    The Wiz solution works both ways round. There’s nothing at all to stop you from installing Windows on a VM under Linux.

    But, of course, you knew that all along.

    Didn’t you, genius?

  119. Wizard Emeritus says:

    “M$ has lots of policies it appears to feel free to disregard, as well as laws. ”

    I am glad that you have proof of that, perhaps yous should provide it to the entities that are investigating the bribery charges, because so far, the worst case seems to be a rogue manager in microsoft Rumania. But there is still no indication that microsoft used bribery as a policy.

    “As the market gets more open and less corrupt, M$ is finding itself at more and more unable to compete. The net will get quite interesting when M$ no longer can finance its shills.”

    Again, Mr. Fish, you do not fail to amuse. Presumably you have noticed all those pesky commercial desktop applications that people use on their computers. Those applications are not going to go away just because someone offers a competition a so called zero cost equivalent. And unfortunately for you and all like you, most if not all of those applications require windows to run. So I think it safe to say that Microsoft is not going anywhere for a while.

    Of course the whole new market for mobile computing has opened up, and microsoft is like everyone else scrambling for a piece of the pie. But that market is an adjunct to the desktop market, not a competing one.

  120. DrLoser says:

    Actually, my dear Doctor, I solved that particular problem long before by banishing whatever Linux distribution to a VMWare virtual machine under windows.

    Indeed. I did consider offering the obvious solution, Wiz, but I felt I should go through the gnarly alternatives first.

    If only to prove that Luvr is a self-admitted useless incompetent.

  121. DrLoser says:

    And I am still the only man here (with the possible exception of RAM) who has actually written a boot-loader (from a set of CDs, for what that is worth) for a *nix derivative — ie, QNX.
    It’s relatively straightforward, though it takes a bit of time. First, create your RAMdisk. Then, load GNU-like tools to that RAMdisk. (You will be pleased to know that QNX likes, or at least in 2000 used to like, the “pure” and “bloat-free” ksh.) Then, munge partitions.

    Now, in the second phase, you install the kernel. And move things around. A lot. From memory, this is the phase when you mess with the MBR. Also from memory, this is when you reboot.

    Now, here’s the clever bit. In the third phase, you’re not actually going to use the kernel. What you actually do is to spin the RAMdisk up again, use ksh again, and drag everything useful off the other six or seven CDs. Then you commit the kernel to the MBR. Then you reboot.

    It was a bundle of fun.

    And it took me six to eight weeks to get it straight. I suspect I know more about this stuff than the pathetic likes of Luvr.

  122. Wizard Emeritus says:

    “Oh, and btw, as late as 2005 I was still having problems with my dual-boot Linux system because Grub moved the other OS boot roughly two bytes sideways.”

    Actually, my dear Doctor, I solved that particular problem long before by banishing whatever Linux distribution to a VMWare virtual machine under windows.

    And I haven’t looked back since.

  123. Wizard Emeritus says:

    “So, here’s a suggestion: I let you use Windows, and you let me use Linux, and we’re both happy. Problem solved.”

    But Mr. L. what you use has that has never been an issue here. IN fact I do not care what you choose to use. I am responding to Robert Pogson’s blog posts, calling him out for what I feel for posts that are misleading, ignorant, and many times opinion masquerading as truth.

    I you insist on commenting on my comments, then you will be responded to in kind.

  124. DrLoser says:

    Oh, and btw, as late as 2005 I was still having problems with my dual-boot Linux system because Grub moved the other OS boot roughly two bytes sideways.

    The distro in question was Mandrake or Mandriva or whatever the French were calling it at the time.

    I don’t imagine things have gotten any better since.

  125. DrLoser says:

    That’s an incredibly stupid remark. I used to keep running into trouble whenever I had to reinstall Windows on a dual-boot system, because Windows would forcibly overwrite the MBR, with no way whatsoever to prevent that. I HATED that. I wouldn’t have minded if overwriting the MBR had been the default option, but a real operating system should at least offer the option to leave the MBR intact–as any serious Linux distribution does (assuming, obviously, that you know what you are doing).

    Yes, assuming you know that you are doing. A big assumption that.

    As a sort of cliff-hanger, I will explain the pain-free obvious solution at the end of this post. But, before I do that, Luvr, let us go into some detail about “knowing what you are doing.”

    It’s true that all Windows installation setups (as far as I am aware) make the assumption that the user wants them as the primary operating system. I say “primary” because some of them play nicely with “secondary” operating systems … that secondary operating system being a variety of Windows. This shouldn’t come as a surprise.

    Now, they can do this quite easily because, not to put too fine a point on it, they don’t use Grub. Nor Lilo. For the last 15 years or so, they have used NTLDR. Think about that. Think about the requirements for NTLDR to interoperate with Grub or Lilo. Don’t think too hard — too many brain cells rubbed together and your brain might catch fire.

    I might very well have the record here for the number of bootable OSes on a single desktop: five. In no particular order:

    * Windows (2000)
    * Solaris 7
    * BeOS
    * OS/2 Warp
    * FreeBSD

    At some point, I can’t remember when, I tried to force a 2000-era Linux into that set. It didn’t play happy.

    However, the principle is fairly straightforward. You choose the awkward ones first (in this case it was OS/2 Warp, oddly enough) and you leave the more amenable ones until last (in this case it was FreeBSD). It takes a bit of trial and effort, but if I, a troll, could achieve that in 2001, I’m pretty sure the Resident Geniuses can manage it.

    The thing about Grub2, you see, and as a man who self-confessedly knows what he’s doing, Luvr, you will no doubt agree with me that the original Grub was an abortion … the thing about Grub2 is that it doesn’t actually fit into MBR memory. Consequently it has to play all those gruesome tricks (four layers of tricks, from memory, and I cannot be bothered to look it up) in order to force itself to fit into limited space.

    It’s far easier to ignore all that amateur trickery and install Windows first. (Or in my case OS/2 first.) But, if you have a prior Linux installation of several years, then here’s what you do:

    1) Take a working Linux desktop system. Make sure you have an extra, bootable, partition on the master disk. (We are all Linux aficionados here. I presume this step is trivial. Consult your friendly local fdisk manual.)
    2) Set up a partition for “other.”
    3) Set up another partition called “garbage” or whatever, I don’t care.
    2) Install Windows. Point the installation at the “other” partition. Yes, it’s just that easy. Now, here comes the tricky bit.
    3) Insert absolutely any Linux Distro DVD of choice. Or floppies. Works either way.
    4) Go through the Linux Distro installation choices (assuming it’s not Slackware, and it has a GUI installation menu). Install to “garbage.”
    4a) Make sure that the boot loader will recognise the Windows partition, and the partitions with your original Linux OS.
    5) Reboot.
    6) Reconfigure Grub so that it no longer recognises “garbage.”

    Looks to me like you have a perfectly serviceable dual-boot Linux/Windows system there, matey. Of course, you can save disk-space by deleting the garbage. And indeed the original OS, but I sense you would go all prickly on me if I were to suggest this seriously.

    Oh, the pain-free, obvious solution? Install Windows first and Linux second.

    What, what, you’d lose all that precious Linux Desktop user information if you did that?

    Back it up, my man, back it up first.

    You … do … back up your /usr and /home and possibly even /etc directories, don’t you, luvr?

  126. kurkosdr wrote, “It still doesn’t install drivers it has in it’s own pool (which are loaded in Live mode)”.

    So one bug in one distro means GNU/Linux sucks? By the same logic TOOS blows chunks.

  127. luvr says:

    kurkosdr wrote, “You see, Windows 10 may be an annoying little twat, but at least it’s a working annoying little twat. Desktop Linux is just annoying.”

    Don’t worry, that’s OK. To me, however, it’s the other way around: Desktop Linux may be an annoying little twat, but at least it’s a working annoying little twat. Windows is just annoying. And that’s OK, too.

    So, here’s a suggestion: I let you use Windows, and you let me use Linux, and we’re both happy. Problem solved.

  128. kurkosdr says:

    That’s an incredibly stupid remark. I used to keep running into trouble whenever I had to reinstall Windows on a dual-boot system, because Windows would forcibly overwrite the MBR, with no way whatsoever to prevent that. I HATED that.

    On the other hand, Windows isn’t marketed as being a dual-boot-happy OS like most Desktop Linux distros…

    Anyway, ignore the stuff about the MBR.

    The main point of the post was: Desktop Linux still sucks. In fact, it sucks even if you don’t consider the dual boot problems. It still doesn’t install drivers it has in it’s own pool (which are loaded in Live mode), the Driver Manager thingie doesn’t work and you have to manually install kernel loaders and kernel modules, performance is crap (hello X.org and Mr. Poettering), power management is god-awful, there are conflicting settings between X.org, desktop environment and shell which are not even hidden from the GUI (with conflicting tools messing with each other), the updater fails to work in perfectly good connections, and then there is my personal favorite screw up: the fact updates don’t install automatically and that you need a sudo account (password) to do updates at all.

    The last screw up is particularly indicative of how much of a toy OS Desktop Linux still is: Let’s say we have Jane who is a manager of a small-medium business or a parent in the home, and has a sudo-capable account. We also have Mark who is an employee (in the SMB context) or a teen (in the home context) and hence has a non-priviledged account. If Jane doesn’t login to Mark’s computer too often (because she has her own), she either has to login once a week (or so) when Mark isn’t looking to install updates, or “roll her own scripts”.

    You see, Windows 10 may be an annoying little twat, but at least it’s a working annoying little twat. Desktop Linux is just annoying.

  129. luvr says:

    kurkosdr wrote, “Because GRUB replaces the MBR and offers no way to undo it. Boo.”

    That’s an incredibly stupid remark. I used to keep running into trouble whenever I had to reinstall Windows on a dual-boot system, because Windows would forcibly overwrite the MBR, with no way whatsoever to prevent that. I HATED that. I wouldn’t have minded if overwriting the MBR had been the default option, but a real operating system should at least offer the option to leave the MBR intact–as any serious Linux distribution does (assuming, obviously, that you know what you are doing).

  130. luvr says:

    kurkosdr wrote, “You can’t cd to ~/Desktop if your language isn’t english. You have to cd to the translation of Desktop”

    That’s the only real issue I agree with. I find it annoying that I must take into account the language settings that are in effect, just to address directories such as what I prefer to call simply “Desktop”. All it takes is, indeed, a few logical links to support both the default English and the localised directory names.

    The rest is just a matter of knowing what you are doing. There are annoyances, alright, but nothing to keep whining about. If I wanted to, then I could whine on and on and on about the neverending Windows annoyances, but I don’t bother. I’ll leave Windows to those who (think they) know how it “works”.

  131. Modular sunfish says:

    Wizard Emeritus wrote, “Also microsoft has been very clear about their policy on corruption”

    M$ has lots of policies it appears to feel free to disregard, as well as laws.

    About the bribery court cases, things take time and it’s only been a few years. The case against M$ in Romania, where M$ paid a €4 million bribe, has only just gotten moving. The courts can take a while yet before the Russian case is settled. Slowness is an old M$ tactic and especially true in situations where M$ is not going to willingly cooperate. Even back in the 1990’s Gates equated obeying the law with the destruction of M$. A view that he seems to keep even today and, yes, 30% of his time is still spent on M$.

    As the market gets more open and less corrupt, M$ is finding itself at more and more unable to compete. The net will get quite interesting when M$ no longer can finance its shills.

  132. kurkosdr says:

    USA = USB *sigh*

  133. kurkosdr says:

    MAR = MBR (autocorrect)

  134. kurkosdr says:

    Also, why does TOOS always need to be repaired?

    Because GRUB replaces the MAR and offers no way to undo it. Boo.

    That’s usually because the manufacturer puts a licence on the firmware which permits only the end user to install the stuff.

    As I said, driver was working in Live USA mode. If the OS could install it in Live mode, it should be able to do it at first boot after installation. No sir, this is about the 4 freedums.

    I did stuff like that in my first months of Linux using tar, fdisk and mke2fs…
    Whatever… What I got is that if I tried to move the start of the partition some blocks down (so the windows partition can grow), gparted fired ominous warnings about GRUB. Unless of course you mean truncating the Linux partition from it’s bottom, and creating a new NTFS partition after the end of the linux one. Fragmenting the drive more than it already is. Behold! Solution!

    So, you want BASH to translate? Yes, I can see how well that works
    In windows, I can address the Desktop (and Downloads, and Music etc) folder both with their English and translated names. All it takes is one symbolic link (for each folder). Now I hate the practice of translating folders (which was born from the PHD-enforced notion that “the localised version should have no English text”), but at least windows does it right. Ubuntu ‘ way is asking for trouble.

  135. kurkosdr wrote, “The broadcom wifi worked in Live CD mode (or more accurately Live USB) but not after installation. The installation disk included the proprietary driver, but is not installed.”

    That’s usually because the manufacturer puts a licence on the firmware which permits only the end user to install the stuff. Otherwise that firmware would be part of the Linux kernel. Lots of manufacturers make their firmware available to Linux.

    kurkosdr also wrote, “Really annoying pet-peeve: Nobody has made a tool to uninstall the damn thing without losing Windows. You have to manually nuke the partitions from windos and make a windows repair disk… *sigh*”

    Sigh. Ever heard of fdisk or dd or grub? Also, why does TOOS always need to be repaired? You should keep GNU/Linux and chuck that POS.

    kurkosdr wrote, “There is no way to shink the linux partition after installation without some serious GRUB-fu. I just nuked the linux partition and re-installed.”

    Gee! I must be a genius. I did stuff like that in my first months of Linux using tar, fdisk and mke2fs… That was LILO, though, but grub isn’t much different in a simple setup. Of course, I was working on ~800MB hard drives in those days. RTFM, I guess.

    Kurkosdr wrote, “You can’t cd to ~/Desktop if your language isn’t english.”

    So, you want BASH to translate? Yes, I can see how well that works. Pretty soon you have multiple names for multiple items or multiple items with the same name… There’s a reason why English is so widely used in IT. It’s to avoid such problems.

  136. kurkosdr says:

    they are currently downvoting those Dell XPSes = they are currently downvoting those Dell XPSes sold with Ubuntu

  137. kurkosdr says:

    Fate makes from strange coincidences… Yesterday, an old Acer Aspire One fell into my hands. Since I wanted a little laptop to take quick notes and run basic MATLAB scripts, I accepted it.

    Now, as you probably know, dual-booting linux on your netbook is like taking a shot of absinthe after your friends take you one. You know it’s pointless, you know it’s going to cause a headache, but you must do it anyway because of some unwritten law or something.

    Ubuntu was my poison of choice (because whatever little proprietary software support there is for desktop linux out there, it’s for Ubuntu). So, bottoms up… today I tested desktop linux on netbooks.

    All in all, I can confidently say Ubuntu (and desktop linux in general) is just as crappy as I remember it from a while ago. Here are the highlights:

    1) The broadcom wifi worked in Live CD mode (or more accurately Live USB) but not after installation. The installation disk included the proprietary driver, but is not installed. You have to follow these intuitive instructions: http://askubuntu.com/a/553619 I guess getting taint by proprietary firmware is OK only if done in Live mode. Anyway, thank Zeus I didn’t go for some extra-freedomite distro like Debian and didn’t got trapped in the funny situation of needing wifi to download wifi firmware.

    1b) Personal favorite: You can’t cd to ~/Desktop if your language isn’t english. You have to cd to the translation of Desktop: cd ~/”Επιφάνεια εργασίας” in my case. Attention to detail is paramount in linuxland I see…

    1c) In the Driver Manager thingie (which of course didn’t work, I hate it when gui tools fail error-less), some FOSSie translated “proprietary” as “private-ruled” instead of “being property of someone”. Unless it’s an Stallma-nite ideological term I don’t get.

    2) Going to the Input Method app (the first app that pops out when you search for “keyboard”) to changeyour language input settings will cause two warnings to appear plus one yes/no dialog (recommended option is “no”, so I clicked it) and will cause Unity to poop itself in the next boot! You get booted to a blank wallpaper. I just reinstalled the whole Ubuntu and used Canonical’s tool. Whatever. I guess nobody told the FOSSies that when your desktop is consisted of 2 moving parts (X.org, Gnome and whatever shell your distro uses) with conflicting settings, things may get hairy.

    3) App performance (such as firefox) is crap. YouTube doesn’t play in 360p (it does on windows 7)

    4) Power management is really crap. It always amuses me how FOSSies will whine and whine about RAM usage and boot times when a really scarce resource like battery mAh’s is litterally being wasted in most linux distros because of crap cpu settings – and X.org).

    5) Even though I have a fast internet, updates happen at glacial speeds of 1-2KB/s on my connection. Old bugs die hard, I guess: http://linuxhaters.blogspot.co.uk/2009/11/bad-karma.html

    6) Personal favorite: You need a sudo password to do security updates(uh-oh). I guess if you want low-priviledge users to use the system, you have to “roll your own scripts”. Or pretend to not care, because security is easy when you are ultra-obscure. Even for priviledged users, updates don’t happen automatically (uh-oh again). They do in Windows. LinuxIsInherentlySecure(tm)

    7) Personal pet-peeve: Most Desktop Linux distros (Ubuntu included) don’t ask which OS you want default in GRUB, or if you want it to remember last OS, they just assume they are the best thing since sliced bread and make themselves default. Some sudo gedit of the grub file and some update-grub is needed to fix that.

    8) Personal pet-peeve: There is no way to shink the linux partition after installation without some serious GRUB-fu. I just nuked the linux partition and re-installed.

    9) Really annoying pet-peeve: Nobody has made a tool to uninstall the damn thing without losing Windows. You have to manually nuke the partitions from windos and make a windows repair disk… *sigh*

    All in all, there is a reason netbooks originally started out as desktop linux devices but later became Windows devices. Hint to Pog: Users downvoted linux with their dollars and euros.

    Just like they are currently downvoting those Dell XPSes. But hey, retailers should devote self-space to linux so that some FOSSies idea about “justice” is fulfilled.

    Homework: Find out how many of the highlights apply to all desktop linux distros, so Pog can’t retort UseDistroX(tm).

    PS: If you know of some distro with good power-management, not the usual “it’s so light it should not draw too much power” nonsense, post.

  138. Wizard Emeritus says:

    “The original GNU/Linux netbook did decline but ChromeBooks both x86 and ARMed are doing pretty well, even in USA.”

    Yep the crapbook seems to have been picked up by some K-12 schools who see it as a panacea. That will only work so long as google can keep up with their needs or until it is realized that google is not their friend and that they are as locked-in as bad as any other solution. But I also think that before that happens happens Google will find a way to “merge” their so called ChromeOS into Android, allowing the Chromebook market to just melt into the Android market.

    All of this of course has very little to do with the Russians latest load of malarkey about Linux. The corrupt kleptocracy that is russia is trying to deal with the fact that they cant get software technology that they can use by attempting to force all of their people to use FOSS on linux.

    Were it not for the people that this is going to hurt the collision with reality
    that results with this latest imposition of FOSS and linux on a captive audience could almost be fun to watch.

    I

  139. Wizard Emeritus wrote, “the netbook market was dealt a blow by the success of the iPad, and killed outright by tablets and smartphones with the commercial OS named Android.”

    Not exactly correct. The original GNU/Linux netbook did decline but ChromeBooks both x86 and ARMed are doing pretty well, even in USA. The netbook in some ways was superior to the tablet, having a proper keyboard. There is a place for all kinds of hardware/software in IT but M$ should never be the sole judge of merit.

  140. Wizard Emeritus wrote, “My name changed because I retired, quite comfortably I might add.”

    Welcome to the club. I’m pretty comfortable too although I imagine a different number of zeros are required to support your lifestyle where you live…

  141. Wizard Emeritus says:

    “That may be a bigger barrier than the others.”

    It may indeed, however there does not seem to be any updates beyond 2 years ago. None of the investigations have so far shown anything beyond misbehaving representatives.

    Also microsoft has been very clear about their policy on corruption

    https://www.microsoft.com/en-us/legal/compliance/anticorruption/reppolicy.aspx

    outlines their policy. Feel free to disregard this, but it stands nonetheless. And of course you can if you feel that you have real evidence, call in to report it.

    Or do you prefer to just repeat your opinion as fact, Mr. Fish.

  142. Wizard Emeritus says:

    “Not going to bother. Its obvious you are using yet another name yet again. Different name same old shit.”

    My name changed because I retired, quite comfortably I might add. You on the other hand are still using the same name and still grumbling the same old angry geek shit. And probably working the same low level tech jobs.

    But Whatever…

  143. Modular sunfish says:

    Bribes and kickbacks also look to have held back the previous upgrade to GNU/Linux there,

    http://www.computerworld.com/article/2484041/technology-law-regulation/feds-expand-inquiry-into-microsoft-bribery–kickback-allegations.html

    That may be a bigger barrier than the others.

  144. lpbbear says:

    Not going to bother. Its obvious you are using yet another name yet again. Different name same old shit.

  145. Wizard Emeritus says:

    A Mr. Bear you never fail to entertain. Your evidence consists if a 7 year old article from the time period by one of your fellow “advocates”

    “Craig is a freelance UK web consultant who built his first page for IE2.0 in 1995. Since that time he’s been advocating standards, accessibility, and best-practice HTML5 techniques. He’s written more than 1,000 articles for SitePoint and you can find him @craigbuckler”

    Of course one should not be naive, ASUS did manage to generate a market whioch microsoft responded to by adapting windows XP to the netbook form factor, the resulting system was undeniably and unfortunately for you, so more usable that the Xandros version died.

    AT any rate the netbook market was dealt a blow by the success of the iPad, and killed outright by tablets and smartphones with the commercial OS named Android.

    (Now of course cue more insults and hit me with the assertion that Android OS is linux…)

  146. lpbbear says:

    Nice try at rewriting history numbnuts…….

    http://www.sitepoint.com/microsoft-killed-linux-netbook/

  147. Wizard Emeritus says:

    “M$ has never succeeded when the market has been open to competition. ”

    Please do give us examples in the desktop PC/Portable PC market where this was so. We await with baited breath.

  148. Wizard Emeritus says:

    “They began arm twisting manufacturers to use WindowsXP on the systems.”

    Was it arm-twisting Mr. Bear, or was it the fact that those OEM, when faced with a slow motion crash in the sales of the Xandros based netbooks, panicked and went to microsoft for a solution, any solution, that would save their asses.

    And Yes Vista was too big for netbooks, but Windows XP apparently could be “gotten down” in size to fit. Microsoft did so e viola, a netbook that could actually run windows binaries. The rest was history.

    But feel free to believe in your version of history. It is no doubt more comforting to you to do so.

    (Now cue the gratuitous insults…)

  149. Wizard Emeritus says:

    “Talking about OEMs seems to have struck a nerve. It would be smart for the Russians to notice the shills’ flurry of activity. I wonder if a solution would be to outbid M$ and pay the shills to agitate for ReactOS or HaikuOS.”

    What flurry of activity is that Mr. Fish? You have three posters here who have decided to point out Robert Pogson for the fraud that he is. As far as being paid to agitate for obscure garbage like React (Its like running windows 2000) OS or an OS (HaikuOS) based on an OS (BeOS) that never lived, no thank you. We are getting payment enough laughing at Tinfoil Hat wearing nutters like You Mr. Fish.

  150. lpbbear says:

    “The real problem Pog is the market down-voted those Linux netbooks and instead preferred XP (an OS which had already been superceded by Vista) to Linux, to the point Asus had to abandon their original plan of ultra – minimalist storage space and Linux and starting offering XP”

    That is not what happened. Xandros was the OS behind the Asus Eee PC. At the time Microsoft did not see this coming at all. Microsoft had forced Xandros out of the desktop OS market by arm twisting Xandros into another (cough cough) “partner” agreement. Microsoft was at the time busy trying to foist “Vista” on consumers and had planned on EOLing WindowsXP. Xandros had found a loophole in the MS/Xandros partner agreement and was working with Asus to provide the Linux based OS originally used on the Eee PC mini laptop. Once the popularity of the device began to get the notice of Microsoft they realized Vista was too resource hungry to run on an Intel Atom system and backtracked on ending WindowsXP. They began arm twisting manufacturers to use WindowsXP on the systems.

  151. Modular sunfish says:

    kurkosdr wrote “Asus had no problem shipping a Linux laptop (Asus Eee PC) and marketing it.”

    And it sold great until stocks were gone. Before they could produce more, M$ came down on Asus like a ton of bricks and stopped production of any more. The rest were hobbled with XP which, of course, sold poorly. At the same time, M$ diddled the definition of netbook hardware to ensure that they stayed out of competition with even low-end notebooks.

    M$ has never succeeded when the market has been open to competition. They, their shills, and their fifth-columnists play for the long game though. We haven’t even begun to feel the pain that restricted boot and UEFI are causing. Russia really will have to get its own hardware supply if they want to gain technological independence.

    Talking about OEMs seems to have struck a nerve. It would be smart for the Russians to notice the shills’ flurry of activity. I wonder if a solution would be to outbid M$ and pay the shills to agitate for ReactOS or HaikuOS.

  152. ch says:

    Mr Pogson, it’s understood that you don’t understand a lot of things because you don’t want to understand. Now, however, this has really backfired:
    (Of course you still don’t want to understand, but for any interested readers around I shall explain anyway.)
    First, what does “force the OEM to preinstall” mean in this context? Preinstall Windows instead of Linux? Of course not, since Linux was even less of an option in 1994/1995 than it is today. The author of your quote is obviously referring to “preinstalling Win95 instead of Win3.x”.
    Second, let’s translate the whole quote into plain English: “We have to make Win95 so good that customers will WANT to have it on their machines instead of Win3.x, only then will the OEMs have to preinstall Win95 instead of Win3.x”
    So your quote proves quite clearly that MS itself did NOT have the power to force OEMs to preinstall Win95 – that power lay with “we the (vast majority of) customers”. Thus MS had to vow us to influence the OEMs. (You might recall that Win95 did sell quite well in retail.)
    In other words, your “proof” actually proves what the “bad guys” around here have told you all along: That OEMs bundle Windows because we the customers want our machines to come with Windows (because we will run them with Windows anyway). Thank you for putting that straight!

  153. kurkosdr wrote, “Outdate documents shown as “proof” aside, Asus had no problem shipping a Linux laptop (Asus Eee PC) and marketing it.”

    The truth is never outdated. Why did it take Asus a decade or more to ship such a wonderful product? GNU/Linux certainly was usable in the mid to late 1990s. The distro I used, Caldera eDesktop, was just about perfect in 2000. It worked flawlessly on the identical hardware that Lose95 shipped by M$ to Computers for Schools in Montreal crashed daily. I’m sure RedHat or Debian GNU/Linux were in good shape in those days too. All ASUS had to do was pick one but they were afraid of M$.

  154. kurkosdr says:

    Outdate documents shown as “proof” aside, Asus had no problem shipping a Linux laptop (Asus Eee PC) and marketing it.

    And Dell ships Ubuntu laptops, as advertised on the Ubuntu page.

    The real problem Pog is the market down-voted those Linux netbooks and instead preferred XP (an OS which had already been superceded by Vista) to Linux, to the point Asus had to abandon their original plan of ultra – minimalist storage space and Linux and starting offering XP and later Windows 7 in their netbooks on proper hard drives and even beef up the specs to take Windows XP and later 7.

    Again, the market decided that. Linux netbooks and Windows netbooks were sold aside, and in fact the Linux ones had a lead.

  155. Further to this matter of M$ forcing OEMs to install TOOS:

    This is from one of the exhibits in US DOJ v M$. I love this database…

    “We must make Memphis compelling to OEMs at retail – only by making Memphis compelling at retail will we be able to force the OEM to preinstall.”

    That was about Lose ’95 and they definitely had it in their minds to lead OEMs around by the nose by fair means or foul. Over the years they used many levers: exclusivity (with OEMs, ISPs, and even web sites), per PC prices (regardless of OS), no No-OS PCs, shifting file-formats, ISV lock-in, and of course, hiding the price of the OS under an NDA so consumers had no idea of the price they were paying, the ultimate anti-competitive act. ie. an OEM couldn’t advertise GNU/Linux as a lower-cost OS because they could not reveal the price of M$’s stuff. Dell and others did that briefly but only for a few days, no doubt because some M$-goon phoned them up.

  156. Ivan says:

    I’m sure it will be a roaring success. Just like Red Star, watermarking everything it touches. Freedom!

  157. DrLoser says:

    One hundred Syrian Freedom Fighters dead.

    But a thousand Russian Government desktops converted to GNU/Linux!

    Bubba, have I got a deal to sell you!

  158. DrLoser says:

    M$ arranged that so it did not have to compete on price.

    No, M$ did not so arrange. Please restrain yourself from being a buffoon.

  159. DrLoser says:

    Here’s a challenge for all you clueless nutters, then.

    Show me a single government department, anywhere in the world, that has been rebuffed by an OEM when trying to source 10+ PCs (desktops or laptops) from any OEM at all.

    Just a single instance, from a single OEM.

    Do that small thing, and I will admit that I am wrong. Microsoft has a lock-down on OEMs.

    Fail to do that small thing, and I will just continue to laugh at you as a bunch of tepid failed psychotic paranoid dingbats.

    I’m pretty sure I’m going to win on the dingbat thing.

  160. DrLoser says:

    He’ll [the head of Lenovo] soil his pants for fear of having to choose between losing the big contract or losing the continued blessings of M$.

    You are insane, Modular Sunfish.

    And in good company around here, apparently.

  161. DrLoser says:

    Just out of interest, Robert, why is it so important to you that the Putinist State claims to be “adopting” GNU/Linux?
    I mean, it’s hardly like you have a deep desire to see Russia approach the IT perfection of, say, the Italian Defence Department.
    Because, if for some as yet unexplained reason, GNU/Linux might achieve that efficiency whilst being used by the Russian Military Machine …
    … The only plausible result would be that Putin would be able to kill more Syrian opponents of President Assad, much more quickly, and at a cost/benefit price you cain’t beat!

    Do, just for once, get your silly little senile paranoid story straight, Mr Robert Pogson.

  162. Wizard Emeritus wrote, ““Why must the price of TOOS be hidden in free markets?” Because its not Robert Pogson.”

    I have never seen the price of TOOS listed on any website along with the PC where it is pre-installed. That information is certainly not listed on any local retail shelves. M$ arranged that so it did not have to compete on price.

  163. kurkosdr says:

    Providing FreeDOS avoids having to prove support for wireless, graphics and other peripherals.

    Yeah, because modern laptops have such complicated peripherals… oh wait, most of them are an Intel CPU+GPU bundle and Intel HD Audio.

  164. Wizard Emeritus says:

    “As to FreeDOS, that’s just a work-around for the ban on selling whitebox systems and not what anyone these days would consider a proper OS. ”

    Really? I thought it was just a place-keeper to demonstrate/test the basic functionality of the computer before it is sent out the door by the OEM.

    But do please provide some evidence of this assertion, won’t you?

  165. Modular sunfish says:

    DrLoser wrote:”what do you think the boss of Lenovo is going to say?”

    He’ll soil his pants for fear of having to choose between losing the big contract or losing the continued blessings of M$. Maybe the order from Russia would be big enough to overcome that, maybe not. Only when M$ gets even weaker will the OEMs be able to move more freely. For ages, GNU/Linux offerings on notebooks and netbooks have been fairly ephemeral. What’s for sale for a while is not necessarily available a short time later. What M$ did to Asus over the EEE is a case in point.

    As to FreeDOS, that’s just a work-around for the ban on selling whitebox systems and not what anyone these days would consider a proper OS. Even for Apple systems, many shops keep them displayed behind shelves or other obstacles, if displayed at all. Other tricks include keeping a password on the screen saver. In some you have to threaten to walk out and order from the net instead to get them to even admit they have some in stock or offer one for sale.

  166. kurkosdr says:

    The Assad regime is illegitimate. They were dropping barrel bombs on civilians, remember?

    You do realize you just gave me the opportunity to ask you some very interesting question?

    1) How do you define a legitimate (non-illegitimate) regime?

    2) Is the Saudi Arabia regime legitimate?

    3) If the Saudi Arabia regime starts dropping bombs on rebels trying to overthrow them inside their territory (if that happens sometime), will they still be a legitimate regime?

    4) If the Saudi Arabia regime starts dropping barrel bombs on rebels trying to overthrow them inside their territory (if that happens sometime), will they still be a legitimate regime?

    You have to answer all those questions, otherwise the default definition for your “legitimate regime” term will be: “a regime that is cozy with the US, not with Russia”

  167. kurkosdr says:

    The monopoly the M$ has on the OEMs also means that it is all but impossible to order x86 hardware without TOOS.

    I bought my lenovo Z70-80 with FreeDOS (later installed Windows 10 on it). And yes it was from a real retailer ( e-lenovo.gr ). The Z product line is no longer sold, but other product lines like the Y product line have DOS variants, as you can see by visiting the site. Intelligence is not your strong thing, right?

    You know why brick-and-mortars don’t offer FreeDOS machines? Because space. It’s like going into a bookstore and asking for uber obscure titles, stuff you should be ordering from Amazon or some other online retailer.

    Hence, I have good-faith belief that you are a pathological moron and a despicable smearer.

  168. Wizard Emeritus says:

    “No, they didn’t. They would have gone without my help, and rightly so, because I wouldn’t be able to help them anyway.”

    Fair Enough.

  169. DrLoser says:

    The usual paranoid gibberish, I see. Just to take a representative example:

    The OEMs are another problem. The monopoly the M$ has on the OEMs also means that it is all but impossible to order x86 hardware without TOOS.

    All but impossible? Surely not. It takes a little assiduous googling, as an individual, but let’s be honest — it’s really not hard to find Dell’s Ubuntu offerings if you’re looking for them.

    Government organisations don’t tend to do this, however. They have staff dedicated to buying equipment en masse. They are one of the prime targets for hardware salespeople, because — guess what? They generally buy in bulk, which makes the commission very appealing. If (as Dell, say) your star salesman in a territory tells you that the government is looking to buy 1,000 new laptops, but they better work with Linux, you make sure that you have 1,000 laptops with Linux preinstalled to sell them.

    But maybe I’m wrong about the sales staff thing. Maybe that doesn’t work. OK then — explain this to me. If (say) the IT Department Head for Moscow phones the boss of Lenovo — or even a white-box Chinese/Korean also-ran — and asks for a pricing on 1,000 laptops, but they better work with Linux … what do you think the boss of Lenovo is going to say?

    But maybe I’m wrong about the market-driven greed of capitalists worldwide. Only one further exhibit, M’Lud:

    Dell’s head of China told The Wall Street Journal that NeoKylin Linux is shipped on 42 percent of the PCs it sells into the country, primarily for the commercial and government PCs that Dell specializes in.

    Now, I happen to prefer that the paranoid nutters on this site remain, endearingly, paranoid. It’s a source of endless amusement.

    But it would be even more amusing if you were all, for the sake of a temporary argument about Monopoly and Slave OEMs and so on, prepared to completely ignore and/or trash this little item of information.

    Please. Be my guest.

  170. luvr says:

    Wizard Emeritus wrote, “So all because of your bigotry, a personal who actually had so run applications instead of operating systems went without help.”

    No, they didn’t. They would have gone without my help, and rightly so, because I wouldn’t be able to help them anyway.

    The best that I can do for anyone coming to me for help, is to point out that, if they consider Windows important, then they will have to look for help elsewhere. Which is exactly what I do. If I tried to <i"help" them with Windows, I would only make them more miserable than they already were. And I wouldn’t be any happier myself, either.

  171. Wizard Emeritus says:

    “Why must the price of TOOS be hidden in free markets?” Because its not Robert Pogson. Anyone who buys a desktop or portable knows that windows comes installed, so that all they have to do is load or download and install the applications that they purchased their computer for (including FOSS BTW) and go.

    OF course being and inveterate cheapskate, you think that they are going to recoil in horror over that $50-100 added to their new computer, toss it (and their windows based applications) away and install your favorite non commercial, “free” OS.

    Riiiight.

    You are as usual living in a fantasy world.

  172. Wizard Emeritus says:

    “CORRECTION: If they needed Windows, then I could NOT help them, of course.”

    So all because of your bigotry, a personal who actually had so run applications instead of operating systems went without help. NOt a nice thing to do IMHO, but then again given your attitude thew were probably lucky.

  173. Wizard Emeritus says:

    “Same goes for resellers, some will go to great lengths to keep customers from ordering a real OS.”

    How about the real applications that their customers use the computer for?

  174. luvr says:

    CORRECTION: If they needed Windows, then I could NOT help them, of course.

  175. luvr says:

    Deaf Spy wrote, “Windows, Mac OSX and MS Office are just the natural choice for people who want to get the job done.”

    Hmmm… I’m not too sure about Windows. In my experience, there’s a growing number of people who find that Windows does not work for them. Obviously, my observations aren’t based on a representative sample of the population, so if you want to consider the number infinitesimally small and the growth rate insignificant, then that’s fine by me.

    All I can say is that I know of three people that have recently dumped Windows and switched to a MacBook, and they are incredibly relieved that they can finally get work done. Their only regret is that they didn’t make the switch much earlier.

    Then there are these five families that came to me because “their computer was broken”, when in fact, Windows must have turned itself into a mess. I told them that if they “needed” Windows, then I could help them, and they would have to seek help elsewhere. They replied that if I thought I had a better option, they would be willing to give that a try, so I zapped their Windows system, and replaced it with a Linux distro (for two of them, I installed Ubuntu with the Unity desktop, for two others I set up Linux Mint Cinnamon, and for the fifth I opted for Xubuntu). I don’t get any complaints from any of them. Quite on the contrary, they are thankful that their computer finally works right.

  176. Deaf Spy wrote, “Ready to accept the bet? “

    Nope. I’ve been busy diversifying my investments in the face of mini-panics in the stock market. I’m about gambled-out… 8-(

  177. Modular sunfish says:

    Deaf Spy wrote, “EU decrees nothing yet”

    Actually, the EU has been illegally insisting on TOOS. It’s a long term, ongoing problem.

    https://joinup.ec.europa.eu/news/many-software-tenders-eu-maybe-illegal
    http://www.networkworld.com/article/2201337/careers/eu-bodies-plan-upgrade-to-windows-7-without-public-tender.html

    The OEMs are another problem. The monopoly the M$ has on the OEMs also means that it is all but impossible to order x86 hardware without TOOS. As Robert points out Android/Linux is already common. Via Chromebooks it’s coming into the notebook and desktop space. GNU/Linux is increasingly available though the availability comes and goes, not always from the same vendor.

    Same goes for resellers, some will go to great lengths to keep customers from ordering a real OS. When Russia breaks the grip on the OEMs, M$ will have lost most of its traditional hold over them.

  178. matchrocket says:

    They can change the software alright, but it’s probably too late for the firmware. For some time now Intel and most likely other vendors have been building spyware into their devices. I’m talking printers, motherboards, video cards and just about anything connected to your system. It will be very difficult for the Russians or anybody else to scrub their systems clean. Maybe even impossible.

  179. Deaf Spy says:

    I think 2016 will be the year of ARM on the desktop and GNU/Linux will be there.

    Robert, let’s make a bet. I claim that 2016 will not be the year of ARM on desktop.

    If I lose the bet, I will donate $200 to your favorite charity. If I win, you will do so to mine. Proof of transaction: copy of the bank order or PayPal transaction.

    Ready to accept the bet?

    I can even go further for another $200 that in 2016 we’ll see an increasing boom of mini PCs with Atoms and CoreM.

  180. Deaf Spy says:

    So, if EU decrees TOOS…

    This is where you go terribly wrong, Robert. EU decrees nothing yet. Windows, Mac OSX and MS Office are just the natural choice for people who want to get the job done.

    See how Anroid thrives in EU without having anyone to decree it. But Linux on desktop can thrive only if being enforced.

  181. Modular Sunfish wrote, “is Russia willing or able to make a break from Intel at the same time?”

    The whole world has accepted Android/Linux on ARM. It’s a tiny step to use GNU/Linux on ARM. The advantages are immediate: smaller/cheaper/cooler chips. Folks are putting 8 2gHz ARMed cores in smartphones these days with 2gB RAM. For desktops all they have to do is increase the RAM and storage options. I think 2016 will be the year of ARM on the desktop and GNU/Linux will be there.

  182. kurkosdr wrote, “Russia “invaded” Syria with the permission of the local government?”

    The Assad regime is illegitimate. They were dropping barrel bombs on civilians, remember? The vast majority of Syrians are not happy that Russia is giving Assad air-support. Russia claims to be attacking ISIL at Aleppo but ISIL is not in that neighbourhood.

  183. kurkosdr says:

    So, Russia “invaded” Syria with the permission of the local government? I guess the US “invaded” Saudi Arabia when they sent troops there during the Gulf War with the permission of the local government. Oh I forgot, being friends with non – Democratic governments is good only when the US does it…

    Russia did invade Ukraine though.

    But when it comes to Syria, Russia did the right thing, protecting the not- yet – overrun by extremists parts of Syria. The only parts of Syria where extremism doesn’t rule are the parts controlled by Assad and the Kurds. Do you really think the Syrian “opposition” will establish a democracy and a functioning state if Assad falls? Question: Did it happen in Libya after Gaddafi fell?

  184. Modular sunfish says:

    I agree, even before all those concerns you list there isn’t any better choice than FLOSS. Those concerns may be enough to push things over the hump and get rolling forward after all these years.

    The largest barrier is that M$ has been able to prevent a free market there and elsewhere in the past, blocking all other Operating Systems. It would help all nations if more were published by M$ victims about the specific tactics used in regards to both OS and applications. Right now the nations are operating more or less in the dark, unable to learn from the experiences of the others. That gives M$ an advantage in that it can often re-use the same tactics against new victims /chumps.

    About the chips, is Russia willing or able to make a break from Intel at the same time? Or would Intel be willing to help in a move that hurts M$ and support real Operating Systems? If not, there were some MIPS based computers out of China for a brief period. Sparc used to be open(-ish). POWER, being IBM, is probably out though.

  185. Modular Sunfish wrote, “This is something like the third or fourth time they’ve made such an announcement. The first one I recall was that they would move to FOSS by 2010, but what ended up happening was a purge of staff capable of actually setting up a FOSS infrastructure. Then again in 2010 they made a similar announcement, aiming again for a few years in the future.”

    What’s changed and is really pushing this is a flood of pressures:

    • sanctions over Ukraine/Syria/whatever preventing M$ from doing as much business as they usually do in Russia,
    • lower price of oil due to USA/Saudi reactions to ISIL and Russia,
    • lots of economic cooperation with China giving hardware less dependent on USA,
    • the Snowden revelations that USA is spying/sabotaging everyone,
    • news that several states, not just USA are actively exploiting TOOS to spy/sabotage IT and infrastructure globally, and
    • the wide adoption of FLOSS and ARM as economically and practically viable technologies in government, business, and personal IT.

    Given the global issues and Russia’s particular concerns there isn’t any better choice than FLOSS and GNU/Linux. They can change that in short order. The hardware change is much more difficult but China will be only too glad to sell them lots of chips as needed at a good price in exchange for petroleum/gas.

    Now that China has a serious downturn on exports, they will be driven to cooperate more with Russia and to seek more independence for their own IT. Then there’s India which just can’t afford TOOS but can afford FLOSS and GNU/Linux. The Wintel monopoly is definitely in decline with not much upside.

  186. Deaf Spy wrote, “if EU allows every administration to make its own choice, fine. When EU decrees “use Linux”, this is not a free market anymore.”

    So, if EU decrees TOOS, it’s a free market but not if they decree GNU/Linux? Hypocrisy. EU is choosing GNU/Linux for many reasons including price but also for security, independence, use on old hardware, ease of maintenance, etc. a bunch of really good reasons.

    GOOGLE, M$, and a bunch of other global corporations are in bed with USA spies, willingly or not. The US government has even tapped into cables to/from businesses and whole nations. The EU is wise to try to control their own IT closely, rather than trusting those guys. The best/easiest/cheapest way to do that is to use FLOSS on open hardware, like ARM. It is totally feasible for EU to get a licence from ARM or one of its partners to control all the software and firmware in their IT, greatly increasing the assurance that it’s not a direct line to USA/China/ISIL spies and saboteurs. FLOSS is on the radar and ARM is as well. Within a few years EU and other governments will insist on FLOSS on open hardware right down to the chips that drive peripherals. They could start from scratch but why do that when GNU/Linux is available and working? That saves $billions in R&D and permits much more rapid migration to secure IT. Sticking with TOOS guarantees continued insecurity with no hope for secure IT.

  187. Deaf Spy says:

    GNU/Linux succeeds wildly there for governments who have greater considerations than individuals and businesses, like freedom and security and price.

    I can’t speak for India, but I surely can for EU. There, we witness a very Minuch-like attitude. Leftist, which EU is currently flooded with, as obsessed by punishing bad evil corporations (see actions against Microsoft and Google), and put everything under a political considerations, and never economic one. EU bureaucracy is ridiculous and is costing taxpayers fortunes. No one at EC is even thinking of costs. They only think of how to sound “free” and “liberal”, and so on crap.

    When you enforce politics, you are no longer operating in a free market. See, if EU allows every administration to make its own choice, fine. When EU decrees “use Linux”, this is not a free market anymore.

    Finally, Robert. No government has ever been more concerned about costs than an individual or business. Never, ever. Explanation is simple. Individuals and businesses spend their own money and are very cautious about that. Government spend other people’s money. Easy.

  188. Modular sunfish says:

    I hope this goes forward, but will take the announcement with a grain of salt. Backing up the words with deeds is needed to make a difference from the other announcements.

    This is something like the third or fourth time they’ve made such an announcement. The first one I recall was that they would move to FOSS by 2010, but what ended up happening was a purge of staff capable of actually setting up a FOSS infrastructure. Then again in 2010 they made a similar announcement, aiming again for a few years in the future.

    That said, if they put some strength behind the move, there is a lot of potential for progress. And this time there are several prime examples, like Italy and Germany, to model.

  189. Deaf Spy, quoting 2+2=5, wrote, “There is no way FLOSS can succeed on desktop in a free market.”

    So, EU and India are not Free Markets? GNU/Linux succeeds wildly there for governments who have greater considerations than individuals and businesses, like freedom and security and price. Consumers, largely, are not even aware that they pay for M$’s OS because it’s hidden in the price but we can see M$’s balance sheets and there’s $billions annually taken in as a tax on IT. Why must the price of TOOS be hidden in free markets?

  190. Deaf Spy says:

    Actually, not quite so. The real reason is (from your own source, Robert):

    “Microsoft, Google and other U.S. companies “reached the point of no return” when they complied with sanctions over Putin’s annexation of Crimea by halting all business with the peninsula, according to Klimenko”

    In other words – no one wants to sell us their services and software, and we are broke because oil is cheap, and no one wants our gas anymore.

    Nothing to do with freedom, productivity, or whatever. Just – we can’t keep using services from Google and Microsoft.

    I told you, Robert. There is no way FLOSS can succeed on desktop in a free market.

Leave a Reply

Your email address will not be published. Required fields are marked *