Workstation Computers From HP

This is kind of a different sentiment from some OEMs we know…
“It’s nice to have choices

All HP Workstations can support a variety of operating systems. HP engineers work extensively with Windows® and Linux operating system providers to verify top performance, flexibility, reliability, and compatibility with HP Workstations. We conduct joint engineering collaboration with industry partners long before systems are introduced.”

Further, when you look at the products we see things like this:

Form Factor Rackable minitower
Available Operating Systems Windows 8 Pro 64-bit
Windows 8 中文版 64-bit
Windows 8 Pro Downgrade to Windows 7 Professional 32-bit
Windows 8 Pro Downgrade to Windows 7 Professional 64-bit
Windows 7 Professional 32-bit*
Windows 7 Professional 64-bit*
Windows 7 Ultimate 64-bit*
HP Linux Installer Kit
Red Hat Enterprise Linux Desktop/Workstation* (1 year paper license; no preinstalled OS)
* This system may require upgraded and/or separately purchased hardware and/or a DVD drive to install the Windows 7 software and take full advantage of Windows 7 functionality.

The point is that the OS is not bundled with the workstation and users can get to see the price of that other OS. They do supply that other OS pre-installed but you can see the price difference between that and installing GNU/Linux.
With identical hardware, getting that other OS costs about $100 more.
HP_workstations_comparison I would bet the cost of tweaking the OS would amount to more with that other OS too. That has been my experience, hours instead of minutes with GNU/Linux.

Now, obviously, HP can do that for all their PCs and they could bundle GNU/Linux as well. Pick any of the major desktop distros and consumers would be happy. It costs HP nothing to give consumers a choice. It is a distortion of the market to give M$ the bye on retail shelves. Clearly there is a market for GNU/Linux PCs.

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.

108 Responses to Workstation Computers From HP

  1. That Exploit Guy says:

    @oiaohm
    Does the state of the drive…
    The problem you are trying to raise (as I understand it – I am just not really that good at reading gibberish) is, simply, “What should I do if the power goes out when SDelete is wiping the drive?”
    Again, to the contrary to your assertion, the space-hogging files contain only blocks that are otherwise unallocated. What is between you and the solution, thus, is simply the delete button.
    Exactly why does sdelete have to claim all of free space at 1 time.
    As far as I am aware, none of SDelete, cipher, and CCleaner claims “all of free space at 1 time”. Space-hogging files grow in size only as the program proceeds with wiping more and unallocated blocks.
    Because twice in the running of cipher the disc will report full. So anything need to write a file at that time cannot.
    The main and only concern here is I/O saturation, which you have yet to explain as to why it’s worth putting up with. If you are by any chance dealing with sensitive information that warrants secure wiping, chances are you have already got workstations with removable hard drives (in mobile rack enclosures) that personnel run software and store their work on and keep in safes when not in use. The only times you want to wipe free space on your hard drive are when you don’t want your wifey to uncover the dubious hobbies you have been engaging in on the internet using basic file recovery techniques. The need to wipe free-space on a server shared by people that are not meant to know each other’s secret should not even arise to begin with.
    Linux solution is managed disruption
    No. It’s just disruption – one that has no justifiable reason for.

  2. Dr Loser says:

    Dr Loser have you ever used fuser before to allow you to change file system modes. I guess not. Its not something your normally script.

    I’m not asking you to script it for any practical purpose, oiaohm. I’m asking you to pick a single configuration, entirely at your own choice, and to step us through the shell commands which achieve your goal on that configuration. Feel free to add comments in the cases where you need to side-step “road-blocks.”

    There now, that isn’t too difficult, is it? You are, after all, a self-acknowledged 99.999 (or is it 99.99999) System Admin! A mere handful of minutes should suffice to demonstrate your innate superiority in this field.

    Or is it?

  3. oiaohm says:

    If you do script the fuser bit you script it to match the system its will be used on.

    Particular blocks are not simple to get past. Only admins with skill can work around them.

    Neither Dr Loser or That Exploit Guy are 5×9 Linux Admin. Or you would know how todo this since 5×9 linux admin has to have this your bag of tricks for filled file systems.

  4. oiaohm says:

    Dr Loser not everything can be scripted. I did give you the individual commands.

    This stage can be simple bash script.
    mount –bind mountdirusersareusing duplicatelocation
    mount -t overlayfs overlayfs-olowerdir=duplicatelocation,upperdir=writeablelocation mountdirtobeworkedon

    The sorting out of fuser results requires intelligence. Some of the write files will be services that require restart. Some with be users that solve when they log out or stop processes.

    Some is taking into account the bugs/limitations of overlayfs. Stage 4 is the tricky bit master of bash is not in fact good enough.

    I have not stated something unskilled personal can do.

    Dr Loser have you ever used fuser before to allow you to change file system modes. I guess not. Its not something your normally script.

    Ok I could do a fuser -wkM That is savage will disrupt users and will result in services not restarting.

    Dr Loser my method is not magic. Close to the same method is used when you have a drive that will not unmount to find the problems and resolve them.

    There is a difference in level here Dr Loser. Skilled 5×9 uptime admins can pull off what I am describing and its run of the mill simple. General admins not so much.

    Running out of disc space on a volume triggers the same kind of problem Dr Loser.

    This is straight out of the 5×9 admin list of dirty tricks. Tricks that will allow you to get to the next restart without having to restart the system before then.

  5. Dr Loser says:

    Drops in DBAN/EBAN and then tosses the drive into the shredder, no worries.

    This may come as something of a surprise to you, Mr Dougman, but I agree with you 100%.

    If we ever set up a two-man security consultancy (I’ll handle the Linux stuff; I trust you with the Windows stuff), we could absolutely paste oiaohm.

    This might be the beginning of a beautiful friendship!

  6. Dr Loser says:

    Dr Loser so you want point by point.

    Indeed I did, oioahm, and I specified the particular order.

    First, your supposed aim.

    Next, how you would achieve that aim (whilst avoiding “roadblocks,” which as far as I can see equate to “no operating system on earth would ever in a million years be able to magic these roadblocks away.”)

    What I was looking for, as the simplest possible request, was a single paragraph stating what you want to achieve.

    You fail.

    Regrettably, I wasn’t clear enough about the steps to achieve this. No matter, we are all of us here experts in the Bash shell!

    There, that’s the thing! Give us a Bash script (assuming installed components — I don’t wish to be too hard on you) that will achieve your goal. Should be simple to test, really. You wouldn’t want us to take you on trust, would you?

    You can miss out the intermediate steps. Ten or twenty lines should be enough.

  7. oiaohm says:

    That Exploit Guy
    –As it has been explained to you by Russinovich, the file is merely there to hog space, i.e. to prevent other process from being allocated the same blocks SDelete is wiping–

    Does the state of the drive being filled get written to disc under Windows yes it does is windows early boot smart enough to delete it no its not. A equal to a O_tmpfile has not been used. Major error in fact.

    That Exploit Guy so if it only to prevent SDelete area its working on from being allocated to another process why does the file have to be so large. Remember defrag under windows can claim ownership of groups freespace blocks without claiming them all. The said reality here it is possible to implement the ideal under Windows. Yet cipher does not.

    cipher under Windows is written wrong. Every statement from Russinovich does not change that fact. Exactly why does sdelete have to claim all of free space at 1 time. There is no functionally of sdelete that in fact requires this.

    Linux wipe tool forces you to address these by closing those with write

    –1) interrupting everyone’s work–
    This will happing running cipher on Windows. Because twice in the running of cipher the disc will report full. So anything need to write a file at that time cannot. This is random uncontrolled disruption. Crashes at those times the full state can be synced onto disc.

    Linux solution is managed disruption.

    Ideal need to take disc wiping on active file systems like defraging. Small blocks at a time. Neither Linux or Windows implements the ideal. Solution Linux is technically better.

    That Exploit Guy you are not comparing to what I have to compete against. I said users could keep on using the machine. I did not say it was 100 percent disruption free.

    Yes order of preferred.
    No Disruption
    Controlled Disruption.
    Totally not preferred random disruption that cipher is.

    –2) changing a file system that is intended to be read/writable to read-only.–

    But the reality its changeable. Your arguement that change it was impossible was in fact lieing.

    Changing a file-system to read only with a new write file-system over the top has another usage out of disc space. Its a basic Linux trick.

    That Exploit Guy basically I have not said the Linux solution is perfect. Early on I did say another solution was ideal that neither implement.

    The huge file filled with clearing data in fact comes from Unix. You see some Unix grey beard recommending like /dev/random to file on btrfs to so call clear the free space or worse /dev/zero (yes btrfs with compression max size file on btrfs filled with zeros consumes 4 blocks. So yes totally does not work. btrfs method is horible.

    Under btrfs you are meant to use it transparent volume add and remove functions. So remove volume dban it readd volume to filesystem. Perfectly transparent to users. Completely horible on time and resources consumed.

    Windows default tools have some quite major bugs.

  8. dougman says:

    Drops in DBAN/EBAN and then tosses the drive into the shredder, no worries. 🙂

  9. That Exploit Guy says:

    Two posts. Always two posts. It’s a disease, I tell ya.
    If these are not address and you just fill the drive with data the way cipher in windows does you can have unexpected write failures to applications resulting in bad state
    Cipher/SDelete does not “fill the drive with data”. As it has been explained to you by Russinovich, the file is merely there to hog space, i.e. to prevent other process from being allocated the same blocks SDelete is wiping.
    Linux wipe tool forces you to address these by closing those with write
    Here, you have just completely defeated your stated intent of “secure deletion of files while users are using [the] machine” by 1) interrupting everyone’s work and 2) changing a file system that is intended to be read/writable to read-only. All this yarn you are spinning does not actually solve any problem that you believe it solves – it just makes you look stupid.
    Just go and take your med already.

  10. oiaohm says:

    Doing / is trickier as you used a cgroup filesystem namespace if you used 1 single partition install. But in multi partition install/by book install / is its small partition. With nothing normally writing there. That is the only version of switch from read-write to read only that is major different.

  11. oiaohm says:

    Dr Loser so you want point by point.

    The road blocks up front are the write files. If these are not address and you just fill the drive with data the way cipher in windows does you can have unexpected write failures to applications resulting in bad state. This can be horrible bad state like half written registry.

    Linux wipe tool forces you to address these by closing those with write to where you want to work and providing redirection for those that need to write.

    * to change from file system from read write to read only where new write files have to be used until next reboot so you can perform some task requiring the file system read only.
    1) mount –bind mountdirusersareusing duplicatelocation
    2) mount -t overlayfs overlayfs-olowerdir=duplicatelocation,upperdir=writeablelocation mountdirtobeworkedon
    Those two steps now stop new write files by users appearing on the mountdir you wish to work on.
    3) mount -o remount ro /dev/sda1 or what ever the real device is behind what you want to clean. If it succeeds good.

    4) fuser -m /dev/sda1
    /dev/sda1 is used again for what ever the real device is.
    This is to find out what processes blocked the remount.
    All PID numbers followed with F have write files open and are blocking the mount -remount ro. So now you have a list to work threw. LIke a service restart will close files and reopen them due to overlayfs the opening files will not be on the file system you are trying to change from read-write to read only. Some will be handled by users ending the processes themselves. Since you have list of processes you can in fact contact that users and ask them to save close and reopen. Minor disruption to only those blocking.

    In fact running fuser can tell you drives that are current read write that can be straight up swapped to read only. Deal with the processes in way change do mount -remount ro
    5) start what ever process required the device to be read only.

    There you have it I have gone from a filesystem being read-write to read only.

    This could be done to perform data recovery on drive as well.

    Something That Exploit Guy keeps on saying is prohibited. 4 step process. Ok fuser -m /dev/sda1 could be done before mount -o remount ro but that is more personal choice.

    That Exploit Guy expressly prohibited my ass. Its only expressly prohibited if you don’t take on and address the issue of the open file handles with write status. Conditions are critical.

    The thing is cipher it self is using a space hogging file does not need to be. If you have some process with a huge file that does not have a restart or start/stop function and you cannot wait it out then you run into a road block cannot be beaten. The program would not be reboot safe either. So finding something like this is a worry.

    Yes the only stuff you cannot that work around are problematic and needs to be addressed anyhow.

  12. Dr Loser says:

    The way I describe has road blocks. You have to resolve those road blocks up front.

    Actually, “the way you describe” is nothing more than a fail-train based upon an overactive and under-informed imagination, oiaohm.

    Bet let’s be professional here.

    Describe (in a single paragraph, please) what you intend to achieve.

    After that, please be professional for once, and detail each and every single step you take to achieve that goal.

    Highlight the “road blocks” along the way. Explain how you get around those “road blocks.”

    You can’t do that, can you?

    Cue wall of gibberish.

  13. That Exploit Guy says:

    That Exploit Guy the big point that you keep on missing is that you can change a filesystem under Linux from read-write back to read only.
    I’ll repeat this one more time for you, and no more:
    As stated in your source, this is an action explicitly prohibited by the kernel under the exact circumstances where the space-hogging file (or, in your word “mother of a file”) is need. It is also irrelevant to the fact that, once again, you have been caught lying due to the irrefutable evidence that the free-space wiping software neither achieves your stated purpose nor feature any of the imaginary mechanisms that you insist on arguing about.
    Also, stop bouncing me two waffles of sweet nothing for every reply I give you. I know this is your habit, but I think your time will be better spent directing them to your healthcare professional.

  14. oiaohm says:

    That Exploit Guy in fact under Linux you could do it using normal unlinked files. Its one of the differences between Windows and most other Posix systems.

    With windows files cannot in unlinked state and protected from being overwritten.

    Yes the reason why you can delete files on Linux without closing the processes using them is the fact they are protected by the open process from being overwritten even that they could have already been marked on the file-systems as deleted.

    c.f. Russinovich only said the blocks had to be allocated. Did not say they had to be allocated to a written to disc file. Allocated in ram is in fact good enough as long as the OS will respect this. Yes most Unix and BSD and Linux based OS’s.

    My method obeys c.f. Russinovich limitation. I did allocate the unallocated blocks. Just happens to be to a file that will it self automatically return to be unallocated if the process is terminated by any means. Crash, Reboot, Kill.

    Only issue I have is I need particular physical blocks. So you can start at the beginning of the disc grab the first block of free blocks process them release them then grab the next block along on the disc. Basically take on wiping free-space like defragging.

    So by your stupidity defraging would be impossible as well. Since it moves around free space. That Exploit Guy get what I am meaning about more than 1 way to skin cat.

    If something is impossible be-sure to check if the other-ways are. You never checked all the ways of allocating.

  15. oiaohm says:

    That Exploit Guy the big point that you keep on missing is that you can change a filesystem under Linux from read-write back to read only. Of course you have to get all the write flagged file handles closed and prevent more from opening.

    Read back I said users could still be using the system while a wipe was in process. I did not say the could be altering the file-system that the operation was being performed on. So I nevered lied.

    That Exploit Guy the big thing you miss is block allocation does not have to be written to disc to prevent processes writing to it.

    Allocated in ram saying that is used is in fact enough to stop other processes writing to it.

    When I say new I ready do mean new. That Exploit Guy The feature is Linux 3.11 find a distribution with that.

    http://kernelnewbies.org/LinuxChanges#head-018d712d03d9950dec9e02f0699f15b3e5a28310
    O_TMPFILE Yes you have just created a file that is already deleted. So it never gets written into the base file structs. Yes block allocations to a already deleted file.

    You keep on presuming todo what I need todo I need unallocated. What I need is O_TMPFILE that is already delete so will not exist after a reboot as a file. With the means to assign physical blocks to it.

    There is more than 1 way to skin the cat. Linux feature that a process holding a deleted file prevents it from being overwritten is so ideal its not funny.

    As you say you cannot touch unallocated. Is there any form of allocated that will not remain after a reboot is the question you had to be asking.

    That Exploit Guy I have not being lying. You have been incompetent. You with -o remount presume there was no way to get around the files with write flag when there is.

    You have also incorrectly presumed there is not a way to allocate blocks on a file-system in a temporary way.

    O_TMPFILE yes this automatically becomes freespace if you reboot. One piece is missing. Only 1 piece. A method to assign physical unused blocks to a O_TMPFILE in small enough groups at a time to leave writeable space. So you never run into the case of out of space because the wipe process filled the complete drive.

  16. That Exploit Guy says:

    Learn to read.
    An interesting proposition from someone whose only ability is, apparently, incessant babbling.
    First of all, you have completely failed to appreciate what other people (c.f. Russinovich) has described about the race condition inherent to writing to unallocated blocks (due to the fact that, in most cases, you can’t mandatorily lock an unallocated block – especially not in Linux) and instead invented two that sound just about as stupid as each other.
    Second of all, the free-space wiping software that you links to fails to match your description. The bottom-line here is that the software will only work if you give it a file system that has not been mounted as read/writable, and it does not feature any of the imaginary mechanisms that you have described here. However you attempt to justify, it won’t changed the fact that you have been caught lying about the software (that you have obviously never used).
    Learn to read, indeed.

  17. oiaohm says:

    That Exploit Guy you will most likely ask next how do you take overlayfs out of existence. This is the total amount of noticeable disruption. Called the scheduled reboot. So you have under 5 mins of disruption.

    You can program pam and logind to merge the overlayed upperfs back to the lowerfs on login.

    Secure wipe of freespace is going to be done when a reboot is upcoming. This is not adding another disruption.

    So yes you can leave areas read-write. Just it adds steps you have to clean up latter.

    This is why there is a difference between a Novice LInux user and a Linux Pro user. A Linux pro user knows what is truly impossible. The only reason why I have to reboot/drop to single user is overlayfs does not support lowerfs being altered while overlayfs is over the top. It is possible to have used a lvm snapshot of the filesystem to avoid this as well. But that depends on your system being setup using lvm before you start. Snapshot lowerfs the overlayfs merge the upper back into that then over mount the snapshot over.

    Of course doing this you do generate some mess in mounting information. So at some point you do have to reboot to clean it up.

    Yes 1 reboot required. Operations are able to processed as if you are not cleaning house.

  18. oiaohm says:

    That Exploit Guy
    — Now, how exactly is that supposed to change the fact that the race condition between the wiping software and other processes attempting to acquire the same unallocated block for writing cannot be resolved using your proposed, imaginary solution due to,–
    Learn to read. mount -o remount ro /dev/sda1
    Will proceed if there are no currently open files with write flag. This fixes the issue of applications attempting to write at the same time as the wipe program. If this does processed it safe to use a direct block wipe program.

    Sometimes its not simple to get to the point mount -o remount ro works.

    Not simple would be a location is like /home and need to remain appearing read-write. You can place a overlayfs over it just like a livecd to stop new files with write states from being opened on the lowerfs. The upperfs of the overlayfs takes all the new files opening with write state. Yes this can be a different physical file-system. A physical file-system you are not performing a wipe on. The fun of Linux overmounting. Of course you need to make /home accessible somewhere else. –bind comes into doing that. Note –bind not move. Bind makes the partition appear in 2 places. mounting over the top of somewhere automatically hides what ever is behind. Linux does not audit if the files just mounted over the top of are open. unmount and remount options are the only options that look at what files are open in the mounting options.

    Its a complete combination of steps. Same combination of steps some livecds use.

    Just because 1 option is expressly blocked does not mean there is not a second way to skin the same cat. The second way is using livecd methods. bind to make the directectory accessable else where and ovelayfs mount over the location sending the writes to another location that does not need secure wiping. Then wait for users to log out/restart services as required.

    The result is no file with write flag will be open on the drive. Then mount -remount ro will proceed.

    The prohibited state you are refering to just enforces extra steps. The thing you are missing. VFS stands for virtual filesystem. Its the virtual bit that makes it possible to switch a filesystem read only yet have the location where it was still appear read-write. The key word is appear. The write is happening on a different filesystem.

    Expressly prohibited is not exactly the case. Straight up just doing mount -o remount ro like will work in most cases against /usr why there are no write files open under there normally.

    Only some sections of the file system have read/write. A share that users on a network may not always have users using it. There are times when the simple mount -o remount ro works.

    The complex requires understanding how to exploit bind mounts and mount overlaying and overlayfs and how to track down processes with open write filehandles on particular file physical systems.

    Basically its doable. If you have been forced to use overlayfs you most likely have to line up todo something when no one is using it to merge overlayfs out of existence.

    The way I describe has road blocks. You have to resolve those road blocks up front.

    The problem with cipher under windows is you don’t have to resolve those road blocks. This causes the issue of application attempting to write and no blocks being left. Like installers they query at start of process if there is enough disc space. If they are they presume they will complete. While they are running there are files open in write state. So really unless you can change a file system to read only or have in the OS the means to request physical blocks on a drive that are unused to be allocated to tempfiles(perferablly only recorded in ram as used) you really should not be wiping.

    1 huge file is completely the wrong thing to be doing.

    In fact I decribed how in the prior post
    The simple reality you don’t understand how to work with the Linux VFS.

  19. Dr Loser says:

    So, to recap through your senseless babble, joe:

    There is also window managers that are only available really on GNU/Linux, for instance awesome.

    There are also window managers on Atari that are only available really on Atari. Cry me a river.

    Without awesomewm, I find using a computer frustrating as hell.

    Have you tried Prozac?

    Also you just happen to be good at the GNU toolchain it’s not to have them at your fingertips instead of having to download and install them, often through some emulator as Cygwin.

    It’s not an emulator. I’m far from convinced that you are “good at the GNU toolchain.” And the whole mess has been superseded by any number of scripting languages you care to name. (I recommend Perl for recovering Bashaholics.)

    At last, but for some reason controversal amongst the M$ fans is the fact that GNU/Linux is Free Software and completely modifiable from the top down.

    Not at all. Nothing wrong with free. Nothing wrong with modifiable.

    Whatever gets the job done. Don’t put words into our mouths.

    I can throw you a rock in any direction and find you a company that does in fact take advantage of the open nature of the source code.

    Throw the damn rock, then. I’m tired of waiting for the other shoe.

    Did I miss that rock? Somehow I think you would have triumphantly repeated the trajectory.

  20. Dr Loser says:

    Have you used Kali on the RasPi, btw?

    Enquiring minds wish to know. You did say that platform independence was important to you, did you not?

  21. Dr Loser says:

    So, how’s that working with an off-beat Linux distro dedicated to network security that was only released in March 2013 going for you, anyway?

    What with you not being the sort of network security type who is prepared either to support his own software, or to download and build the necessaries. Must be tough, considering you need dozens of separate unique packages for your everyday (non-specialist) work. I can sympathise.

    Presumably, prior to March 2013, you were stuck with M$ Windoze. Or was that not the point of your bringing Kali up?

  22. Dr Loser says:

    I don’t understand what you M$ trolls are trying to accomplish.

    Nothing at all, joe. Perhaps we differ markedly in our daily goals, but I tend to accomplish things at work. Whilst I am mildly offended at outrageous claims on the Internet, I am not under the illusion that I will accomplish anything by pointing out the bleedin’ obvious.

    It’s just a hobby and an intellectual exercise, that’s all.

    I envy the people who have more patience then me in dealing with them.

    Alternatively you could, just for once, listen to a developer with thirty years’ experience across a broad range of platforms (all three Big Iron Unixen, Windows, etc). I put far more effort into your silly little gcc cite than you did. Then again, I’ve probably been using gcc for far longer than you have — I even replaced CC 5 on Solaris 2.7 with it, just because it worked to purpose.

    Worked to purpose … worked to purpose … you just don’t get that concept, do you?

    Carry on crusading, O blithe ignorant one!

  23. Joe says:

    I don’t understand what you M$ trolls are trying to accomplish. You aren’t getting very far. I think it’s pretty obvious that we remain unconvinced by your spin, and we are not interested in purchasing your medicore proprietary software. So sorry. I don’t understand why you aren’t peddling it to some people who don’t know any better, you’d probably have more success. Has Micro$oft really got this desperate?

    The unfortunate thing is usually when a M$ fan or salesperson comes into your office (as will definately happen if you run a large company and try to move away from M$ or Oracle software, btw, so prepare for it), you can always show them the door when you tire of this kind of crap. But these people don’t get the hint. I envy the people who have more patience then me in dealing with them.

  24. That Exploit Guy says:

    That Exploit Guy basic mistakes. mount -o remount under linux cannot move the mount point.
    My mistake. That’s “mount -M olddir newdir” as stated in the documentation. Now, how exactly is that supposed to change the fact that the race condition between the wiping software and other processes attempting to acquire the same unallocated block for writing cannot be resolved using your proposed, imaginary solution due to, according to your cited source, that it involves a “a saner way” that is expressly prohibited by the system? Again, don’t get the wrong impression here – at this point there is nothing worthwhile left to discuss on the subject. The fact that your purported mechanism is not even employed by the wiping software that you have linked to is by itself a solid reason that no one should take any of your words seriously. Everything else is just a pointless cat-and-mouse game.

  25. oiaohm says:

    That Exploit Guy basic mistakes. mount -o remount under linux cannot move the mount point. Its not an option when you pass a -o remount flag to give new mounting location. The mounting location is in fact locked in the “mount -o remount” process. The inodes and other information displayed to applications also does not change.

    Remount is a name of special operation part of the gnu development mount and libraries that has been included in Linux. The name is deceptive. Its not in fact remounting on most file systems. remount does not unmount or mount in most cases. If remount did unmount file-handles would be voided going from ro to rw as well. File handles remaining are remaining in existence because it has not been unmount and mounted.

    Read the ebusy thing again. Notice something That Exploit Guy there is no mention of read file open. Only files opened with write flag will for sure send a Ebusy.

    That Exploit Guy and you cannot read. You cannot open a file for write on a read only filesystem ever.

    The ebusy is talking about going from read-write to read-only mount. The condition is that there must be no write files open on the filesystem. Yes you can keep on hitting mount -o remount ro until it does not return ebusy.

    http://forums.gentoo.org/viewtopic-t-915736-start-0.html

    Yes mount -o remount ro is in fact performed in the shutdown of particular distributions. Mostly to draw line in sand so processes cannot open up any more files for writing past that point.

    Read only filehandles can remain open through all remount operations on supported filesystems. Cannot remain open with a unmount. This is why remount is not unmount.

    mount -o remount ro and mount -o remount rw gets more warped than what you think.

    mount -t auto /dev/sda1 /wrong1 -o rw
    mount –bind /wrong1 /wrong2
    mount -o remount ro wrong2

    Interesting enough /wrong1 is read rw and /wrong2 is ro at the same time.

    Locking a device for cleaning you would run a mount -o remount ro /dev/sda1. When you apply mount -o remount rw /dev/sda1 after cleaning to restore to normal. /wrong2 still would be read only.

    What is going on here. Where do you think read only and read write is implemented. VFS layer in fact.

    Funny enough is you can move the mount point under Linux with the mount –move command. Using binding you can also overlay existing mountings.

    There are many ways to skin the cat. Lots of them require you to understand what linux can do.

    The complex way to skin the cat. Is exploit the overlayfs mount. This has been in Linux since version 3.0.0
    https://kernel.googlesource.com/pub/scm/linux/kernel/git/mszeredi/vfs/+/overlayfs.current/Documentation/filesystems/overlayfs.txt

    between move and bind and rbind.

    Overlayfs creates a nice bit of magic. where is overlayed no longer has new opening write file handles ie lowerdir. The lowerdir is the drive you want to take read only. The upperdir gets all the writes and alterations from that point on other than existing handles that are still open.

    It is in fact possible for the directory that was exposed from a drive that is being cleaned to remain read write by using overlayfs to join it to another area of disc that has not had to be turned read only.

    So That Exploit Guy yes its all possible to be cleaning a drive and user going about there business without knowing what is going on. Does user really need to know there files are not being written to X drive but are in fact being written to Y drive because X drive is busy being cleaned.

    Yes the magic of bind mounts and overlayfs mounts and remounts.

    The fact is you can overlayfs mount a location straight away. Wait for users to end processes so over time there will be no write file handles open. Yes what is in overlayfs can be merged back after you are done.

    That Exploit Guy Linux mount operations are extremely flexible.

    There are very short displayed windows of read only protection really required if you know what the heck you are doing with Linux.

    Yes the time to merge the overlayfs upper and lower back into one piece is the only time you have to prevent the user from having read-write access. Yes this is a lot shorter time than running a file-system clean.

    Yes -o remount ro and tell the user this section of disc is read only for a while is simpler than having to merge overlayfs later on.

    So yes That Exploit Guy there is ways around having file handles open for writing while the file system with most of the data on is locked read only. Not the most pleasant. Its simpler to just wait for a window when no filehandles with write are open on a partition.

    If that is not possible then you use overlayfs to create that window. Yes overlayfs comes with annoying price. Way Way less than a freespace wipe program. Freespace wipe program you can be looking at 8 hours+. overlayfs sort out might be 1 hour.

    Using Windows while the huge file is being created also create fragmentation.

    So yes changing from read write to read only is always possible under linux the question is how much framework do you have to set up to enable it. The change is not always click of fingers you might have to wait a few hours before you can start the freespace wipe.

    Users don’t have to see a thing with careful planing.

    That Exploit Guy basically I just described everything that allows you to turn / under Linux read only. Now how do you do that with Windows C:

    There are some areas where Linux is insanely more flexible. Filesystem handling is one of those things. Yes it is possible to overlayfs mount /

  26. Dr Loser says:

    I notice, joe, that you’ve given up on the whole “ineffable magic of the GNU/Linux toolchain” thing.

    At least we’re getting somewhere. Very slowly.

  27. Dr Loser says:

    I’m struggling to understand “DrLoser”‘s retort to the Kali thing.

    It isn’t really that difficult. You quote a little-known distro in support of your thesis that Windows is generally “unpleasant to work with.” I reply, in time-honoured fashion, with a link to Wikipedia, which I think credibly highlights four of the major packages included in the distro.

    There’s nothing special about any of them.

    He pulled out four random packages from their repository (which contains hundreds of netsec packages.

    Not random, just the four quoted in Wikipedia. There may be further magic in Kali, but it would be up to you to highlight it, wouldn’t it?

    I know enough about these four packages to be aware that they represent a good chunk of “netsec” functionality, and that said functionality is also present on Windows. Or on OS/X, for that matter. Or on NetBSD.

    If Kali was just four packages, it would be kind of pointless wouldn’t it?

    A bit of a straw-man argument, isn’t it?

    It’s quite possible for Kali to be kind of pointless in any case. It’s quite possible for it to be invaluable. You provide no evidence either way.

    All I’m saying is that using a corner-case distro to make unsupported and unsupportable arguments about Windows’ “bloat” or “lack of package management” or “failure to provide tools” reeks of desperation. You do realise that exactly the same argument would apply to Debian, don’t you?

    (I’m not making that argument; it would be silly. I’m simply applying a reduction ad absurdum.)

  28. That Exploit Guy says:

    That Exploit Guy the reality here is you don’t understand how Linux remount it working.
    At this point I am not entirely sure there is still any reason to reply to your obvious horsecrap, but I’ll bite anyway…
    Let’s start by revisiting your assumptions, shall we? You assume:
    1) That at some unspecified stage the Linux kernel changes the mounted root file system from read-only to read/writable without unmounting.
    2) That this supposed mechanism is applicable to your hypothetical free-space wiping software.
    For Point 1), I have already helped you clarified the “unspecified” mechanism by pointing out exactly what it is (i.e. remounting). Whether it involves atomically unmounting and mounting the file system (which “-o remount” must do if it involves moving the mount point elsewhere) or just changes the flag does not matter – it’s just an implementation detail (since it has to be one single move from other processes’ perspective either way). All we need to know here is that it involves remounting the file system.
    Regardless, this still
    won’t help you in anyway support the second point. Why? That’s because your proposed mechanism demands switching a file system from read/writable to read-only. In fact, your entire argument is hinged on the procedure being possible. As far as what you have cited to support your argument is concerned, when there are opened file descriptors (for writing, specifically), the OS can atomically remount the file system only from read-only to read/writable, not the other way around. With this glaring contradiction in mind, is there supposed to be any reason for anyone to give a toss about what you have to say? I think not.

  29. oiaohm says:

    That Exploit Guy there is a big difference between deceptive and lieing.

    What I wrote was deceptive it suggested the possibility of read write. I never stated read write. So yes getting up me for being slightly deceptive would have been valid. Slightly deceptive is not lieing. Bad things don’t happen using the Linux method. Either you can start the secure wipe and its safe or your cannot under LInux.

    Windows you can start the secure wipe but its not ensured to be safe.

  30. oiaohm says:

    That Exploit Guy basically give I can just throw more and more pages in the manuals about remount at you. Right down to the Linux documentation for writing filesystem drivers.

    The reality it is valid under Linux to change from read-write to read only without closing all files. Its also perfectly valid to change back the other way. It is important to know this when you do have to perform maintenance.

    There are a lot of places where Linux does not follow the posix standard but follow other standards of other Unix’s. Remount handling is one of those areas.

    That Exploit Guy wordpad is not a office suite because you don’t have excel or powerpoint like items. A suite demands a collection of tools. LibreOffice like or not is a Office suite.

    MS trolls are very good at abusing english.

  31. oiaohm says:

    That Exploit Guy the remount changes the flags controlling a file system. It does not nuke the currently open file-handles unless those will be conflict with the change or the file system is not able to perform the change.

    Remount is in fact a specialist command give to the file-system driver under Linux That Exploit Guy and solaria s by the way.

    That Exploit Guy the reality here is you don’t understand how Linux remount it working.

    EBUSY
    The device is already mounted.
    The mount point is busy. (E.g. it is some process’ working directory or has a filesystem mounted on it already).
    The request is to remount read-only, but there are files open for write.

    http://www.gnu.org/software/libc/manual/html_node/Mount_002dUnmount_002dRemount.html
    It is not unspecified behaviour. Yes you get a ebusy if the filesystem does not support remount command and unmount and mount will be performed and there are open files.

    GNU libc was the first to define the means to swap from read-write to read-only without closing all files. Read-write files have to be closed. This comes for the first Unix’s that GNU Libc was designed to support.

    Also your stage in the init is also wrong. The change by remount to read-write happens after the fsck stage is past of the init process. Yes the fstab is in fact read twice. It is not remounted straight away.

    That Exploit Guy I did not say remount but that was really my only error. Remount under LInux does not equal close all files. Like remount with change some other flags other than read-write can leave all files open. Like adjusting journaling options.

    Linux remount and a lot of the Unix remounts are in fact just how to change settings on a file system most of the time.

    That Exploit Guy problem is I am out right lieing. You are accusing me of out right lieing because you don’t understand how Linux remount and most Unix remounts work most of the time. Ext2 to Ext4 only give busy on change to read-only from read-write by remount if a read-write file is open.

    So this is not some magical function. In fact this behaviour of remount is very hard to have work in BSD.

    Linux you can also use remount to flick on and off noexec flag on filesystem. That is intersting any executable bit running from the file-system has to be closed when turning noexec on or it returns ebusy.

    There are some very interesting features to the Linux/GNU remount.

  32. That Exploit Guy says:

    It comes with a subpar web browser
    Which I suppose is the reason that I have come up with more links to support my arguments than you have ever supported yours.
    no office suite (unless WordPad is an “office suite”)
    Which I think is pretty on par with LibreOffice or whatever you consider an “office suite”.
    maybe some crappy games and various random utilities.
    Then perhaps “fsck” is about as random “chkdsk”, by your definition.
    C’mon, why stop there? How about “fdisk” vs. “diskpart”? Or “ln” vs. “mklink”? After all, the reason we are here is so we can share and compare this kind of arbitrary metric of usefulness till the cows come home, isn’t it?
    To make it useful you actually have to install a whole bunch of random crap, which isn’t easy to begin with because Windows doesn’t have a package manager
    What would be your precise definition of a “package manager”? Does this constitute a “package manager”? Isn’t it kind of a shame that that was the same neat idea that got rejected by Debian but taken up wholesale by, I dunno, Windows (c.f. WinSxS)? Have fun figuring out whether you should symlink libblah.so to libblah-0.1.0 or libblah-0.1.1 in /lib, though. It’s good brain exercise.
    Windows is just not all that, even on a purely technical level. It’s merit comes mostly from the fact that it’s popular, therefore has a large ecosystem. The OS itself is just years of random legacy crap piled miles high, hence the huge footprint.
    I like unqualified hand-waving. They are so cheap to make you can fill pages with them without even batting an eye.
    Here, you have made three assertions:
    1) That Windows has had no merit aside from popularity.
    2) What OS has accumulated for years is “legacy crap”.
    3) That the “legacy crap” is random.
    Immediately, it’s obvious that 1) is begging the question as to how Windows has become popular to begin with (don’t look at me – Joel raised the point first). 2) and 3) are related to the first and suffers the same problem that one will never see when it’s viewed from your angle. Of course, why attempt to look at something from a broader perspective when you can dwell in this “my mile is longer than your kilometre” sort of nonsensical, and, quite frankly, self-absorbed arguments?
    Yes, most of us here have used Windows.
    Here’s news flash – I have been on your side as a Linux advocate. I also have been a youngster just out of teenage years and thinks he knows everything, and, in my book, you are about as typical as a Linux advocate can get.
    Also, “we understand Windows”? You mean as in “we don’t know changing the shell in Windows involves nothing more than modifying a couple of documented settings”? Seriously, you have got to be kidding me. You have been a novice user of Windows. You don’t know much of the OS but just enough to get you by (see also: office workers, grocery check-out clerks). You think you are a pro but the fact is, rather, that the OS has made things easy enough for you to not need to think of what’s going on in its internals when you do something. Your “understanding” of the subject matter is therefore not worth considering, no matter how strong you believe otherwise.
    I don’t want to bother collecting and maintaining all these tools myself when there as an entire community of excellent security practitioners doing it.
    I am not sure you are aware, but at this point you have already shifted the goalposts from “difficult to emulate” to “hard coded” Linux dependencies and then to mere appealing to (ill-defined) authority (c.f. “excellent security practitioners”). Is there an actual point you are trying to make here, or is it too hard for you to admit that you don’t know what people have been using WinPE for?
    The way we work with GNU/Linux is far deeper then what is possible with Windows because GNU/Linux is Free Software that is free to modify and free to distribute.
    Your knowledge in Windows has already proved to be barely below skin, so is this the kind of canned statements that you think will impress anyone?
    Kid…

  33. ram says:

    That is a pretty feeble machine for HP to be calling a “workstation” regardless of OS.

  34. That Exploit Guy says:

    Have you never follows a Linux boot and shutdown. On boot filesystem are first mounted read only with files open using remount changed to read write.
    Here you have asserted two things:
    1) That at some unspecified stage the Linux kernel changes the mounted root file system from read-only to read/writable without unmounting.
    2) That this supposed mechanism is applicable to your hypothetical free-space wiping software.
    It’s kind of a shame that Wikipedia does not help clarify that “unspecified” part, does it? Fear not – I’ll fill in the gap for you – once “init” has been loaded, /etc/fstab is read, and then the root file system is remounted accordingly. You don’t believe me? Here’s how OpenSuSE does it.
    And the second point… Seriously, is there anything worthwhile to left be discussed in this so-called “conversation”? I have already exposed you for outright lying about how a particular piece of wiping software works in actuality, and now you want to save face by distracting everyone from your dishonesty with more dishonest yarn? I am sorry, but have you no shame?

  35. dougman wrote, “what is it in Windows, that is such a convoluted mess”

    The answer to that is as complex as the software. It all stems from M$’s number one priority being to make outrageous profits. To do that they have to screw all competitors and lock in users. That’s a heavy burden to put on every programmer in the place. Every extraneous API, restriction in EULA, obfuscation of code, restriction in performance etc. are designed by salesmen to lock in the whole world of IT, not to get the job of computing done by the most effective means. Fortunately for us, that complexity prevents M$ from releasing two reasonable releases in a row or having a reasonable upgrading/updating process and giving us half a decade between reasonable releases. So, the world has innovated and worked around the problems M$ created by using FLOSS on ARM or Intel/AMD. For me, even trying to understand the EULA and to comply with it within a budget was impossible whereas I could bring world-class IT to the whole LAN for $0 on the budget. For that low price I got wonderful software that actually did what it was supposed to do: create, find, modify and distribute information. The world doesn’t need to slave for M$ for free when the world can have software like Debian GNU/Linux and all the data of the web for (nearly) free instead.

  36. oiaohm says:

    Joe and dougman

    https://www.google.com/patents/WO2004034626A3?cl=en
    Most Linux distributions use LZO but as you can see by link its patented. Its one of those items with a free license for open source products.

    http://bashitout.com/2009/08/30/Linux-Compression-Comparison-GZIP-vs-BZIP2-vs-LZMA-vs-ZIP-vs-Compress.html

    Windows is in fact still using compress in the form of MSZIP compression. And LZO does better than LZMA on binaries.

    So Windows is over 50 larger download than what it should be. Due to poor compression to start off with. Just to add insult. Windows does support LZX yet default packaging from Microsoft is either uncompressed or mszip.

    Using LZX would have saved at least 20 percent.

    Fairly much if Microsoft could use there own packaging correctly it would be smaller. Or better Microsoft bites the bullet accepts MSI and CAB tools need to be open source so they can use LZO and other very effective compressions.

    Microsoft has a hand tied behind there back for online provide. Crappy compression.

    Yes android packaging is gzip, bzip2 or lzo. All better at compressing than lzx or mszip.

    Yes the fact that a section of that 3.5 GB is uncompressed when its compress-able is down right insulting.

    Yes to interface with web sites Microsoft has to have gzip equal installed anyhow. Yet they don’t support it in their package management.

  37. Joe wrote, “We don’t use GNU/Linux because “we don’t understand Windows”. We understand it, and we don’t like it. It’s horrible.”

    Amen! The difference between not liking something and finding it really horrible is the total lack of redeeming features. There is absolutely nothing I like about it and a long list of things that I hate: US DOJ v M$/Comes v M$, BSODs, re-re-reboots, malware, slowing down, the price, and of course, monopoly (now going away, thank Goodness). That’s the short version…

  38. Joe says:

    It’s funny because despite being many times larger than an average Linux distro, a fresh Windows install is nearly useless. It comes with a subpar web browser, no office suite (unless WordPad is an “office suite”), maybe some crappy games and various random utilities. It’s not even secure by default, you have to be batshit insane to use Windows without a full blown anti-virus and security suite.

    To make it useful you actually have to install a whole bunch of random crap, which isn’t easy to begin with because Windows doesn’t have a package manager.

    Windows is just not all that, even on a purely technical level. It’s merit comes mostly from the fact that it’s popular, therefore has a large ecosystem. The OS itself is just years of random legacy crap piled miles high, hence the huge footprint.

    Yes, most of us here have used Windows. We don’t use GNU/Linux because “we don’t understand Windows”. We understand it, and we don’t like it. It’s horrible. GNU/Linux is just overall, far more pleasant to work with. We don’t have to tie ourselves forever to the licensing and business whims of a sole proprietary vendor either. The way we work with GNU/Linux is far deeper then what is possible with Windows because GNU/Linux is Free Software that is free to modify and free to distribute. This is important to me, and until Windows can say the same it will never be very appealing to me.

    I’m sorry that apparently this fact makes you so angry you M$ fans actively have to seek Linux blogs to troll on, but it’s not surprising. As Ghandi said “First they ignore you, then they laugh at you, then they fight you, then you win.”

  39. dougman says:

    Ok, so lets get this straight….3.5GB for a full-blown Windows operating system? The last Linux ISO I downloaded and installed was only <1GB (915MB) and you still get ALL the drivers you need. Also, with that said, it only takes 5 mins to install then 5 mins to apt-get update/upgrade.

    So, the question is, what is it in Windows, that is such a convoluted mess??

    3-5 hours to install/upgrade an operating system is rather stupid and a total waste of time, it seems you Windows folk do not value your time and enjoy wasting your life.

  40. oiaohm says:

    That Exploit Guy 3.5 GB does not include all features that you can install in Windows 8.1 that is the average download that will happen based on the common level of features enabled. Your presume is off. Some people are finding this out they go into windows 8.1 and go to enable some extra features then it goes and has to download other bits from the Microsoft store.

    That Exploit Guy the sad part is the 3.5 GB is poorly compressed. There are many times better compression but Microsoft would have to pay patent licenses or release code open source to use them. Do Linux distributions have this problem no their package management is open source.

    Patents are hindering Microsoft quite badly.

  41. That Exploit Guy says:

    He pulled out four random packages from their repository (which contains hundreds of netsec packages. If Kali was just four packages, it would be kind of pointless wouldn’t it?)
    I suppose it’s not his responsibility to come up with a purpose of or any sort of minutiae about the distro for you, is it? After all, “Kali Linux” is something that you have brought up to support whatever argument that you think it is supporting, so don’t you think it is your responsibility to do that bit of homework yourself? Also, let you give you a hint: have you noticed that Kali Linux is not just meant for network pen-test? So much for regurgitating someone else’s talking point from some forum…

  42. oiaohm says:

    That Exploit Guy Linux remount is interesting enough more selective than you are giving it credit.

    –So mounted read write can be switched to read only while doing secure wipe
    No. “-o remount” implies unmounting and mounting the file system again, and you can’t do that when there are files opened from within the file system (“mount” will just tell you the device is busy). The tool you link to simply exits when it encounters a device that is mounted for read-write (“rw”), and it’s not hard to see why that’s the case.
    In fact —

    Have you never follows a Linux boot and shutdown. On boot filesystem are first mounted read only with files open using remount changed to read write. The reverse is also true and is used in some Linux systems shut-downs.

    A remount does not destroy the already loaded items with Linux kernel. Cannot do it with files open who are you trying to kid That Exploit Guy.

    That Exploit Guy the limitation that any file open on a remount could trigger remount failure is a posix define. Linux read only files open don’t trigger when the kernel is Linux. Same with many other Unix kernels.

    Remount is a remount is not a mount or a unmount. Its a unique operation to its own right. The min posix requirement is that a remount is implemented as unmount followed by mount. But in fact this is not required to happen and does not happen with most Linux filesystems.

    This is the problem you know posix min requirements but not how the world is implemented.

    Dr Loser
    –Whoops, it’s in an unstable state, somewhere between full existence, partial existence, and nonexistence. I’m sure my reads will be really, really stable. Not.–

    In fact the disruption is identical to another application using full disc io. Swaping.

    — This best method would require alterations in vfs in Linux to allow sections of discs to be temporary claimed without ever having to enter that claim to disc.

    No, that’s about the worst method imaginable, although I give you props for your imagination. First of all it doesn’t exist (for a very good reason). And secondly it completely defeats the purpose of having a filesystem in the first place. —

    Dr Loser Linux delete files don’t instantly disappear while a process has them open either. There are already ghosts. I was not going mean enough to mention Linux these days does in fact support a claimed only in memory for temporary files. Yes the blocks left marked as free in the filesystem even that its being used for a temporary file.

    So yes you can implement a secure wipe tool on a mainline Linux kernel using a rolling method. Distrobution used kernels most don’t have the temp-file without claiming the space on disc option yet because there kernel is not new enough. The option of creating temp files without writing claimed blocks into the on media storage extends the life of SSD drives by reducing writes. It removes 2 writes. First to create tempfiles second to delete them. Remember SSD have a limited write count before sectors start failing.

    The fact that the feature is implemented is in fact implemented in current Linux vfs shows how little you know Dr Loser. Yes you have just came out and claimed a option does not exist when it does.

    That Exploit Guy and you Dr Loser. Really don’t know Linux.

  43. That Exploit Guy says:

    Windows IS a bloated PIG, I mean 3.5GB for an upgrade and one still does not get all the drivers with it?
    The more pertinent question would be whether you could fit all device drivers for Windows into 3.5GB, don’t you think?
    Also, the 3.5GB download size would include all features present in Windows 8.1, as this would be in line with how Windows handles component enabling/disabling as since Vista. In other words, that download presumably is the full image of the operating system – and I shall leave this and this here for a bit of “contrast”.

  44. Joe says:

    I’m struggling to understand “DrLoser”‘s retort to the Kali thing. He pulled out four random packages from their repository (which contains hundreds of netsec packages. If Kali was just four packages, it would be kind of pointless wouldn’t it?) He showed that one of those four can be installed on Windows, which isn’t surprising because not everything in Kali or any Linux distro for that matter is Linux only. He then decided to throw a couple of insults at the packages themselves. Great I guess, but irrelevent. Then he threw a couple insults at me (par for the course), and ended with something that can be summarized as “victory is mine, haha loser [this is not ironic considering my name]”. His alleged victory is apparently to show that Kali has no merit over Windows.

    But all he showed is that you can install one of the tools included in Kali in Windows. That was never part of the question, as you can install many of the tools included in Kali on Windows. The question is why would you? I don’t want to bother collecting and maintaining all these tools myself when there as an entire community of excellent security practitioners doing it. They are working together to produce a domain specific OS that does all the legwork for you – you just work on doing your job.

    I’m not sure if this guy just trolling or is he for real. Like anything it’s really hard to tell these days, especially over the Internet.

  45. That Exploit Guy says:

    It’s great that you can sortof customize Windows by hacking up system files in unsupported ways
    Obviously a youngster I am talking to.
    What you refer to as “hacking up system files” is nothing more than a one-line change in the system settings. In the old days of Win9x, this would be “Shell=…” in System.ini. This was exactly how the series of OSs provided the choice between Start Menu and Win 3.x style Program Manager. Nowadays, newer NT-based systems use the registry equivalent of ini file and the appropriate settings are as documented here.
    By the way, how would you define a “supported” way of changing something in Linux?
    Mind you, they are rarely available as nice little MSI files.. often they are distributed in source code, hard coded to POSIX or even Linux.
    As Dr. Loser has already, these “rarely available” tools are easily obtainable for the Windows platform. The Aircrack-ng website gives you a link to a Windows binary package right on the front page, Wireshark is as popular as ever on the platform, and nmap has already had a Windows version since 2000. What is this complaint of yours about “rarely available” tools again?
    I can tell you what I lost though: access to a lightweight, secure, multi-architecture OS.
    Lightweight: a metric that serves no purpose in a pen-test. If you want a stripped down platform that boots from a USB stick and runs your applications, you have already got WinPE for that purpose.
    Secure: grsecurity developer Brad Spengler uses Windows 7 for his work. That should tell you a lot about that typical talking point of your.
    Multi-architecture: an argument that flies in the face of your notion of software hard-coded for specific platforms, for obvious reasons; also completely useless for pen-tests.
    To save my time and your, let me ask you question: do you have a point that is actually worth considering?

  46. dougman says:

    Loser, demands that one ‘shut-up’ and not mention fact about Windows, but fails to realize this is a public blog.

    Windows IS a bloated PIG, I mean 3.5GB for an upgrade and one still does not get all the drivers with it?

    http://www.pcworld.com/article/2056292/windows-8-1-launch-weekend-plagued-by-some-show-stopping-installation-issues.html

    “Windows code is simply too cumbersome. It’s too subject to corruption and when things go wrong, they go really wrong.

    Who builds and releases an upgrade without putting in some type of failsafe that rolls the system back automatically if something goes wrong? Also, doesn’t this upgrade CHECK the hardware and drivers first to ensure compatibility?

    What a train wreck.”

    “This is totally reprehensible. I’m sick of Microsoft using its customers to beta test their releases.”

    “….took almost 5 hours for the update to install on my Windows 8 machine.”

    M$ is doooomed.

  47. Dr Loser says:

    It’s the novice hour here, isn’t it?
    Every hour here is novice hour.

    Haven’t you noticed?

  48. Dr Loser says:

    I plead guilty through ignorance to those two charges, Robert. (The handbook was delivered CoD, and sadly my troll account didn’t cover that.)

    Those are Rule 12 and 19 of the troll’s handbook, issued by M$.

    What are the other seventeen or so? Let your imagination run riot!

    After all, unlike joe, you can count up to a hundred.

  49. Dr Loser wrote, “do you really need “multi-architecture?” First time that little nugget has popped up.”

    That’s been around a long time. It’s one of Debian’s strengths. Device-independence also gives independence from Intel and M$, a good thing in my book. Besides less expensive software, multi-architecture permits using less expensive hardware, also a good thing. If AMD, ARM and MIPS weren’t around, how much higher do you expect Intel chips to be? 300%? 200%? Those are the kind of premiums Intel charges for newer devices. With no competition they would have no need to drop those prices.

  50. Joe wrote, ” Oh and I shouldn’t be using GNU/Linux, because they said so. But nothing about why Windows is any good, or why Linux isn’t. “

    Those are Rule 12 and 19 of the troll’s handbook, issued by M$.

  51. Dr Loser says:

    cipher program on windows can in fact render the computer unbootable That Exploit Guy due to the huge motherfile. Why its a real written file to disc.

    Random idiocy quoted as a chance for all of us to have a good laugh before continuing.

  52. Dr Loser says:

    Obviously things like “arguing facts” is a lot harder than trolling and throwing tamper tantrums…

    Which temper tantrums would those be, Dougie? I see no temper tantrums on either side. Your evidence here being?

    As for facts, I’ve repeatedly asked joe for them. The best he could come up with is an (irrelevant) link to the git repository for gcc. Now, I happen to think this is irrelevant to the point in question, but I’m always happy to pursue “facts” wherever they lead me — thus my quantification of those “facts.”

    I don’t expect you to rebut that quantification, because you depend upon selling second-rate Windows anti-virus kits for your livelihood. I was assuming that, out of those 57 or so “facts,” joe might be able to pounce on one.

    After all, despite his distaste for people on the other end of the phone from third-world countries/cities that he might or might not be able to guess the first letter of, and I’m absolutely certain (it’s a fact!) that he doesn’t despise them just because their English is obviously better than his, the man knows his facts.

    As evidenced by this excellent and completely irreplaceable minority Linux distro he references.

    Tell me, Dougie, have you experienced Kali?

    Are you experienced?

    If not, then don’t butt in to a private conversation. You don’t have a grasp on the “facts.”

    You never do, do you?

  53. Dr Loser says:

    I’m a Pinko-Nazi-Communist-Racist-Jew-Idiot who is also literally Hitler.

    Sadly, no, Dougie. You’d be far more interesting if you even had an iota of those qualities.

    You’re just a tedious uneducated slob who sells snake-oil for a living and whines a lot.

    Hey, sometimes TV is short on proper comedy. It’s always nice to be able to depend upon sincere ignorance and cupidity, and where else would we be able to get it?

  54. Dr Loser says:

    Oh, and Joe?

    That “multiple invoices from Microsoft” thing. I assume you have evidence to back this up?

    After all, you can get free software to do all of that (apart from your rather peculiar choice of windows manager), so I beg leave to doubt the invoices.

    But I’m sure you tried, and you have proof. Just like you had proof for all your other assertions.

  55. Joe says:

    dougman,

    Funny stuff! All I’ve learned so far is that I’m according to these M$ fans (so it’s true, apparently they exist!) I’m a Pinko-Nazi-Communist-Racist-Jew-Idiot who is also literally Hitler. Oh and I shouldn’t be using GNU/Linux, because they said so. But nothing about why Windows is any good, or why Linux isn’t. Obviously things like “arguing facts” is a lot harder than trolling and throwing tamper tantrums for these people. It’s also interesting to see how they seem to just come out of the woodwork and mysteriously stumble on a GNU/Linux blog.

  56. Dr Loser says:

    Nice diversion, Joe. I assume you’ve conceded on the various points of throwing rocks, counting from one to a hundred, the potential value of being able to alter to source as it pertains to SMEs, and let’s just leave the casual “thick accents” to one side, shall we?

    Given that, I appreciate your honesty. It’s down to personal requirements with Kali, isn’t it?

    And there’s nothing wrong with that. As long as you stop pretending that anybody other than yourself and possibly other people in your field care. Because, you know, none of us could care less.

    However.

    Care to share (besides gaining a bunch of invoices from Microsoft)? I can tell you what I lost though: access to a lightweight, secure, multi-architecture OS. There is simply no reason to even consider Windows in this use case.

    “Lightweight” is irrelevant. “Secure” is, unless you can prove otherwise, an entirely personal judgement. And do you really need “multi-architecture?” First time that little nugget has popped up.

    But heck, let’s review what Kali has to offer.

    (1) nmap
    (2) Wireshark
    (3) John the Ripper (I so love these cute names, particularly when they reference serial killers of prostitutes)
    (4) Aircrack-ng

    Do you seriously believe that this is functionality (albeit without your non-headache-inducing windows manager of preference) that is unavailable on other platforms?

    If so, then you’re more ignorant than I thought.

    Take number 4. It’s available on Windows. As a matter of fact, it’s available on every other Linux platform out there.

    What, specifically, is so special about this Magic Distro of yours? Nothing, that’s what.

    Now, try again. Throw one more rock and hit a number between 1 and 100. I’ll abide by the rules. Find a single update to gcc, off your cite, that has the faintest relevance to an SME.

    You fail, mate. You fail horribly.

  57. Joe says:

    Yeah sure, write my own Windows equivalent to Kali Linux, except I’d have to install hundreds of different netsec tools. And keep tabs on them and maintain them when they are updated. Mind you, they are rarely available as nice little MSI files.. often they are distributed in source code, hard coded to POSIX or even Linux. I’m sure that’s just a make/make install away from working on Windows? So after all this effort, what would I have? Did I gain anything? Care to share (besides gaining a bunch of invoices from Microsoft)? I can tell you what I lost though: access to a lightweight, secure, multi-architecture OS. There is simply no reason to even consider Windows in this use case.

    Also, linking to a bunch of random “Windows shells” is not really an answer to awesomewm. It’s great that you can sortof customize Windows by hacking up system files in unsupported ways, but I rather not. They are usually half-assed hobby projects meant for looking uberl33t then any actual usability.

  58. dougman says:

    Certificatons to support Windows? LOL…. All the so-called ADMINS and IT Idiots that boast Windows are just in it for the money, and could care less abour real computer work.

    3-Hours to restore Windows 7? Wow….what a bloated mess. I just redid my / partition in 5 mins and apt-get upgrade took another 5 mins, for a sum total of 10-mins. Windows is a nightmare!!

  59. Dr Loser says:

    Dr Loser in fact do tech support for Windows operating systems.

    No you don’t, oiaohm. Cut it with the “VAR Guy” nonsense already. You know nothing at all about Windows systems, and the best you can do is offer first-level tech-desk support for changing passwords and such.

    Of course, you can prove me wrong by quoting your certifications.

    ==Which Windows “issues” would those be, oiaohm? ==

    Name one really there is more than 1.

    You sort of fail on analytical logic, don’t you? Proposition, counter-position, resolution.

    What you fail to understand (and there are so many things that you do not understand) is that it is up to me to question your thesis (“see here, I can quote you a Windows issue”) yet it is up to you to quote an example that supports your claim.

    Now, as it happens, this is quite a simple thing to do. I will help you out by offering an anecdotal personal example. Today I had to restore my Windows 7 machine from the backup partition, followed by updating the entire thing through 116 (to memory) updates. Took me about three hours, although I enjoyed watching Northern Exposure whilst it happened, and luckily I was stocked up on good coffee.

    See? That’s honesty.

    You wouldn’t recognise honesty if you tripped over it, would you?

  60. Dr Loser says:

    Without awesomewm, I find using a computer frustrating as hell.

    Don’t you think that’s a rather embarrassing admission? I guarantee you that 99% of Linux users have never heard of the thing. I’d lay odds that Mr Pogson has never even tried it.

    There are probably about three billion computer users in the world (I’m going with the flow here and accepting mobile phones and tablets as computers for the purpose of this discussion).

    If there’s more than a couple of hundred who are so mentally constipated that they need something with a cute little name like “awesomewm” to be present before they can actually concentrate on the job, I’d be very surprised.

    Nice to meet one of those couple of hundred, on the other hand. Congratulations: you are a very special person.

    Shame about all those third-world people with slightly different accents you have to rely on for support, though.

  61. Dr Loser says:

    Also you just happen to be good at the GNU toolchain it’s not to have them at your fingertips instead of having to download and install them, often through some emulator as Cygwin.

    Yes, joe, I “just happen to be good at the GNU toolchain.” (I’d describe it as the *nix toolchain, but freedom is freedom, I guess.) I’ve been using it for, let me see, seventeen years. Typically on Solaris, sometimes on Windows. And Cygwin doesn’t “emulate” it, btw: it’s a port. I can explain the difference to you if you’re still confused.

    find and grep are part of my everyday arsenal of tools. When I need to import some fatuous flat file from a *nix system, I use cut and awk and sort and uniq. If I want to write a “little language” I can use flex and bison (although I strongly recommend you don’t use the hairy version) … these two, and indeed quite a lot of the rest, are far less useful now that we have actual proper parseable scripting languages such as Perl and Python, but what the heck, struggle on with the hacky drivel you’re used to.

    I’m not presuming to guess what skills you have, joe, although you’re not exactly showcasing them, are you? So why presume to guess what skills I have? Indeed, why presume to guess which platform I prefer? (Clue: try it and I absolutely guarantee that you’ll guess wrong.)

  62. Dr Loser says:

    I see we’re parading our knowledge of filesystems now, oiaohm. You seem to have accumulated an enormous amount of opinions on the subject: almost all of them incorrect. So let’s start with your first paragraph.

    You have to take into account that while mounted you can switch a filesystem between read only and read write on Linux and most Unix’s. So mounted read write can be switched to read only while doing secure wipe. So users still can be using the machine and access the data without risking disrupting the process.

    Let’s say I have a file open for read/write. Let’s say my sysadmin umounts/remounts the partition as read-only.

    I submit that this would be quite disruptive to my ability to write to that file.

    But, fair enough, the purpose of this *nix-fu is to allow my sysadmin to “secure wipe” the file. Splendid. Let’s say I have the file open for reading.

    Whoops, it’s in an unstable state, somewhere between full existence, partial existence, and nonexistence. I’m sure my reads will be really, really stable. Not.

    Now let’s consider “secure wipe.” Is this going to defeat the CIA or NSA or anybody? Sadly, not. It’s a worthless hack, although one befitting the typical *nix filesystem (not ZFS or anything modern, though) which makes a habit of lying to you.

    As per usual for Linux, it’s just prestidigitation. Not something you could make use of in any real-world scenario.

    This best method would require alterations in vfs in Linux to allow sections of discs to be temporary claimed without ever having to enter that claim to disc.

    No, that’s about the worst method imaginable, although I give you props for your imagination. First of all it doesn’t exist (for a very good reason). And secondly it completely defeats the purpose of having a filesystem in the first place.

    Now, settle down, breathe deeply, and propound one thought at a time. And bear in mind that That Exploit Guy knows far, far more about filesystems than you are ever likely to Google.

  63. That Exploit Guy says:

    You have to take into account that while mounted you can switch a filesystem between read only and read write on Linux and most Unix’s.
    By remounting it. In Linux, the option is “-o remount” if you want to do that with the “mount” command. It’s the novice hour here, isn’t it?
    So mounted read write can be switched to read only while doing secure wipe
    No. “-o remount” implies unmounting and mounting the file system again, and you can’t do that when there are files opened from within the file system (“mount” will just tell you the device is busy). The tool you link to simply exits when it encounters a device that is mounted for read-write (“rw”), and it’s not hard to see why that’s the case.
    In fact the saner thing is flick the filesystem to read only while doing secure wipe than creating a huge file.
    The mechanism that you describe is unlikely to be implementable and certainly does not exist in Linux, for the obvious reason that by doing so you will end up disrupting all processes currently accessing the file system. Don’t you feel tired writing all these lies just to justify a factually invalid point again? Because I am sure as heck sick of reading them.

  64. oiaohm says:

    That Exploit Guy
    –The “one huge mother of a file” is needed for resolving a race condition that occurs when the wiping software and other processes attempt to acquire the same unallocated block at the same time.–

    Really is this the only race condition you have to avoid. Answer is no its not.

    Exactly what is an application going todo if it magically cannot write a file that it was able to open read-write in the first place. This is the second race condition. Filling the media causes a race condition of applications failing because they cannot write to disc.

    Linux can change a mount from read write to read only while OS is running. It will not allow this change while any application has a read write file open. This is resolving the second race condition at the start of the wipe process. Changing the file-system to read only also resolves the reason for creating the huge mother file in the first place as well. Due to the face it removes applications from being able to acquire free blocks on the file-system.

    Users can still be using a Linux system with file-systems on disc switched to read only. If Linux could not operate this way livecds would not work as well as they do. This is another feature of Linux over mounting. So Linux can on fly change from write to disc media OS to a write to ram OS and back.

    cipher program on windows can in fact render the computer unbootable That Exploit Guy due to the huge motherfile. Why its a real written file to disc.

    Ideal for wiping local file-systems freespace you need a means to claim disc blocks on disc that are only recorded as claimed in ram so prevent other applications going after them. But will return to unclaimed when the OS reboots. Also allow freeing them back as you are done.

    Most cases you are needing to do a secure wipe as a Admin is because people have deleted files normally and forgot to use a secure wipe tool. That Exploit Guy, So you are secure wiping the freespace because that is where the non correctly deleted data is.

  65. oiaohm says:

    You have to take into account that while mounted you can switch a filesystem between read only and read write on Linux and most Unix’s. So mounted read write can be switched to read only while doing secure wipe. So users still can be using the machine and access the data without risking disrupting the process.

    In fact the saner thing is flick the filesystem to read only while doing secure wipe than creating a huge file.

    What is the risk with the cipher program under windows is an application attempts to write to disc only to find out that its now full due to what the cipher program has done filling the disc. The application was able to open a read-write file on the media.

    The huge mother file solution is in fact unstable. That Exploit Guy

    Linux way is safe applications that have opened files read-write will block switching to read only from read write mount.

    Yes you resolved the applications writing to the partition before you can start with the Linux solution.

    The 1 huge file method is unsafe. Due to all free-space being consumed at some point with programs with read-write file operations in play.

    That Exploit Guy the race condition is solved in such a simple way. Take owner ship of block perform operation on it then return to to free. Ok this generates a huge stack of create and destroy file messes. Not worth the head ache if you are writing these filesystem changes to disc.

    This best method would require alterations in vfs in Linux to allow sections of discs to be temporary claimed without ever having to enter that claim to disc.

    That Exploit Guy the huge mother file does resolve the race condition then creates another one. Due to the fact its really being written to disc as an existing file it creating excess wear and tear as well.

    That Exploit Guy are you also not aware that the windows cipher program can render a windows machine non bootable. Power outage just as cipher huge motherfile gets to max size. Result no free space to create temp files(like backups of registry) on boot up. This would not happen if the motherfile was a virtual item never written to disc.

    Yes That Exploit Guy I said users could be using the machine while the wipe was going on not that they could write the the partition currently being cleaned.

    Read my words fair more carefully in future.

    That Exploit Guy yes the easy way out is safe. No stupid surprises. No risk the OS will be rendered not boot-able.

    Yes I would like vfs altered to allow rolling claiming of free-space for clearing. So that a clean tools claimed file only exists as long as the OS is running.

    Huge motherfile was never required its just a huge unsafe hack. LInux read only mount is a hack but is a safe hack.

    Really yes this is a area both Windows and Linux can do better. But Windows is the worst here.

  66. That Exploit Guy says:

    Ok secure deletion of files while users are using machine. http://wipefreespace.sourceforge.net/ simple for Linux on the most common filesystems.
    Windows still depends on fill up all freespace with one huge mother of a file.

    Right off the bat, you have already confused the difference between securely wiping a file and securely wiping free space (the software tool you link to is for the latter).
    The “one huge mother of a file” is needed for resolving a race condition that occurs when the wiping software and other processes attempt to acquire the same unallocated block at the same time. The particular tool that you use as an alleged example of Linux overcoming this simply takes the easy way out by refusing to work when the target file system is mounted as read-write (see wfs_e234_chk_mount() in source file “wfs_ext234.c”).
    Is there anything at all you actually know about, oiaohm?

  67. That Exploit Guy says:

    I suggest you slow down a bit with the rapid-fire insults and try to add something productive to the conservation.
    It’s not exactly other people’s fault that you happen to have blurted out whatever ugly little thoughts that have been floating around in your head, is it?
    Also, productive? Sure, then, let’s examine how “productive” you have been since the beginning of this “conversation”:
    Linux has a lot of specialized distros that are difficult to emulate in Windows.
    Kali Linux for one

    Difficult for those not knowing how to build WinPE, that is.
    It’s not exactly an engineering feat to load some command line tools onto a stripped down environment, after all.
    There is also window managers that are only available really on GNU/Linux, for instance awesome. Without awesomewm, I find using a computer frustrating as hell.
    Shell replacements for Windows exist and have already existed for longer than some Linux advocates on the Internet have been alive. If your WM does not show up on this list, then whoever coming up with this “awesome” shell of yours is certainly to blame.
    Also you just happen to be good at the GNU toolchain it’s not to have them at your fingertips instead of having to download and install them, often through some emulator as Cygwin.
    SUA (or Interix) is a POSIX subsystem that is present in parallel to Win32. Not a terrible lot of people (including *nix fanatics) even know it exists.
    Also, I found it highly doubtful that support for Unix-like tools are in popular demand among Windows users. After all, Interix was merely an optional component even in its hayday.
    Besides, what precisely do you mean by “GNU toolchain” anyway. Is it coreutils? binutils? “make”? GCC? What exactly is the benefit of using “chown” in an environment that does not use the owner/group/other permission scheme anyway? Perhaps you should elaborate before attempting to imply that your expectation is any more than strictly personal and non-functional.
    Continue with your “productive” ways, by all means, if that’s what you mean by “productive”.

  68. oiaohm says:

    Dr Loser in fact do tech support for Windows operating systems.

    ==Which Windows “issues” would those be, oiaohm? ==
    Name one really there is more than 1.

    Ok secure deletion of files while users are using machine. http://wipefreespace.sourceforge.net/ simple for Linux on the most common filesystems.

    Windows still depends on fill up all freespace with one huge mother of a file. This takes hours and you cannot have user running the machine at the same time. Otherwise bad things happen on windows.

    This is also why ghosting for rapid restore is a good thing. This is why WGA is such a pain.

    Linux has options on for network users not to store any data on the harddrive of the local machine while running the application in the ram and cpu of the local of machine. Windows does not have this options. Again this make secure clean up harder. Because data has been stored on the local machine.

    Damage caused by Mult anti-virus installed. Yes Linux does not get sick because you install multi commercial anti-viruses at the same time. Windows can develop all kinds of issues. Mostly because the anti-virus hooks in Linux were designed on the fact multi applications would be using them at the same time. Windows was design on the idea of only one using the hooks at a time so when you get two you get cat fight. Worst part is some malware also hooks into the same hooks under windows and linux. Linux it causes some issues but Windows it causes data destruction.

    Basically attempting to secure a Windows system and its kicking you in the teeth at every turn.

    ==Bear in mind, incidentally, that your latest pathetic effort (re: 9x replacement) involved a little-known and even lesser-used SourceForce foolishness that required a VxD driver.==
    VxD driver was extended life to 9x line. If you go look at the web stats you will see the 9x fads out rapidly in 2011 not 2007. Yes there is a rapid end of life reaction to end of life of kernelex were there was not to Microsoft end of life of 98se. Sorry VXD item was used enough to show up in web stats at a few percent. Little known might be a claim interesting that it still showed. More used in webstates than Linux.

    Dr Loser this is the problem you have not done your homework on what historically happens. Without knowing the history you cannot predict the future.

    There are a set of high risks with the end of life of XP. Just to be more interesting Dr Loser there is a prior example with Windows 3.11 as well. I am not referring to a solo freak event. List of conditions will be meet by Windows XP end of life for something that has happened twice in history already to happen again. I might be wrong and it might not happen again. The odds of history are on my side.

    In fact I think this is partly why Windows XP support was extended so many times by Microsoft in the hope the market share would drop under the trigger level.

  69. Joe says:

    If at all possible, I suggest you slow down a bit with the rapid-fire insults and try to add something productive to the conservation. Thanks kindly.

  70. Dr Loser says:

    But when I’m looking for technical support, I just rather get it from someone a little more domestic. Hopefully also someone without such a thick accent that I can literarly not understand 35-45% of the words coming out of his mouth.

    Then perhaps you should demonstrate your expert ability to distinguish one accent from another and politely request the Balngladeshi call-center worker to forward you to a more appropriate Bengaluru advisor.

    Astonishingly enough, these poor saps can count from 1 to 100. They also understand what an SME is. Only rarely do they throw rocks in all directions.

    You might not feel at home with them, and in all honesty if you just stuck with a licence for Windows, you would probably not even have to deal with them.

    But, what the heck? You’re clearly incompetent. You can’t even distinguish one nationality from another, let alone explain why on earth a mom’n’pop business would benefit from being able to read, examine, modify, and redistribute the software available on a $3000 HP workstation!

    Go ahead, my man, knock yourself out!

    Alternatively, and I realise this is going to be a bit painful for you, you might care to explain why you brought Bangladesh into the conversation in the first place.

    Note: You did that. Not me. And for reasons yet to be explained, you were quite offensive about an entirely innocent 200 million or so people.

    Not one of whom, as far as I am aware, have sullied the git history of gcc, SME or otherwise.

  71. Joe says:

    I got no problems with various South Asian cultural groups (although they technically not a seperate race), by the way. Especially their food.

    But when I’m looking for technical support, I just rather get it from someone a little more domestic. Hopefully also someone without such a thick accent that I can literarly not understand 35-45% of the words coming out of his mouth.

  72. Joe says:

    I thought you were leaving?

  73. Dr Loser says:

    But you typically get better quality hardware and tech support that isn’t out of Bangladesh. That to me, is worth it.

    Way to go on pointless and inaccurate racism. Can you say “Bangalore?”

    Probably not. You can’t even pick a number between one and a hundred.

    Your sort is doomed, and good riddance.

  74. Joe says:

    *for the higher end hardware stats

  75. Joe says:

    Anyway, back to the original topic.

    Now, obviously, HP can do that for all their PCs and they could bundle GNU/Linux as well. Pick any of the major desktop distros and consumers would be happy. It costs HP nothing to give consumers a choice. It is a distortion of the market to give M$ the bye on retail shelves. Clearly there is a market for GNU/Linux PCs.

    Yes, they can. But the problem is will they? The answer is probably “no”. Major OEMs have some complicated ‘business relationships’ with Microsoft.

    On the other hand many smaller OEMs have no problems with unbundling Windows on all their machines. And some OEMs like System76 exclusively focus on GNU/Linux.

    The costs are about par for the hardware stats, but smaller OEMs typically compete on the higher end and not at the bargin bin end of the market that HP and Dell mostly live in. But you typically get better quality hardware and tech support that isn’t out of Bangladesh. That to me, is worth it.

  76. Joe says:

    I’ve found examples of companies that aren’t IBM and Oracle employees, some of the most latest commits come from such people. I’ll give you a hint, start with the latest five commits from 87f13849b (current HEAD) and you’ll find at least one.

  77. Dr Loser says:

    But why did you decide to ignore the companies these people work for, except in cases where they worked for IBM and Oracle?

    I didn’t. I waded through your “all your base belong to us” source, and I identified the only ones I could, based upon the domain.

    Evidently you quoted the gcc git list without even bothering to examine it. Clearly, you have not returned to it, otherwise you would have picked a number between 1 and 100. I’d even accept a “cheat,” if you could find a single SME in there.

    You can’t, can you?

    You’re boring, ignorant, and pointless.

    I’ve given you several opportunities to state your case and back it up. You have failed every single time. (And, God knows, much as I love gcc and have indeed worked with it since 2.95 and before, on Solaris and various other *nixes, it’s not exactly a poster-child for reading-examining-modifying-redistributing, is it?)

    You tire me with your insolent ignorance.

    And I have no wish to sully Mr Pogson’s blog by pursuing this matter any further.

    Fire away! Pick a champion! I’ll let you bask in your victory!

  78. Joe says:

    But why did you decide to ignore the companies these people work for, except in cases where they worked for IBM and Oracle?

  79. Dr Loser says:

    Just to remind you, Joe:

    I think the original Four Freedoms were meant to benefit individual users rather than large corporations, but let’s run with that one, shall we?

    Never mind. I’m sure you are absolutely desperate to spend $3000 on a workstation you don’t need, just because it doesn’t come with Windows installed. (And quite easily blatted over, but of course “only if your time is worthless.”)

    Curiously, I didn’t see your name on the gcc git list. Nor have you quoted a receipt for this absurdly expensive piece of hardware. Nor can you quote a single SME at us.

    Never mind. As Douglas Adams once said, “Keep bashing the rocks together, guys.”

  80. Dr Loser says:

    You should be looking at their domain names.

    I did, joe, I did. I even added a couple in parentheses.

    However, it’s not for me to pick horses in this particular course. You, too, can look at domain names (for what little that is worth).

    I genuinely don’t care which domain name you pick. As I said before, anything between 1-100 is fine.

    Pick one, and we can continue with this analysis.

    Failing that, you are basically spinning fairy tales out of whole cloth.

  81. Dr Loser says:

    And … back to oiaohm.

    Dr Loser
    “$179 per annum for the full-on Red Hat experience, including support.”
    The fact that Windows issues will cost you more than that in a year on average anyhow. Kinda makes the $179 null and void. TCO is something it does not pay to pick on Linux for.

    Which Windows “issues” would those be, oiaohm? As far as I am aware, you steer well clear of that poisonous operating system, and therefore you have no experience of such “issues.”

    Go on, name one. And while you’re at it, you might put your best foot forward and support poor joe in his doomed attempts to throw rocks in various directions and have a 100% hit ratio for SMEs that have a serious and non-negotiatble requirement to view, modify and (well, maybe not) modify the source,.

    Bear in mind, incidentally, that your latest pathetic effort (re: 9x replacement) involved a little-known and even lesser-used SourceForce foolishness that required a VxD driver.

    You are aware that a VxD driver operates at a binary level, are you not?

    So much for “viewing the source, not the proprietary closed version.”

  82. Joe says:

    You should be looking at their domain names.

  83. Dr Loser says:

    Maybe I’m wrong, but I feel it shouldn’t be very difficult for you to follow the link I posted to get a more wholistic picture of the who contribute to GCC.

    Well, I did ask for a single company you’ve thrown rocks at. And I did ask for a single example of the usefulness of FOSS in either the GNU toolchain or the kernel to an SME. And I did make the brave offer to respond to a single random git update from 1-100.

    I think at this point I hold the moral ground and you are clearly not able to defend your thesis. However, to be fair, let me take the top 57 off your cite and allow you to pick the star SME amongst them.

    Let’s get (w)holistic on your ass!

    IBM: 2
    Hubicka: 2
    Timshen: 3
    Paolo: 4 (Oracle)
    Rsandifo: 2
    Daily Bump: 3
    Danglin: 3
    Wschmidt: 1
    Mrs: 1
    Olegendo: 1
    Uros: 5
    Ebotcazou: 4
    Law: 2
    Emsr: 1
    Amacleod: 4
    Tejohnson: 2
    Ktkachov: 1
    Ian: 4
    Jbglaw: 1
    Rguenth: 1
    Clyon: 1
    Jgreenhalgh: 1
    HP: 1
    Tmsriram: 1
    Meissner: 5 (IBM )
    Edlinger: 1

    Note that we are not discussing whether a single one of these submissions are useful, accurate, or relevant.

    Note that we are not even discussing the rather depressing fact that gcc is so broken that it requires constant nannying and burping.

    Note that several “contributors” are in fact employees of large corporations. There’s nothing wrong with that. I’d expect it.

    But your (unproven) theory is that you can throw a rock in any direction and hit somebody (I would assume an SME or equivalent, but maybe I’m being unfair) who absolutely requires the ability to “read and examine and modify.”

    My contention is that you are deluding yourself. I’ve given you enough ammunition off your own cite.

    Prove me wrong.

  84. Joe says:

    Sorry, I don’t desire to single out people or companies. Maybe I’m wrong, but I feel it shouldn’t be very difficult for you to follow the link I posted to get a more wholistic picture of the who contribute to GCC.

  85. Dr Loser says:

    Lack of means to provide required features yes. If not providing required features then Linux TCO blows out. If Linux features will do the job lower TCO.

    Let’s get this straight, oiaohm.

    Your argument, apparently, is that Linux has a lower TCO “if you don’t need required features.”

    I can’t help but agree with that one. Do you think, possibly, that the lack of “required features” might be a significant factor in large companies not adopting Linux?

    Perish the thought.

  86. Dr Loser says:

    Of course this requires a view of capitalism that isn’t entirely adversarial.

    Not really. It merely requires a model of the world that isn’t insanely simplistic.

    Corporations, governments, even individuals trade stuff based on “commercial secrets” all the time. Perhaps you haven’t noticed this, but it has in fact been the basis of modern capitalism since the time of David Ricardo, who explained how both sides benefit from such transactions.

    But that’s not really relevant, is it? Because you seem to have sure-fire proof that being able to “examine the code” (and again, I’ll let you off on the redistribution thing. Heck, Google don’t redistribute Goobuntu for some reason) is a huge benefit to the poor sods you spend your spare time throwing rocks at.

    Just name one of them. Stop theorising. Stop throwing rocks. Name one.

    A simple request, I think.

  87. Joe says:

    It’s not github. It’s git. They are different things you know.

  88. Dr Loser says:

    Hugely enjoyed the link to github for various perversions of gcc, btw.

    Care to nominate one? Pick a number from 1 to 100 at random, and I trust you to be honest.

    I’m quite prepared to analyse it in this context.

    Not much of an advertisement to the average SME, though, is it?

  89. Joe says:

    Another thing to note is the indirect benefit due to the sharing of useful code.

    If company X needs feature Y, they can contribute Y to the project.

    Now if a company Z also needs feature Y, they can get the feature because Z wrote it previously. But Y might not even exist if it wasn’t for company X.

    Of course this requires a view of capitalism that isn’t entirely adversarial (ie. “if X made Y, only X should benefit from it”, which isn’t always the optimal method to advance technology as a whole).

  90. Dr Loser says:

    But if your method of “gotcha” is to pick any arbitrary criteria for what constitutes a company or contribution. I’m sure you can find some other arbitrary criteria to “win” on. For instance, I’d have a hard time finding any companies who’s CEO is named Sally bin-Saddam that has exactly 14 employees that edit shortwave wave drivers in the Linux kernel.

    Quoted in its entirety, because it’s self-evidently hilarious.

    First of all, I’m not talking about “winning.” I don’t even know what “winning” means in this context.

    Secondly, I proposed a perfectly believable basis for the fact that these behemoths are sold without an OS. It won’t persuade any of us here to buy one, and it has nothing at all to do with FLOSS. (It has a lot to do with Linux cannibalizing the *nix base, and I can address that if you wish.)

    And thirdly, you claim to be able to pick a company, let’s assume an SME, more or less at random, whose chosen IT involves “examining and modifying the code.” (I’ll leave out the redistribution under the GPL just to make it easier for you.)

    Do you have one? Can you quote one?

    Apparently not. Take your time. I can wait.

  91. Joe says:

    Loser,

    The great thing about Free Software is the people or organizations putting code into the system isn’t secret. It’s actually cryptographically verifiable in some cases (Git repos). Go ahead and read the git logs for GNU GCC and you’ll see quite a few domains that aren’t all big companies.

    http://repo.or.cz/w/official-gcc.git/log

    But if your method of “gotcha” is to pick any arbitrary criteria for what constitutes a company or contribution. I’m sure you can find some other arbitrary criteria to “win” on. For instance, I’d have a hard time finding any companies who’s CEO is named Sally bin-Saddam that has exactly 14 employees that edit shortwave wave drivers in the Linux kernel.

    The point is the amount of companies modifying and contributing to FOSS is non-trivial.

  92. Dr Loser says:

    A lot can be done with servers instead of workstations too. The hardware is similar except for the graphics/sound cards.

    Oh, I absolutely agree, although the thing about workstations (and this goes back to the glory days of SGI and Irix) is that they basically sell themselves into narrow vertical markets (eg CAD/CAM) via those not-particularly cheap graphics and sound cards.

    A happy result for Linux enthusiasts is that there’s still a huge amount of Posix/*nix commercial closed-source software that is out-of-box ready to run on these $3,000 behemoths.

    I contend that this is why they are available without an OS.

    It also seems a little peculiar to boast about the lack of an OS on a $3000 machine that you (and I, and obviously Dougie and oiaohm) will never buy. I can see how $30-$50 might sting on a $300 machine, but for the life of me I can’t imagine anybody who would stump up the cost of a decent second-hand car, if only it didn’t have that pesky “Microsoft Serf Tax.”

  93. Dr Loser says:

    I can throw you a rock in any direction and find you a company that does in fact take advantage of the open nature of the source code.

    I think the original Four Freedoms were meant to benefit individual users rather than large corporations, but let’s run with that one, shall we?

    Having thrown your rocks, can you name the SME (rough definition, less than 500 employees) to which this claim applies in any meaningful sense? And by “meaningful,” I’m stipulating either the GNU tool-chain or the kernel itself.

    Rather an unlikely proposition, I think. And if you reduce the stipulation from 500 to under 10, I would suggest that you’d have to heave an entire quarry of rocks before finding a suitable candidate.

  94. Dr Loser wrote, “after you’ve rushed out and bought your $249 ChromeBook, you will also be acquiring a $3,000 GNU/Linux workstation?”/span>

    About the only thing I do that might cause a blip in its wait-queue is building a kernel. My old Beast does it in ~1/2h and that’s done every few weeks, nothing urgent at all. I just make -j a few and keep on using the system. I occasionally index my file-systems but it doesn’t seem reasonable to buy a new machine to speed such things up. I do sleep and can put things to work then. The last time Beast really sweated, I updated my local copy of Wikipedia. That got half way through before power failure knocked it off… Nothing vital needing that kind of power here. I expect such workstations are needed for those multimedia creators/editors, CAD-types and bug-hunters who bisect the commits to find what broke what… A lot can be done with servers instead of workstations too. The hardware is similar except for the graphics/sound cards. There’s certainly no need for a more powerful workstation just to build kernels. A server could do a better job cheaper. e.g. instead of a bigger graphics card, one could have a few more multi-core CPU sockets and tons of SSD arrays or a cluster of more modest servers. GNU/Linux can do it all.

  95. oiaohm wrote, “You are dismissing the cost of Maintaining windows systems as zero cost bw. 5 extra hours of labour maintaining a computer is the 100 dollars in licensing gone.”

    Lots of businesses budget >$1K per seat per annum just because of that other OS. Folks who use GNU/Linux spend way less than $100 per seat per annum. At one school I had 153 seats of GNU/Linux and once it was all installed and running, there were maybe two or three issues per annum for the whole system. I used to start my day checking things out and it was so boring. Every day was the same. Stuff just worked. My recollection is that in the first several years the only significant issues were two RAM modules and a hard drive on the servers. The client machines never had any issues. I doubt system administration cost more than 15 minutes per day, a few dollars at best, a few $K per annum for 153 seats, ~$10 per seat per annum. In systems running XP, I had issues almost every day (slowing down, malware, re-re-reboots and failures to boot) and I often worked late fixing them, hours per day. It was always that other OS messed up. Hardware problems were likely two or three issues per annum. Do the maths.

  96. bw wrote, ” If someone wants to put some ersatz version of Linux on it, then they are on their own.”

    Linux is Linux. RedHat and everyone else contributes drivers and other code to the same source-tree. It’s not ersatz to take the latest release from kernel.org and expect it to run on current hardware supported by HP and RedHat. Folks who actually make the pieces and the finished goods really do want it to work with GNU/Linux of any distro. The worst that can happen is that some distro may not have a particular driver built in or built as a module. A simple compilation fixes that. It’s no sweat, especially if you have a bunch of identical machines. In schools I had many different configurations of hardware old and new and the latest kernel in Debian worked just fine. People using a distro are not on their own. The folks creating the distro, using the distro and all kinds of upstream projects like Linux all work for the end users.

    The only problem with using random distros on some machine is that “the latest version” may have a bug not yet fixed or discovered. That’s why distros like Debian GNU/Linux have “stable” versions which are very well tested in all kinds of uses. Anyone really concerned about this will do their own testing and not rely on RedHat or anyone else to do the job. If RedHat loses their data, what recourse do they have? None. Read the contract. Neither does M$ take care of lost data. People have to take care of their own stuff or pay someone to do that. RedHat only makes sure the software works. They are not going to maintain your system and its data for ~$100 per annum.

  97. Joe says:

    There is also window managers that are only available really on GNU/Linux, for instance awesome. Without awesomewm, I find using a computer frustrating as hell.

    Also you just happen to be good at the GNU toolchain it’s not to have them at your fingertips instead of having to download and install them, often through some emulator as Cygwin.

    At last, but for some reason controversal amongst the M$ fans is the fact that GNU/Linux is Free Software and completely modifiable from the top down. The typical response is “grampa doesn’t need that”, but many companies don’t ignore this. It’s impossible to write code once and have it applicable to every possible use case. That’s where Free Software comes into play. I can throw you a rock in any direction and find you a company that does in fact take advantage of the open nature of the source code.

  98. Joe says:

    Linux has a lot of specialized distros that are difficult to emulate in Windows.

    Kali Linux for one. You can spend weeks trying to make a similar environment (not complete) on Windows, or you can just download Kali Linux and have a working netsec system in 15 minutes.

  99. bw says:

    I think bw is being a bit mealy-mouthed here

    I think that I stated the facts fairly directly, loser. Up front, HP says you can buy this machine for less if you install your own Red Hat or SUSE and presumably they will guarantee fitness of purpose. If someone wants to put some ersatz version of Linux on it, then they are on their own. Sure that comes cheaper.

    What is it, though, that someone as sophisticated as dougman might do with such a beast using Red Hat that he could not do with Windows. Or Pogson for that matter. I am interested in the answer here.l

  100. oiaohm says:

    Dr Loser
    “$179 per annum for the full-on Red Hat experience, including support.”
    The fact that Windows issues will cost you more than that in a year on average anyhow. Kinda makes the $179 null and void. TCO is something it does not pay to pick on Linux for.

    Lack of means to provide required features yes. If not providing required features then Linux TCO blows out. If Linux features will do the job lower TCO.

    bw also beware SUSE and Redhat are both to operational after install from disc inside half a hour on that level workstation hardware. Windows OEM setup in fact stuffs around for longer.

    Yes its wrong when time to userablity is won by the machine you install the OS clean.

  101. oiaohm says:

    “It is just as misleading to claim that the freebie versions of Linux are as potent and efficacious in workstation roles as the products from SUSE and Red Hat.”

    In fact what you are saying is missleading bw. That is the problem. Like Scientific Linux in built particularly for high in science area workstations. So that freebie in a far better fit for particular roles. Scientific is Redhat compatible.

    Apples and Oranges is exactly right. The desktop users requirements chooses the required distribution. So high level science workers are going to be scientific linux in most cases that is a freebie so all the math tools they use are well and truly integrated. Something Redhat and SUSE does not do. So giving those staff Redhat or SUSE in crippling them. So the freebie in this case is more potent and efficacious than the charge for competitors if you are doing high end data processing.

    Shock horror right Redhat and SUSE not suitable for desktop usage for a lot of users.

    This is why we laugh at you so much. 1 size does not in fact fit all bw. Please stop thinking commercial Linux in fact fits. There are a lot of users where SUSE and Redhat don’t fit.

    Yes the horse and apples is very true. The problem here is the customised to market Linux solutions will take less than half the setup time of a non customers to market Linux. Even so a redhat or a suse take less time than a Windows install in a lot of ways.

    bw the french police and other use freebie Ubuntu. Report lower TCO due to lower operational failure rate and better remote repair ability.

    So even if you have to pay the same on Linux as Windows buying a new windows license Linux is still cheaper on TCO.

    You are dismissing the cost of Maintaining windows systems as zero cost bw. 5 extra hours of labour maintaining a computer is the 100 dollars in licensing gone. Yes Windows on average needs lot more maintenance.

    bw the means to ghost images effectively equals reduced install time. Faster repair.

    Basically MS WGA stuff makes windows much more costly to repair. Then MS 3.5 GB stupidity on top.

  102. Dr Loser says:

    Gotta say, those are pretty neat-looking machines for $3,000+. At that price, I for one would not balk at paying $179 per annum for the full-on Red Hat experience, including support. I think bw is being a bit mealy-mouthed here.

    I’m as excited by this development as you are, Robert. May we assume that, after you’ve rushed out and bought your $249 ChromeBook, you will also be acquiring a $3,000 GNU/Linux workstation?

    I’m looking forward to an unbiased review of this new Beast.

  103. Joe says:

    It is just as misleading to dismiss the effort required to install and maintain Linux as being zero cost

    That’s a great observation. But, the cost of maintaining Windows isn’t zero either.

    And with the OEM license, you are just getting the right to use the OS and get automated updates – nothing more. The OS itself is cheap, but not when you want central management, tech support, or e-mail, or office apps. Those cost extra, a lot extra. Once you start adding in all the various stuff that is included in a Red Hat subscription, it will make Red Hat’s fees look like a bargin compared to the same thing from Micro$oft.

    Micro$oft is really good at the whole “drug dealer business plan”, they’ll make their stuff look cheap at the start (less than what you pay for “toilet paper”!) and reel you in to hundreds of thousands of dollars of license fees over time. All while getting you locked in (“addicted”) to their products to make it even more expensive to even think about switching to alternatives like GNU/Linux.

    Lost opportunity and switching costs due to vendor lockin are of course, never part of their TCO calculations. Because why would you ever want to use something not from Micro$oft? That’s just silly!

  104. bw says:

    Of course that’s misleading

    It is just as misleading to dismiss the effort required to install and maintain Linux as being zero cost as opposed to paying the $105 to have it done for you. It is just as misleading to claim that the freebie versions of Linux are as potent and efficacious in workstation roles as the products from SUSE and Red Hat. You are not comparing apples to apples here. You are comparing apples to horse apples.

  105. bw wrote, “To install your Linux OS, you must purchase a Red Hat Linux box set or SUSE Enterprise Linux OS”.

    Of course that’s misleading. Folks can just install Debian GNU/Linux from a flash-drive. Some wouldn’t mind paying more for GNU/Linux. It’s not about price but FREEDOM. With GNU/Linux you have the choice to maintain the system yourself or not. M$ does not maintain systems as far as I know. Their “updates” keep slowing them down and introducing more bugs. Many folks who care about their systems turn M$’s updates off.

    So, no, subscriptions don’t start at $179 per annum. It’s $0 or $49 depending on how comfortable one is with GNU/Linux. I know a lot of people who are very comfortable with it.

    The difference between “Desktop” and “Workstation”?
    “Desktop
    Designed for task workers. Typically requires a limited set of applications. End user has limited administrative control over the system. System administrators need to be able to pre-configure and remotely manage consistent images across a large number of systems.
    Primary applications in use are productivity applications like Firefox, Evolution/Thunderbird, OpenOffice, Planner/TaskJuggler. Primary services in use would be network file systems like SMB/CIFS, NFS, printing services, IM.
    Workstation
    Workstations are designed for advanced Linux users. In addition to the tools provided in the Desktop variant, the Workstation variant supports a stand-alone development environment. End users are expected to have local super-user privileges or select super-user privileges.”

    is a development environment, of no use to many users of powerful PCs.

  106. bw says:

    You pay $105 bucks less to HP, but the fine print says

    Your HP Workstation was configured without a Linux Operating System (OS) as you requested. Hewlett-Packard has offered this solution to allow you more flexibility in installing the Red Hat Linux OS or SUSE Enterprise Linux OS of your choice.
    The HP Installer Kit does not include an OS. To install your Linux OS, you must purchase a Red Hat Linux box set or SUSE Enterprise Linux OS from the following Red Hat or Novell web sites and install according to their instructions:

    A workstation OS from Red Hat starts at $179/year. How many years are you going to keep the machine? There are no extra payments for Microsoft.

  107. Aieee a Ballhogg says:

    It’s interesting that none come by default with an optical drive. I guess those days are soon over. USB keys or OEM installs will be the only way forward.

Leave a Reply

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