GNU/Linux Adoption Grows in California

I originally wrote this as a comment but I think it’s worth a post. I reported that Distrowatch saw a sharp increase in activity for the popular distros and was replying to a comment.

===========================================
NetApplications is seeing something similar:

week of September 30, 2012, USA GNU/Linux =1.91%

month of September 2012, USA GNU/Linux= 1.59%

What big roll-out of GNU/Linux happened last month?

NetApps:
California, September, 8.81%
week of September 30, 10.46%

Yipee! It’s Mountain View, CA which went from 10% for September to 52% for the week of September 30. Did Google expand? Did a neighbour take up GNU/Linux?

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.

27 Responses to GNU/Linux Adoption Grows in California

  1. Chris Weig says:

    No wrong. 1972 Bayer Tree was invented then. There are a few variations off it that happen quickly after. That are also called b-tree’s.

    Dr. oiaohm, even your regular trusted source, Wikipedia, disagrees with you. Why don’t you just give up and retreat to the Australian Bush? I mean, yes, you’re already there, but please cut off your internet connection. Go get some new certificates, perhaps in rhetoric.

  2. That Exploit Guy says:

    ‘Computer Surveys’ -> ‘Computing Surveys’

  3. That Exploit Guy says:

    @oiaohm

    ‘1972 Bayer Tree was invented then.’

    No, as the article tells you on p. 122, what “B” in “B-tree” stands for is never revealed. There is only one “B-tree”, and that’s the B-tree proposed by Bayer and McCreight.

    ‘There are a few variations off it that happen quickly after. That are also called b-tree’s.’

    Look at the footnote. The acticle you reference to was published in Computer Surveys in 1979 and contains clear and distinct nomenclatures for each and every B-tree variant discussed there. The alleged confusion of B-tree existed only in your imagination and all the claims you have built upon can thus be conclusively disproven by the very existence of the article itself.

    The rest your comment is mostly rubbish built upon the same faux premise, and I am not going to waste my time reading it.

  4. oiaohm says:

    That Exploit Guy
    –B-tree was invented in 1972 by two Boeing employees, and it had been well defined for two decades even before NT existed. Have you heard that story about the boy who cried wolf?–

    No wrong. 1972 Bayer Tree was invented then. There are a few variations off it that happen quickly after. That are also called b-tree’s.

    That Exploit Guy read
    http://people.cs.aau.dk/~simas/aalg06/UbiquitBtree.pdf
    By 1979 B-tree has become a mess of variants all called B-tree. This is still the case by 1991 when windows NT starts using and design the registry Hive. With people fighting over the most effective way to make a b-tree.

    Read D Comer deletion. If less than d x 2 keys remain then leaf gets removed and merged with is next door leaf. This does not match the first define you just showed.

    Now under D Comer you set d key value to 0 and you never ever delete a leaf once its created. Yes concatenation trigger is d times 2.

    D Comer B-tree is Ubiquitous B-Tree that is not Bayer Tree(B-tree) exactly. Close minor differences.

    The differences between the B-trees is how and when delete is defined. D Comer B-tree allows for never deleting leafs.

    That Exploit Guy yes B-tree was not the only confused term without a solid locked define. Notice that B*-tree got called B-tree by some and A define of a B-tree got called a B*-tree by some. Even that they are nothing a like.

    So from 1972 to 1979 just 7 years we go from a well define thing of what a b-tree is to a confused mess.

    From 1979 to 2000 the confusion reduced but was around that complete time. But it took a long time to get everyone on the same page. Made worse by some using B-tree to mean binary tree.(not if there was not enough confusion)

    From 1979 to 1998 a few others still throw in new variations to b-tree as well.

    That Exploit Guy D Comer wrote his complete book trying to sort out what was and what was not a b-tree with the word abuse going on that the time.

    That Exploit Guy the critical question with Windows Hives is what B-tree are they using. It looks a lot like a D Comer B-tree with d set to 0. But there are about 4 others it could be.

    D Comers have a nice mathematical bug you have 1 leaf with less than d x 2 entries no where to merge what do you do now because you are meant to perform a merge. Yes there are some implementations doing D Comers you find a good coding bug right there.

    The major difference is delete processing. You can detect what most of the variants you are dealing with by looking at how the delete is done. This is where most of them go different ways.

    –Cormen, Thomas; Leiserson, Charles; Rivest, Ronald; Stein, Clifford (2001), Introduction to Algorithms.–
    This is when the modern day define starts and the dog fighting between implementations stop and what we call a modern day b-tree starts. This is basically lets return to the 1972 define and say that is a b-tree and say everything else defined needs its own name. Even in the 2001 book they define 2 ways of doing delete in a b-tree.

    The page is index in the borland book look up b-tree and see where it takes you. The result is completely wrong. That Exploit Guy I did not say the borland books was ever right. Its a historic bug in those source books.

    That Exploit Guy you are the inexperienced here. I lived that time frame with all the confusion.

  5. That Exploit Guy says:

    @Oiaohm

    ‘This is the problem b-tree define has evolved with time. NT starts using b-tree in registry before its define is rock solid.’

    Hah! What a yarn you have spun there. B-tree was invented in 1972 by two Boeing employees, and it had been well defined for two decades even before NT existed. Have you heard that story about the boy who cried wolf?

    Actaully, scratch that. If you ask me, I can even give you scans of the pages of that data structure book I cited in my previous comments with no questions asked. Can you at least tell me which pages of those books “written by Borland” I am supposed to look at? No?

    What are those hundreds of words in your comment good for, then?

  6. oiaohm says:

    That Exploit Guy
    –Only someone who’s clueless enough to write something like “the B-tree is a binary tree” would take your comments seriously.–

    On this you need to go to your local library look for books written by Borland pre 1996. You will find in their Pascal and C++ books that a B-tree is a binary tree and what we call a b-tree today is called a balanced tree in those books with how todo delete undefined. This is the problem b-tree define has evolved with time. NT starts using b-tree in registry before its define is rock solid.

    Robert Pogson has stated that he is a pascal programmer that means his terms might come from the Borland books.(or most likely will) So getting B-tree wrong is what you would expect if he has been telling the truth.

    The term as we know today B-tree only really officially locks in the year 2000. Before that you have the term b-tree used for binary trees in some books. This is language evolution with time.

    By the way point out word usage based on cultural grounds is not a english lesson. Its a language history lesson focused on english. History lesson is not teaching someone to read write english. It teaches you why its so screwed up.

    This is something a person who has just goggled would normally never know.

    When dealing with older programmers and you say B-tree make sure they are thinking the right thing.

    That Exploit Guy the other thing with defrag on write is under memory pressure. If you have been forced to drop your file cache reading it back from disc is slow if the files are fragmented.

    The the file-system fragmentation hits you from many sides. There are many reason to get rid of it. Of course you want todo it with the least number of extra read/write to what general operations do. The lowest is always defrag on write but it has limitations. Like the files having to be cached before able to defrag.

  7. oiaohm says:

    That Exploit Guy
    –Sure, because obviously the cost of moving the file from one region of the platter to anothor (or moving from one platter to another) and then appending it is somehow faster than just appending.–

    Note I said write to file. Not just appending. NTFS you rewrite the full file from start to end since it was an existing file its rewriting back down fragmented. Programs do this. So chances to defrag at no cost appear as part of normal operation. NTFS under windows is not looking for them.

    So there are a percent of defrags of files that are basically free if you do them as part of the write operations. Even writing the disc struct down is basically free due to journal allowing many operations to be bundled and dumped to drive in one hit.
    Robert Pogson
    –Obviously defragging a bit at each I/O operation increases the cost of that operation buy may allow a thousand others to go much faster.–

    This is also true and false defraging on write does not always cost at each I/O operation compared to what you would have had do if you did not defrag at write.

    That Exploit Guy you need to understand how the Linux system pulls it off. When you read/write a file its cached in ram. So now to move that file its already in ram. Was there as part of normal operations. Same is true under windows. So you have already copied the file off the disc.

    –the cost of moving the file from one region of the platter to anothor– So it not move.

    Its all about how fast can you write the file out. A file fragmented in 1 place is 2 operations a file fragmented in 2 places is 3 operations.

    That Exploit Guy you must not have done OS operation theory yet covering how it caches stuff. Linux defraging on write would be costly if the files were not already being cached in ram.

    While you have a dma read/write going on the cpu can go on and do other things. The cpu has to stop to que up IO operations. Lower you IO hit the more time your cpu has to process stuff. Ok write to disc might take a little longer the saving on future reads pay for this.

    That Exploit Guy there is only an obviously the cost to an idiot who does not understand what operating systems are doing in background.

    This is the problem it appears expensive to move a file. This is why running a defrag program is expensive. The files you are moving with a defrag program normally are not already in ram due to other operation requirements so have to read from disc and written. This is the other shocking reality doing it on write you have halved you defrag file cost due to the file already being in memory due to other operations. Even better you get to subtract off the write cost.

    If you want the most optimised defrag for IO costs you have it on write. When you write you fix fragmentation.

    Lets say a file is fragmented and it has not fully been read into ram. Ext 2/3 will not defrag it on since its not in the disc cache. This is when shake is required to tell the OS to pull the file into cache and rewrite it.

    What if you don’t have ram big enough to cache it. Yes ext 2/3 + large files bigger than ram equals I am going to fragment and not be able to fix it

    This is a weakness of ext 2/3 and why ext 4 has a background defrag program.

    That Exploit Guy ext is not a perfect file system. Ext is fragmentation resistant not fragmentation proof. There are things you can do it fragment it.

    That Exploit Guy the other thing you miss is most file writes are not appends. They add something in middle/start and something to end. Different locations on disc of the start end end means doing completely independent maths for both. Getting the file de-fragmented saves on cpu operations as well.

    That Exploit Guy really when you look at doing write down each I/O and CPU you operation you will perform doing with write defrag and without write defrag with reads mixed in. Write defrag always wins.

    This is a case where what appears to be common sense is completely wrong. You miss seeing the true common sense because you failed to see that the file would already be in ram done to other operations. And the only time you avoid having to read the file from disc specially to defrag is if you do it at write.

  8. That Exploit Guy says:

    @Robert Pogsonm

    “TEG needing reminding that he wrote…”

    I remember what I wrote, and instead of reply to the following like a grown-up:

    ‘Of course, we are absolutely in the position to be certain if those “thousand other operations” need the same kind of shuffling around or not.’

    You simply decided to whine about me rightfully ridiculing your for claiming credentials less-than-fascinating subjects like clicking the “defrag” button in Windows Explorer.

    Pathetic.

  9. TEG needing reminding that he wrote, ““because obviously the cost of moving the file from one region of the platter to anothor (or moving from one platter to another) and then appending it is somehow faster than just appending.””

    I replied and he criticises my credentials. I guess he was just trolling.

  10. That Exploit Guy says:

    @Robert Pogson

    ‘I have worked with fragmented file-systems and defragging does improve most I/O operations.’

    What a credential! I am pretty sure no one else having used a defragmentation tool can claim the same thing, now can they?

    ‘Obviously defragging a bit at each I/O operation increases the cost of that operation buy may allow a thousand others to go much faster’

    Of course, we are absolutely in the position to be certain if those “thousand other operations” need the same kind of shuffling around or not.

    Mr. Pogson, you have either just solved the halting problem in its entirety, or you are one of the most clueless, desperate, loud-mouthed individuals I have seen for a while.

  11. TEG wrote, “because obviously the cost of moving the file from one region of the platter to anothor (or moving from one platter to another) and then appending it is somehow faster than just appending.”

    I have worked with fragmented file-systems and defragging does improve most I/O operations. Obviously defragging a bit at each I/O operation increases the cost of that operation buy may allow a thousand others to go much faster. e.g. Transfer rate =100MB/s, average seek= 8 milliseconds, throughput lost per seek is 100MB/s X 8 ms = 800KB and that applies equally for reading or writing and rotational latency adds to the problem. So, a fragmented system can be many files behind while booting or loading an application, a real drag on morale. The cost of occasionally fixing some fragmentation during a write may well improve performance for many other read/write operations. Typically proximity of blocks will help reads much more than writes because many files are written once and read many times but writes are improved too.

  12. That Exploit Guy says:

    @oiaohm

    ‘Art of War I have the book. You need to read contents of the chapters to find the quote.’

    I wonder what exactly you have learned from all those 25 years you’ve spent in China – this is, of course, supposing that you have actually been to that country before.

    Anyway, can you hear that? It’s the sound of about one sixth of the world population laughing at you for claiming to have read the Art of War but not even being able to recognise one of the most famous sayings from it.

    If you want to lie about something, at lease have the sense to choose an obscure subject to start with. Claiming to have read an ancient piece of literature that most Chinese speakers and many non-Chinese speakers are familiar with without actually having read it is just downright disingenuous.

  13. That Exploit Guy says:

    @oiaohm

    ‘Reality defragging on write reduces IO operations so enables you to write faster to disc.’

    Sure, because obviously the cost of moving the file from one region of the platter to anothor (or moving from one platter to another) and then appending it is somehow faster than just appending.

    Why do I have this nagging feeling that you have trouble to even just add one to one?

    @dougman

    ‘If M$ registry wasn’t buggy, why are there so many cleaners available?’

    What’s the relevance between registry backup and registry “cleaning”? Ah, I see… You sell registry cleaners for a living, hence of course the last thing you want is someone casting doubts on them. Am I correct?

    ‘To overcome this error message, an updated backup is required for the Windows user.’

    I find it interesting that you often quote random paragraphs from the Internet into your comments without attributions to their respective authors. Are you afraid that people will sooner or later find out you often are just parroting <a href="less-than-authoritative sources in order to scare people into buying your wares?

    Anyway, even if we take “your” (note the scared quotes) word as-is, what do ‘operating system crash, virus attacks, and improper system shutdown’ exactly have anything to do with the registry being “buggy”? Mr. salesman… I mean, dougman, do you care to elaborate on that?

  14. dougman says:

    If M$ registry wasn’t buggy, why are there so many cleaners available? Auslogics makes a nice one and even provides you with backups before you make changes to roll-back. ERUNT is a tool I use to backup registries.

    Corruption in Windows registry is one major reason for loss of data from your hard drive. Few major causes of registry corruption are operating system crash, virus attacks, and improper system shutdown. In most cases of registry corruption, the Windows operating system becomes unbootable and displays an error message.

    To overcome this error message, an updated backup is required for the Windows user. However, the problem continues if the user has not created any valid backup. In such cases, the data saved in the hard drive can only be accessed using third-party Data Recovery Software such as a Linux Live CD

    Windows XP, 7 & 8 still suffer from previous failings. I skipped Vista, as it was a Joke.

    http://social.technet.microsoft.com/Forums/en-US/w8itprogeneral/thread/8336b8a6-0e6b-417e-be55-5803db73ed10

    http://support.microsoft.com/kb/307545

    http://www.forumswindows8.com/crashes-debugging/corrupted-registry-win-8-a-498.htm

  15. oiaohm says:

    That Exploit Guy reality you have screwed up so now you have to attack the person.

    I only worked out late that you were completely full of it when you started talking about defragging on write being costing. Reality defragging on write reduces IO operations so enables you to write faster to disc. There is no disadvantage at all. Keeps disc less fragmented and you get writing done faster. Its win win win. Yet NTFS does not do it.

    Art of War I have the book. You need to read contents of the chapters to find the quote.

    That Exploit Guy you really cannot lie your way out a wet paper bag. If I leave you will attempt to get away with bogus crap. Like trying to claim the registry is not buggy.

  16. That Exploit Guy says:

    @oiaohm

    ‘Go read the Art of War there is a basic idea you are missing on winning people over. Also I have a feeling you are being bogus because I have never seen that exact line.’

    “孫子曰 不知彼 不知己 每戰必敗”

    So you have read about the Art of War on Wikipedia? Fascinating. I am pretty sure you know exactly what the above quote tells you, then. If not, let me ask you one simple question – why don’t you just give up already?

    If people want to know something from Wikipedia, they will read it on Wikipedia themselves rather than from someone who can only paraphrase it poorly.

    If people want to learn more about the English language, they will take lessons of it from someone who can actually write it rather than someone who can’t write anything other than semi-intelligible moonspeak.

    If people want to entertain themselves with fairy tales, they will find someone who can tell a good story instead of someone who can’t even lie his way out of a wet paper bag.

    Bottomline: They don’t need you.

    Oiaohm, don’t you know that your game was up about a week ago, and you are just throwing away what’s left of your own dignity by posting this kind of pretentious comments here?

    Also, take a look at the bottom of each page here. What do you see?

    Only someone who’s clueless enough to write something like “the B-tree is a binary tree” would take your comments seriously. You have no idea where you are standing, and you have no idea how much other people are ahead of you in your own little game. The Art of War… I wouldn’t even trust you quoting from G.I. Joe, to be brutally honest.

  17. oiaohm says:

    Chris Weig
    –Yes, sarcasm on the internet does work. If someone doesn’t get it, then that’s precisely when you know that it works.–
    No its when your creditability goes out the window. Since now they question everything you post and don’t trust you any more because you did not mark it. Also its only sarcasm if a section of its not possibly fact.

    It takes knowing the topic way better than you to make sarcasm about it that is not foot in mouth Chris Weig.

    I would say with you knowledge level stick to what you think are facts. You don’t know enough to do sarcasm.

    –By the way, look at Robert’s tag line: “One Man. Closing all the Windows”. This is a rhetorical device, too. Has anybody ever complained about it?–

    Go read the Art of War there is a basic idea you are missing on winning people over. Also I have a feeling you are being bogus because I have never seen that exact line.

  18. Chris Weig says:

    Chris Weig sarcasm + Internet don’t work. You have been told this before. It pays to mark what is sarcasm.

    Yes, sarcasm on the internet does work. If someone doesn’t get it, then that’s precisely when you know that it works. Most of the time that’s exactly what you want; that someone doesn’t get it.

    By the way, look at Robert’s tag line: “One Man. Closing all the Windows”. This is a rhetorical device, too. Has anybody ever complained about it?

  19. oiaohm says:

    Chris Weig sarcasm + Internet don’t work. You have been told this before. It pays to mark what is sarcasm.

    Its even more funny when what you are saying as sarcasm could be part of what really happened.

    Chris Weig I got that you thought it was a joke. Problem was the joke was more on you because the time frame matches. The change in numbers line up with when editshare started putting out proto released to customers.

  20. Chris Weig says:

    @JR:

    Why? Do you want me to put smilies behind my posts so you understand their intention, too? Pogson recently wrote something about sarcasm and/or irony being unnecessary/unhelpful on the internet. He’s wrong. These rhetorical devices are very useful, as you aptly demonstrated.

  21. JR says:

    @ Chris Weig

    Only someone with your intelligence can turn this:

    “GNU/Linux Adoption Grows in California’

    into this:

    “I guess the porn industry has switched to Kdenlive or Openshot.”

  22. oiaohm says:

    Chris Weig really I never said I was in porn production. In fact I am not. My blender usage is normally for building and production line simulations.

    The timing of things makes your statement questionable. Leading up to a release software testing is always under way.

    The simple reality here your statement was not in-line with what is happening Chris Weig. Editshare was giving some early releases out of lightworks for Linux to some of there major customers for feedback. Some of those will be in the porn industry.

    Blender current is very good with green/blue screen work also does a really multi overlay effects. So its vastly more suitable for lot of movie work than Kdenlive or Openshot.

    Chris Weig kdenlive is responding. http://www.indiegogo.com/kdenlive-re They are seeing they now need to pay developers full time as blender has done for years to get stuff done.

    Openshot is also re-factoring.

    Chris Weig competition is good. There is a problem of course. If high grade is not in a market and everyone in a market is competing with other low grade there is less pressure to evolve.

    We are going to see evolution pressure.

  23. Chris Weig says:

    Don’t joke there is a chance they have got their hands on a pre-release Lightworks for Linux from editshare. Lot of them would be using editshare storage solutions.

    Right. Weren’t you responsible for the visual semen effects in “Mandingo Invades the Australian Bush”? Did you do them in Blender?

  24. oiaohm says:

    Chris Weig
    –I guess the porn industry has switched to Kdenlive or Openshot.–
    Don’t joke there is a chance they have got their hands on a pre-release Lightworks for Linux from editshare. Lot of them would be using editshare storage solutions.

    Videos of Lightworks running on ubuntu are already on youtube and general public release is 30 Oct 2012.

    Video editing is something where Windows Linux and OS X will equal.

    Chris Weig the reality is they might not have change NLE they are using. Just now the NLE they know and love is not longer locked to windows.

    Kdenlive and Openshot both have some very strong open source competition from lightworks and blender.

    When you combine in Octopus 6 things get even more interesting. News rooms being able to run on fully Linux based platforms. Pay for Octopus and support from editshare and you have a newsroom with all the software you need. No need to pay Microsoft a cent.

  25. Chris Weig says:

    I guess the porn industry has switched to Kdenlive or Openshot.

  26. ssorbom says:

    News to me, I’m from California. Explosion of interest could have something to do with Gov Jerry Brown’s Open Textbook initiative (increased interest in Free Culture maybe?).

Leave a Reply