Catch up on stories from the past week (and beyond) at the Slashdot story archive


Forgot your password?
Software Linux

Torvalds Unveils New Linux Control System 527

BlakeCaldwell writes "CNet reports: 'Linux founder and leader Linus Torvalds has launched a new tool, called Git, to manage his software project, after a dispute led him to drop the previous system.' He will start using Git instead of BitKeeper to control the flow of updates and track changes in the kernel." We've covered this previously. Relatedly, ChocLinux writes "Jeremy Allison, who wrote Samba with Andrew 'Tridge' Tridgell, is sticking up for his friend in the row over BitKeeper. "
This discussion has been archived. No new comments can be posted.

Torvalds Unveils New Linux Control System

Comments Filter:
  • Git? (Score:5, Interesting)

    by YrWrstNtmr ( 564987 ) on Thursday April 21, 2005 @08:52AM (#12301611)
    Isn't that a bit of a disparaging name in English as it is spoken in the Olde Country?

    As in "You daft git!"

  • by iggymanz ( 596061 ) on Thursday April 21, 2005 @08:53AM (#12301623)
    maybe the kernel programmers should take 2 weeks and fix the basic flaws of git, like the business of not storing deltas to files
  • Obligatory (Score:5, Interesting)

    by mcrbids ( 148650 ) on Thursday April 21, 2005 @08:59AM (#12301668) Journal
    Now, expect a thundering herd of comments like:
    Great! Now, Linus can be helping to build OSS counterparts to commercial software that can be truly trusted, rather than rely on the whims of a commercial vendor.
    This is just another example of where OSS software is MORE RELIABLE than their commercial counterparts.
    The thing is, they'd be right.

    The only thing is to remember: The terms of Linus' use of BK was noncommercial which is poison to a commercial entity. The combination of closed-source + no charge == noncommercial. If it was OSS, with a GPL-like license, at least the OSS community could give something back to BK that wasn't money, but it wasn't, and BK had no opportunity to profit in ANY WAY from this move.

    I'm not surprised this didn't work out well.
  • Eh? (Score:5, Interesting)

    by Aldric ( 642394 ) on Thursday April 21, 2005 @09:08AM (#12301718)
    Although Torvalds is revered by many within the open source community as the founder of Linux, he also has detractors among the free software movement. There is even a conspiracy theory on news site Slashdot that the anti-Torvalds rhetoric may have the underlying aim of persuading the open source community to switch to Hurd -- an alternative to the Linux kernel that is being developed by the Free Software Foundation.

    Did I miss something? I saw some comments to that effect in the stories, mostly as a joke except for the usual random nutcases that see conspiracies in everything that happens. Terrible journalism from zdnet here.

    The rest of the article wasn't any better, being the most heavily biased piece of crap I've read since the last TCO study by Microsoft. Linus and Tridge both have valid points but the article paints Tridge as a villain breaking BitKeeper copyright (which he didn't) and terms of service (which he didn't agree to).

  • by PeterBrett ( 780946 ) on Thursday April 21, 2005 @09:10AM (#12301732) Homepage the business of not storing deltas to files

    Actually, that's a feature. One of Linus' main objections to existing SCMs was that they're too slow -- and most of that sluggishness comes from the time necessary to calculate and resolve deltas in the ubiquitous RCS files.

    As far as I can tell, it's actually working pretty well so far.

  • Doubled file size!? (Score:3, Interesting)

    by Anonymous Coward on Thursday April 21, 2005 @09:23AM (#12301814)
    From TFA:
    Among the differences: Git can't rename a file; users must instead delete one and recreate it elsewhere with the new name, McVoy said. And it doesn't handle space efficiently;
    a tiny one-character change to a 1MB file in Git will result in a 2MB file, whereas BitKeeper's file will grow only by one byte.
    Does anyone know why diff wasn't used here? 1 byte difference (BitKeeper) is excellent, a 100 bytes (diff) is alright, but doubling the file size (Git) seems really insane to the untrained eye..
  • by kyb ( 877837 ) on Thursday April 21, 2005 @09:39AM (#12301910)
    Now this confuses me. The BitKeeper "free" license says that you can't use BK to work on a competing product after using BK. Isn't this exactly what Linus has just done?

    Will McVoy ask Linus's employer to sack him? He should at least threaten legal action.

  • (Score:3, Interesting)

    by zulux ( 112259 ) on Thursday April 21, 2005 @10:06AM (#12302110) Homepage Journal
    Tridge is not in the right. Tridge has demonstrated his immaturity by disregarding the effects of his actions on others

    Tridge displayed amazing foresight - he was the problem of Bitkeepers vendor lock-in and was doing something about it before Bitkeeper changed their format again.

    Truns out that Tridge was right not to trust the Bitkeeper people with the Linux kernal.

  • Erm, name change... (Score:5, Interesting)

    by Xarius ( 691264 ) on Thursday April 21, 2005 @10:06AM (#12302114) Homepage
    There is a package manager, make-based one at that, out there called git. Site in german, package in English [].

    It's a good tool, which basically monitors source-built programs and creates an uninstallation script for them.

    Won't this mean Linus' new tool will have to have a name change?
  • Re:Tridge Speaks (Score:2, Interesting)

    by Anonymous Coward on Thursday April 21, 2005 @10:08AM (#12302126)
    The problem with your method is that you can't script it. Netcat allows you to pipe input over a socket.

    Netcat is an excellent tool, incidentally. In addition to piping input it's possible to write rudimentary servers with just a shell script. It's well worth taking the 10 minutes or so needed to learn it.
  • (Score:1, Interesting)

    by Anonymous Coward on Thursday April 21, 2005 @10:14AM (#12302169)
    There is a reason why he reverse engineered the file format and not the source control system. BitKeeper stored the code in a proprietary data format, their IP is the only data capable of reading that format. If BitMover ever chose to revoke all licenses to use their IP to read the format it would bar the kernel developers from retrieving the kernel source code stored in the system.

    So by attempting to solve a problem that didn't yet exist, Tridge precipitated that very problem. One might call that self-serving.

    One minor fault in this theory -- BitKeeper withdrew its support but left enough of a grace period to permit Linus and others to withdraw their source from BitKeeper into whatever other package they wanted/needed. So the problem still doesn't exist.

    Oh, and another minor fault in this theory -- Tridge couldn't possibly have reverse engineered the file format after BitKeeper performed some evil source snatching deed?

    Oh, and one final fault with this theory -- If BitMover ever chose to revoke all licenses to use their IP to read the format it would bar the kernel developers from retrieving the kernel source code stored in the system. -- Isn't this theft and/or extortion? It strikes me that a few police complaints and a SMALL CLAIMS LAWSUIT that would take a trivial amount of time could solve that problem.

    It was a nice try at justification, but the theory doesn't fly.
  • by Anonymous Coward on Thursday April 21, 2005 @10:15AM (#12302174)
    Not only that, the license says that you can't use BK and develop a competing product, even if you don't use BK in developing that product.

    However, since McVoy has revoked the license, I don't think he can reasonably expect others to abide by its conditions perpetually.

    In any case, a condition that essentially said "if you ever use this product, you can never do X" would probably be entirely unenforcable in a software license not involving a negotiated and signed contract. Perpetual non-compete clauses aren't even considered reasonable in employment contracts (there's always a time limit), how could such a condition possibly be considered reasonably in what's effectively a software EULA?
  • (Score:3, Interesting)

    by LWATCDR ( 28044 ) on Thursday April 21, 2005 @10:17AM (#12302194) Homepage Journal
    "Tridge is not in the right. Tridge has demonstrated his immaturity by disregarding the effects of his actions on others. Tridge has the skill to write a source control system from scratch, and intentionally chose not to - all claims to moral superiority died with that decision."
    The problem I have with all of this is that the license that bitkeeper used was the least free I have ever heard of. If you used it you could not not "work" on a competing program! If Tridge had not signed that agreement how was what he did any more immoral than Open Office reverse enginering the Office file format? If Microsoft said that if you use outlook express you can not work on any other mail reader people would be after blood.
    Frankly Bitkeeper got more free press from Linus using it and say how great it is than they could ever have afforded to pay for. I admit that I have no problem with none free software but I do have a problem when I am being told what I can and can not work on! Think about it Bitkeeper was less free than Windows! Not only did you not have the source you where not allowed to write a competing program! I was all with using Bitkeeper until I heard about the none compete part. I feel everyone has the right not to give away there work but I have a problem when they tell me what I can write.
  • by Svartalf ( 2997 ) on Thursday April 21, 2005 @10:18AM (#12302206) Homepage
    I'd say that they are. In fact, several SCM tools are looking at what Linus has been doing with git and are giving the stuff due consideration (ARCH is going to pull all the "good" ideas coming from this- I think several other SCM projects are going to do the same...).

    Yeah, I think Larry just stepped in a hornet's nest here- my only complaint about the whole thing is Linus' going on and on about bad ideas, etc. The only bad idea that was going on was his use of BitKeeper in the first place.
  • by Herschel Cohen ( 568 ) on Thursday April 21, 2005 @10:33AM (#12302341) Homepage Journal
    Read the last couple of lines on this link: ux/story/0,10801,101207,00.html
    Indeed, it might be good to read the entire article, much better than I remember Computer World particularly when compared to other articles on this topic seen elsewhere.

    I found this on this morning.
  • (Score:2, Interesting)

    by Klivian ( 850755 ) on Thursday April 21, 2005 @10:35AM (#12302352)
    >and it is what most wanted anyway.
    Actually Linus and many other core kernel hackers did not want this. And the kernel developers was never forced to use BK, and the few who did not like tke BK license managed just fine whitout. What Tridge did was force his political/religious agenda onto the kernel developers. Besides BitMover did not revoke their license on a whim, they tried a reasonable dialog involving Linus. However one try to present the facts one thing remains. That is Tridge is the worst case of zealot preaching freedom of sw, but not respecting other peoples freedom to choose and work with any license they want.
  • Patents on protocols (Score:2, Interesting)

    by tepples ( 727027 ) <tepples@gmail.BOHRcom minus physicist> on Thursday April 21, 2005 @10:58AM (#12302549) Homepage Journal

    Communication protocols are methods, not works of authorship. Methods cannot be copyrighted, but they can be patented. What is your evidence otherwise?

  • by Bombcar ( 16057 ) <racbmob@bomb[ ].com ['car' in gap]> on Thursday April 21, 2005 @11:03AM (#12302616) Homepage Journal
    Heh. A quote:

    And rsync takes care of the network replication.

    So git, which had to be written because of something Tridge did, uses one of Tridge's programs. Whee!
  • Re:Tridge Speaks (Score:3, Interesting)

    by wct ( 45593 ) on Thursday April 21, 2005 @11:39AM (#12302971) Homepage

    He has only claimed to not use the bitkeeper client. There is no EULA governing his use of the publicly facing server. The argument has been that if he used the bitkeeper client to reverse engineer the protocol, he was breaking the terms of using the client. Clearly he wasn't.

    Whether it is an abuse of Bitkeeper's servers to try and interact with them without permission is another debate altogether. Does putting a server on the internet with no access restriction mechanism qualify you for legal protection against access?

  • Uh, how about every time they implement something that Apple implemented first?

    Did you miss "Stealing ideas? Sure, everyone gets accused of that, but nobody in the OSS community with any credibility is going to use that kind of phrase except in jest"?

    The only example that I can think of that might fit your claim is Windows, and the real problem with Windows was not that it was a GUI, or that it could be made to look like the Mac, since both Apple and Microsoft got the idea from Xerox. The problem with Windows (and this doesn't get nearly enough publicity) is that it duplicated the internal structure of Mac OS, even where Mac OS was doing things the wrong way. The original design of Windows, for example, was going to have preemptive multitasking from the get go. It was changed under Bill Gates' urging to use cooperative multitasking. Eller's book documents how Bill Gates acted as a hole in the "chinese wall" between the applications group (who were developing software for the Mac) and the OS group (who weren't contractually allowed to know anything about the Mac), bopping over and telling them to do what turned out to be apallingly stupid things (like said cooperative multitasking) because that's what the Mac did. THAT could be described as theft, because it involved Microsoft violating the NDA they signed when they contracted with Apple to develop applications (like Word) for the Mac.

    It was also a disaster in the long term for Microsoft, if they hadn't done it Windows might have become a viable desktop OS in the '80s instead of having to wait for Windows for Workgroups and Windows 3.11, where they were able to field the first Win32 (Win32s) APIs that let them haul themselves a bit out of the mess they'd made of the 16 bit Windows.

    All your other examples are what's technically known as "irony". It's like "Nobody wanted Multitasking until Microsoft stole it from UNIX"... nobody who says that literally means any illegal activity occurred. It's what we in the industry call humor.
  • Re:Git? (Score:2, Interesting)

    by mink ( 266117 ) <> on Thursday April 21, 2005 @12:10PM (#12303255)
    Looking at the evidence it seams thoe only sane reasonable person in the whole stupid ordeal is Tridge.

    Telnet [servername] 5000

    [servername]> help

    does not seem to me to be evil. But if it will help we can ban all forms of remote system access, to keep companies like bitkeeper safe.
  • by HelloKitty ( 71619 ) on Thursday April 21, 2005 @12:11PM (#12303264) Homepage
    Subversion is awsome:

    I started using the Subversion (SVN) source control tool at home instead of CVS (they have convert scripts to convert the old CVS repos)... I use it to store about 2GB of sound and art data as well as source code and configuration files for my system, basically anything I might want to get a previous version.

    My frontend is TortoiseCVS, which is integrated in the explorer windows - which makes for a really easy and intuitive interface.... No more synchronizing two separate windows. Just use explorer, very simple.

    It handles binary files automatically (by autodetection using heuristics on the data and through configurable mime types), and stores binary diffs so your database doesn't get bloaty. In the recent 1.2 version of SVN they've added locking to the databgase in case you're working with others on binary (not-mergable) files.

    For linux there's GUIs as well if any of you do that thing... And there's other options than just Tortoise on Windows...

    They support several database formats including Berkley DB, as well as a filesystem database (similar to what cvs used)...

    Access can be local filesystem, or remote over WebDAV (http) or over SSH (what I'm using, it's secure, encrypted, good!)...

    TortoiseSVN even comes with a graphical merge tool (TortoiseMerge).

    If you're looking for source control, I recommend Subversion. It's free, cross platform, and awesome... I would almost recommend to everyone, but I still have some features I still want to test before going that far... They still need to release 1.2 final (write-locking for binary files), and TortoiseSVN needs to support the locking feature in the 1.2 client...
  • I'd go as far as to say that really Linus has been irresponsible here by starting Git.

    Git is a hack to suit his needs, which is fine. But Git is pulling in all this publicity and development time from other people - there is basically a team of developers hacking on this hack of an SCM while more proven and mature solution (years in the making) which are only a little bit of dedication away from being ready to host a project as large as Linux are struggling to accomodate Git rather than focus on their (few) shortcomings.

    A good example is Darcs []. When this BK storm came about, the author David Roundy started on an efficiency crusade because the main obstacle to Darcs being an effective SCM for Linux was it's performance. This resulted in a number of improvements (due in the next release) and brought Darcs most of the way to being suitable for hosting the Linux source (it's distributed, easy to use, easy to merge). But now this whole Git thing has pulled everybody in different directions and recent list discussion is focused on how to accomodate Git as the backend storage mechanism for files even though it's really totally unsuited to the Darcs SCM methodology.

    If Linus had said, "Well we're ditching BK as soon as possible. I want to use either Foo or Bar but needs X and Y in them addressing first." This would have brought about the necessary improvements without this Git diversion which may turn out to be permanent to the detriment of general progress.

    He could have gone with Darcs or Monotone or even Arch and brought about their emergence as great distributed SCM tools. Instead he spent days - even weeks - hacking on Git and coordinating things through Git with manual merging. So he's wasted a lot of his own time and the time of others [hacking the hack] for the sake of creating something home grown rather than adopting something that was close to being an adequate replacement for BitKeeper.

    Great men don't always make great choices.
  • by gwait ( 179005 ) on Thursday April 21, 2005 @01:18PM (#12303826)
    Hard drives are cheap,
    deltas are slow
    (doesn't that seem like the roses are red thing?)
    besides, you can bet that git will turn from a quick LT hack into a full blown system within weeks, as a horde of programmers take it apart and put it back together.
    What gits me is - wow, are there really that many patches coming in? Sounds like a disaster waiting to happen :(
  • Using BK's servers (Score:2, Interesting)

    by thaddjuice ( 235568 ) on Thursday April 21, 2005 @01:22PM (#12303856) Journal
    One of the things that a lot of people miss is that Tridge wasn't just reverse engineering some software that reads a Word file. He was reverse engineering the protocol that BK used on their servers.

    What if Tridge wrote something that totally hosed the kernel source on BK's server? People would be screaming bloody murder at BK for letting it happen. One of the reasons BK kept their stuff closed was so they could take accountability if anything went wrong and now exactly how every client was accessing it. That's one of the advantages most managers see with going with a commercial company rather than a OSS solution.

    This isn't just copying functionality, it's putting a widely used system at risk because you don't agree with their practices. That's the same philosophy espoused by a lot of virus writers.
  • by slashdot_commentator ( 444053 ) on Thursday April 21, 2005 @01:23PM (#12303863) Journal
    I'm sorry, Linus' actions are just plain hypocritical here. I can understand how he was pissed at losing a useful tool. I can't understand how he can promote McVoy at the expense of our freedoms, especially to reverse Engineer.

    How is Linus being a hypocrite? He does not introduce himself as an official F/OSS spokesman. He never said proprietary software was intolerable that it never should be used in any circumstance.

    I can't understand how he can promote McVoy at the expense of our freedoms, especially to reverse Engineer.

    Because you're a Stalinist Communist. You are too mentally limited to listen to other points of view. Linus never attacked Tridgell's right to reverse-engineer. Torvalds attacked Tridgell's unilateral decision to scuttle the agreement he had with McVoy. I have the right to own a firearm (in the US). It doesn't mean I would be justified to conduct target practice at a shopping mall or woods near a highway. You are so concerned about your rights, you think nothing of abrogating Torvald's right to choose what software he wishes to use. Much like a Communist would.

    He is NOT indespensible, thanks to the GPL.

    Then fork, you loud-mouthed Communist.

    With all due respect to him, I think it's one of his biggest blunders in the history of Linux.

    Nope, I think his biggest blunder was prematurely releasing 2.4 before he had a stable memory manager. I think it really illustrated the amateur nature or limitations of Linux kernel development philosophy. And I'm not thrilled about many minor decisions he has made. But I equate Torvalds to Democracy. And I'll take him before you anytime, Communista.

  • It's a shame... (Score:4, Interesting)

    by [Xorian] ( 112258 ) on Thursday April 21, 2005 @04:49PM (#12306447)

    ...Linus didn't discover Vesta [] before he started implementing git. The core of the two are very similar.

    • Linus said "git is not an SCM. it's a filesystem designed to host an SCM". The Vesta repository is a filesystem. It has a versioning system (and a builder) built on top of it.
    • git's backing store is indexed by a hash of the file contents, just like Vesta's.
    • git stores complete copies of files, but only one copy of each file, just like Vesta.

    Of course there are some important differences. Like the fact that Vesta's been around for over a decade, and has been in production use for microprocessor design at Compaq and Intel for over 6 years.

news: gotcha