Slashdot Mirror

← Back to Stories (view on slashdot.org)

No More BitKeeper Linux

Posted by CmdrTaco on from the submission-avalanche dept.

An anonymous reader writes "KernelTrap has a lengthy article detailing BitMover's recent decision to drop support for its free version of BitKeeper. Linus Torvalds began using BitKeeper back in February of 2002, a decision that has resulted in frequent flamefests, but also in increased kernel development productivity. Evidently the recent decision was due to OSDL's decision to keep paying a developer who was working on reverse engineering BitKeeper... What tool Linus will move to is still being determined."

15 of 958 comments (clear)

  1. I cant wait by Wizy · · Score: 4, Interesting

    I cant wait for the "I told you so" articles. Lets put money on whose will be best. I have my money on Richard Stallman.

    1. Re:I cant wait by Wizy · · Score: 5, Interesting

      As another reply, this quoting Linus himself:
      "PS. Don't bother telling me about subversion. If you must, start reading up on "monotone". That seems to be the most viable alternative, but don't pester the developers so much that they don't get any work done. They are already aware of my problems ;)"

      Seems he is already looking into using it.

    2. Re:I cant wait by nathanh · · Score: 3, Interesting
      Arch and Subversion are both worthy and usable systems right now, and many projects are already working happily with one or the other.

      Subversion is entirely inappropriate. Linus prefers the distributed revision control tools. Subversion uses a centralised repository.

      Monotone, arch and svk are all options. My money's on monotone.

    3. Re:I cant wait by rpdillon · · Score: 3, Interesting

      This distopia you speak of where people are subscribing to software is in fact a utopia.

      A subscription model to software is in fact how it *should* be. Introversion, the three-man coding team in the UK that codes out of their bedrooms, just released a new game called Darwinia. In their interviews, they talked a bit about their last release, Uplink, and mentioned how odd it was to work 10s of thousands of hours with no pay, and then suddenly stop work, and get tons of money. They pointed out that this was VERY different from just about every other job model, including entertainers like rock bands, who make most of their money touring (making money as they work). Obviously, anyone who is salaried or gets hourly wage also makes the money as they work. Not so with independent developers.

      In fact, I would go even further and say the GPL lends itself to a subscription model. While I don't love Transgaming, they understand how it should work. People keep thinking that software is a "product", one that we box up and sell on shelves in a store. It isn't. Software, more than ever, is an ongoing relationship between the users and the developers. The GPL captures this, and so does a subscription model. I don't like MS anymore than the next Slashdotter, but if they are really looking at subscription, that is a good way to go. The developers get paid as they work, which means they always have an incentive to fix bugs, add features, streamline code, etc., because they continue to get paid to do so. It also allows them to release at least some of their code under GPL, because you're not just paying for the code, you're paying for the ongoing work on the code. Anytime you want to stick with where you are (keep the current version), you can. But if you want to get the most up-to-date fixes, pay for the time they spent to do that work, or, do it yourself with the old codebase.

      I honestly believe that this is the way software should work. Software should never be sold, but people can still make a living writing it. Basically, developers should sell their time and talent, not the software.

      Now, MS might be doing a model more along the lines of "You can only use Office 2017 as long as you pay...if you don't pay, no more Word for you!". This is also a subscription model, and it clearly isn't a good option. But not *all* subcription models are bad, and in fact, as I've said, some are good.

    4. Re:I cant wait by Deusy · · Score: 3, Interesting

      You might also consider Darcs, whose website also keeps a copy of the Linux source in a repo.

      It's decentralized and all that jazz. A darcs repo is hosted over http (or ssh) so it doesn't impose much in terms of hosting requirements.

      The only downside is perhaps that it's written in Haskell and that some distros don't have great support for the Haskell packages Darcs needs.

      --

      Free Gamer - Free games list and commentary

    5. Re:I cant wait by shutdown+-p+now · · Score: 3, Interesting
      This was an individual and company that doubled the output of main-line Linux development over a couple year span and the only thing asked was not to try to reverse the product.
      It wasn't about reverse-engineering the product, it was about a protocol. And yes, it is too much to ask - close your code as much as you like, but give us the specs, or at least let us reverse-engineer. Protocols must be open.
    6. Re:I cant wait by Bruce+Perens · · Score: 3, Interesting
      Divert the resources in the demonstratably finite pool of OS developers to create a tool which met the need. Divert resources to adapt to the tool as it evolves. Lose effort and suffer inefficiencies as long as the evolving tool fails to support critical requirements.

      I've been to this argument before. I remember when it was about why I should accept TrollTech's non-Open-Source license on Qt. People who did not want to accept that started GNOME. And as GNOME came along and threatened to eclipse their work, TrollTech was convinced to Open Source the Qt library. Their company literally took off with that decision. It's much, much larger now. But there never had to be a GNOME if TrollTech had only figured things out earlier.

      Bruce

      --
      Bruce Perens.
  2. Take aim at foot, Fire! by Sanity · · Score: 5, Interesting
    Wow, non-free software vendor decides to drop support for a piece of software leaving their loyal users out in the cold. Thanks BitMover for proving why Linus' decision to rely on a non-free version control system was a mistake.

    Having quickly read the RTFA, it looks like the motivation behind BitMover's hissy-fit was that a contractor of OSDL was working on reverse engineering BitKeeper's protocol in his spare time, and OSDL must have refused to, or failed to make him to stop (ouch, threatening someone's job to make them stop doing open source in their spare time, not cool!). BitMover's CEO claims to be on the side of open source, yet last time I checked interoperability was a good thing, and reverse engineering was a legitimate way to achieve it. Not according to CEO Larry McVoy, to him reverse engineering is evil, and those that do it are "bad apples" that should be punished by the rest of the open source movement.

    Of course, lots of this is my own suppositions based on reading between the lines of the article, I am sure if I have got anything wrong people will be quick to correct me.

    1. Re:Take aim at foot, Fire! by 0x461FAB0BD7D2 · · Score: 3, Interesting

      Although it may be offtopic, but non-free software vendors aren't the only ones dropping support for popular products and disappointing their loyal users. Mozilla recently did that with Seamonkey, so that they could focus on Firefox.

      User loyalty means nothing anymore. It's all about the bottom-line.

  3. In it for the money by Kevster · · Score: 5, Interesting

    Note that Larry McVoy has pointed out that the number of improvements to the commercial version due to suggestions from Open Source developers has been dropping sharply. To me, that means "giving free copies to these guys has been beneficial to my bottom line, but isn't doing much for me lately, in the financial sense". It sounds like this reverse-engineering issue is a smokescreen, a scapegoat for cutting off the "freeloaders" (those contributing to improving the product).

    So, he's in it for the money. Is anyone surprised?

    --
    I always equivocate. Well, almost always.
  4. More info by Virtual+Karma · · Score: 3, Interesting

    Read this discussion on /. for more more info: http://developers.slashdot.org/article.pl?sid=05/0 3/18/0255216&tid=156&tid=162&tid=106

    --
    fuvoo: watch something
  5. Larry never got it by nagora · · Score: 4, Interesting
    Larry goes on about how pro-open source he is but anyone that licences products with a restriction on what the users can do in their own free time is an arsehole. If MS had produced an EULA for Word that said it can't be used by people who use Acrobat Distiller, they would have rightly been scorned. Same goes for Larry and his odious BitKeeper restrictions.

    TWW

    --
    "Encyclopedia" is to "Wikipedia" what "Library" is to "Some people at a bus stop"
  6. Huh? by ABCC · · Score: 5, Interesting

    About 50 posts and nobody has suggested the possibility that M$ could have paid off Bitkeeper in a move to "hurt" linux, has everyone left their conspiracy hats at home today?

  7. Re:Too Obvious Answer by kfogel · · Score: 4, Interesting

    "Unusably slow" over a thousand files or so is totally unexpected, and not my experience at all.

    The Subversion tree itself has more than a thousand files, yet we don't have any speed problems. I'd like to know exactly what you're observing, and what might be causing it.

    keesh, would you mind describing the your slowness problems on users@subversion.tigris.org? Thanks.

    --
    http://www.red-bean.com/kfogel
  8. too slow by Fourier · · Score: 3, Interesting

    Darcs is nice, but it doesn't (yet) perform well enough for regular kernel development. The patch reordering algorithms work by loading the entire history in memory, which does not scale well to large trees.

    Darcs is, at the moment, a nice system for smaller projects.