{"thread":{"id":"23394","subject":"Git and Linux tarball size evolution","startedAt":"2010-04-09T16:33:13Z","lastAt":"2010-04-10T11:31:48Z","messageCount":2,"participants":["Victor Grishchenko","Alex Riesen"],"isPatch":false,"patchVersion":null,"patchTotal":null},"messages":[{"id":"139056","messageId":"r2l1454bf6f1004090933g4e58277dh51c731ca6b097a45@mail.gmail.com","threadId":"23394","inReplyTo":null,"subject":"Git and Linux tarball size evolution","fromName":"Victor Grishchenko","fromEmail":"victor.grishchenko@gmail.com","sentAt":"2010-04-09T16:33:13Z","receivedAt":"2010-04-09T16:33:13Z","isPatch":false,"sender":{"key":"victor.grishchenko@gmail.com","avatar":null},"body":"http://bouillon.math.usu.ru/files/linux-tarball-evol.png\n\nI plotted sizes of official linux kernel tarballs found at\nftp://ftp.kernel.org against their dates. (Yep, the methodology is\nextremely dirty.)\nLong story short: the normal dynamics of Linux development is\naccelerated growth. Apparently, in the end of 2002, development hits\nsome limit, probably a scalability problem. It returns to the\naccelerated mode in mid-2005, which event coincides with the\nintroduction of git (pointed by the arrow).\nSpeculation: indeed, the git lets development scale.\n\nIt is clear that git has changed the release pattern. But was it the\nreason why the development (and tarball size) returned to accelerated\ngrowth? Another possible interpretation is that 2.5->2.6 stage\ninvolved too much of reengineering, so \"normal\" incremental\ndevelopment slowed down for a while.\nDo git developers have any opinion on that?\nThanks!\n\n-- \nVictor\n"},{"id":"139144","messageId":"h2u81b0412b1004100431o8c543d4di7b6489211d053c98@mail.gmail.com","threadId":"23394","inReplyTo":"r2l1454bf6f1004090933g4e58277dh51c731ca6b097a45@mail.gmail.com","subject":"Re: Git and Linux tarball size evolution","fromName":"Alex Riesen","fromEmail":"raa.lkml@gmail.com","sentAt":"2010-04-10T11:31:48Z","receivedAt":"2010-04-10T11:31:48Z","isPatch":false,"sender":{"key":"raa.lkml@gmail.com","avatar":"https://avatars.githubusercontent.com/u/324101?v=4"},"body":"On Fri, Apr 9, 2010 at 18:33, Victor Grishchenko\n<victor.grishchenko@gmail.com> wrote:\n> http://bouillon.math.usu.ru/files/linux-tarball-evol.png\n>\n> I plotted sizes of official linux kernel tarballs found at\n> ftp://ftp.kernel.org against their dates. (Yep, the methodology is\n> extremely dirty.)\n\nCould you try using the sizes of *unpacked* tarballs?\ngzip/bzip2 will offset the real growth a bit (maybe even a big bit).\n\n> It is clear that git has changed the release pattern. But was it the\n> reason why the development (and tarball size) returned to accelerated\n> growth? Another possible interpretation is that 2.5->2.6 stage\n> involved too much of reengineering, so \"normal\" incremental\n> development slowed down for a while.\n\nThere were a lot of cleaning up in 2.5/2.6.\n\n> Do git developers have any opinion on that?\n\nThe was Bitkeeper before Git, but also the development process\nhas changed, with Linus becoming less of nexus of it. He does\nmore merges than ever now, with a large part of integration and\ntesting done by subsystem maintainers and people like Andrew\nMorton.\nBesides, you cannot ignore the developments outside of Linux\nworld, which percipitate into kernel (things like new architectures).\n"}]}