Talk:BitKeeper

NPOV
I'm also suprised to see such an in-depth article about BitKeeper on Wikipedia. Is there any Wikipedia policy on commercial entities editing articles of commercial interest to them? How does Wikipedia intend to address the inevitable issues of differing viewpoints in an article such as this? (For example, I'm sure most commercial customers of BitMover are not interested in the aforementioned CVS server compromise, or the fact that BitKeeper is closed-source.)

If the consensus is that a BitMover representative can feasibly edit this page, I'll address some of the factual inaccuracies without affecting tone or thrust, even though those are arguably not NPOV. Andy 23:35, 14 Apr 2004 (UTC)


 * I'm the author of a fair amount of the current article, although others contributed a lot, before and since I worked on it. There isn't an official policy about "interested parties" editing a page, but I think it's something a lot of people would frown upon (it's similar to people writing golden autobiographical articles about themselves - it's really very difficult to be objective about something one cares about, however hard one tries).  Still, by saying your connected with them, and enquiring openly about the policy, you've already gained a great deal of credibility (as opposed to those who just go in and change stuff they don't like).  If it were up to me (which it isn't, but then it's not up to anyone else either) I'd prefer that declared parties either leave corrections/additions on this page, on a subpage (e.g Bitkeeper/input) or on a page in their personal userspace.  I really don't think the article is very deep at all.  It certainly needs much better technical exposition of the product itself, some better details of how the product is sold (it seems to be a more complex arrangement than just shrinkwrap) and some company history.  Most of what's there now should be in a section called something like "bitkeeper and the linux kernel".  Hmm, I did try pretty hard to keep in NPOV, and I don't think one could say it either has a "larry is evil" character, nor a "larry is a benificient angel" character either.  There's certainly an over-emphasis on linux stuff, which has some unfortunate controversy necessarily attaching, and that's something I'm keen to fix.  The trouble is that a company of BitMover's size really doesn't merit an encyclopedia article (that's a contentious issue here too, so others will surely differ) of itself, and it's the linux connection that makes it "noteworthy" (whatever the heck that means).  And yes, the last paragraph has rather little to do with Bitkeeper and BitMover, but it does go to the bitmover-linux issue.  And I've found that either I put in stuff like this, in hopefully a sane manner, or some maniac will come along and insert some craziness ("Bitkeeper has no security and is run by cretins.  Linux's only security bug was introduced by bitkeeper" or somesuch). -- Finlay McWalter |  Talk 00:09, 15 Apr 2004 (UTC)


 * Addition: yes, you're right that most commercial customers aren't too interested either in the CVS gateway (or specific issues arising from it in the past) or the licence. But then wikipedia is not a marketing brochure, so it's not really intended for commercial customers in particular.  In a more well-rounded article both of these issues would logically be confined to the linux and OpenSource section (where both are significant issues, pertaining directly to the subject at hand).  We certainly don't put lots of whiny stuff in about oracle not being open source or having lots of security holes (just as wikipedia isn't a brochure, it's not an open-source advocacy site either). -- Finlay McWalter |  Talk 00:14, 15 Apr 2004 (UTC)

How about some more meat
The "political" aspect of BitKeeper is no doubt of interest to some subset of the geek community. But honestly, shouldn't this article mostly try to explain some of the technology? Version control management is an interesting problem. For prospective users of BitKeeper, this article contains almost zero information. Psm 21:51, 2 Mar 2005 (UTC)

Have to agree with the comment above. The article says very little about what BitKeeper is and does. Instead, it reads more like a bitch session by Slashdot users. Needs some editing for bias. 203.59.80.62 (talk) 09:23, 24 May 2009 (UTC)

Seconded. It read like a soap opera script and I know nothing more about BitKeeper. — Preceding unsigned comment added by 204.228.129.18 (talk) 21:10, 21 September 2011 (UTC)

Perhaps some more context for the license arguments?
Perhaps it's worth mentioning that the license originally did not limit what other software you could work on (e.g. http://web.archive.org/web/20021012223603/http://www.bitkeeper.com/bkl.txt), and it did, in the early days, include modifiable source (iirc the only limit was that you couldn't disable openlogging). Ah, and I just found http://www.mit.edu/afs/athena/user/x/i/xiphmont/Public/critique.html, which includes an older copy of the license, too. Rweir 04:50, 17 July 2005 (UTC)

Cleanup
Can we remove this paragraph yet? pitr 14:16, 5 August 2005 (UTC)


 * "The removal of the free version makes it very likely that there will be an open source competitor within a few years, since BitKeeper's availability had previously made it unnecessary to develop one and the widely dispersed nature of much open source development provides a very strong incentive to have such a tool."


 * I have been able to convince Sun in December 2006 to make SCCS OpenSource. I still believe that the SCCS history file format is the best idea in a version control software. What we minimally need to add is a network protocol and release versioning. The file format from BitKeeper is mostly undocumented, so it is questionable whether it makes sense to try to support it. 10 years after BitKeeper it is possible to do many things better as it turned out that trying not to break the history format compatibility was wrong. Check sccs.berlios.de --Schily (talk) 21:21, 6 May 2011 (UTC)

checksums are not magic numbers
The first line of an SCCS file is a checksum. Magic numbers are essentially special constants, which are different from checksums. BK may indeed use a different algorithm for checksums, but no WP:RS is given for this statement either. TEDickey (talk) 00:26, 18 March 2012 (UTC)


 * Please try to inform yourself before commenting things.... see http://sccs.berlios.de/man/sccsfile.4.html to understand what the magic number is. --Schily (talk) 13:03, 19 March 2012 (UTC)


 * Your opinions aren't notable, and you haven't provided the requested WP:RS TEDickey (talk) 20:43, 19 March 2012 (UTC)


 * Is this your way to tell others that you don't know what you are talking about? I still recommend you to read the official documentation at http://sccs.berlios.de/man/sccsfile.4.html to understand what the magic number is in contrary to the checksum. --Schily (talk) 00:23, 21 June 2012 (UTC)


 * This is not the official man page. It claims to be SunOS, but it is not from any official SCCS version, but a fork. Some of the information in the man page is nonstandard. --91.52.46.144 (talk) 21:39, 28 June 2012 (UTC)
 * This is the official man page. --Schily (talk) 11:36, 29 June 2012 (UTC)


 * I don't see any mention of it on the standards pages such as http://pubs.opengroup.org/onlinepubs/009695399/utilities/sccs.html - perhaps you ought to inform them that they have to listen to you more carefully TEDickey (talk) 21:39, 29 June 2012 (UTC)
 * Try to inform yourself about POSIX: POSIX does not cover the file format used by SCCS, so an equivalent of the man page http://sccs.berlios.de/man/sccsfile.4.htm of course cannot be found in POSIX. --Schily (talk) 16:24, 20 July 2012 (UTC)


 * I see - then because you are asserting that there is no applicable standard, then it follows that there cannot be any official documentation, because you have not formed your own standards committee. Your comments then are purely hypothetical. TEDickey (talk) 11:38, 21 July 2012 (UTC)


 * You seem to miss that the POSIX standard defines the layout of the p. files - just the s. files have been left out in POSIX. Regarding the s. files, we need to call the AT&T documentation for the s. files "official" - I just pointed to the most recent version of that document, but the part that explains what the magic numbers are, seems to be unmodified since 1977. All real-world SCCS implementations follow that AT&T document for compliance. It seems that you just need to read that manual in order to avoid confusing magic numbers with checksums. Why don't yout read that document? Hint: for SCCS version 4 and SCCS version 5, the magic number in network byte-order is "\001\150". --Schily (talk) 09:34, 26 July 2012 (UTC)


 * As you've noted in other contexts, you have modified the documentation, and it cannot in any sense of the word be considered "official", nor authoritative for that matter TEDickey (talk) 10:50, 26 July 2012 (UTC)


 * Looks like you ned to keep the last word even after you discovered that your demur is void. Just as a note: I did not modify the documentation, I just enhanced it. --Schily (talk) 10:51, 27 July 2012 (UTC)
 * An enhancement is, by definition, a modification. --CRConrad (talk) 11:33, 10 June 2022 (UTC)

Outdated?
The box has been there for so long and there is little clarification as to what information should be included. As such I propose the removal of the 'outdated' box — Preceding unsigned comment added by Plot Citizen (talk • contribs) 19:26, 24 April 2014 (UTC)

Input from Bitmover guys?
BitKeeper developer here. I would love to update some inaccuracies of this article, but obviously anything I say is going to be viewed with suspicion. Suggestions on how to make improvements would be appreciated.

On the magic numbers, for bk-5.0 and earlier the file format uses SCCS plus additional metadata and a non-standard compression format. To prevent SCCS from being confused by our extensions the ^Ah header was changed to ^AH. This was not really a secret.

In bk-6.0 we introduced a new on-disk file format that ditches SCCS completely. This change was made for performance as it allows random access to the history and paging of really large files. In bk-7.0 the high level metadata of repositories in the ChangeSet file was changed again. This allowed dramatic performance improvements especially with very large repositories. Sorry that sounds like marketing. ;-) But it does relate to the discussion about "secret file formats". But yes these new file formats are proprietary. The point for the article is that BitKeeper originally used SCCS for storing data, but doesn't any longer.

On licensing, with the bk-7.0 release the default license was changed to drop the controversial terms and look more like other commercial software licenses.

Overall this entire article is written from the perspective of open source developers that were betrayed by the whole BitKeeper episode, but doesn't anyway say much at all about what Bitkeeper does and how it compares with other solutions. We are fine with keeping a history section that outlines the Linux kernel stuff, but it seems there should be an expanded section explaining the features provided by BitKeeper.

Wayne (talk) 13:15, 2 September 2015 (UTC)


 * As I recall it, there is very little discussion of BitKeeper features aside from primary sources (press releases are primary sources, for instance), virtually no third-party/knowledgeable reviews. Getting reliable sources to cover that is the way to proceed. TEDickey (talk) 23:45, 2 September 2015 (UTC)

External links modified
Hello fellow Wikipedians,

I have just added archive links to 1 one external link on BitKeeper. Please take a moment to review my edit. If necessary, add after the link to keep me from modifying it. Alternatively, you can add to keep me off the page altogether. I made the following changes:
 * Added archive http://web.archive.org/web/20120523134437/http://kerneltrap.org/node/4966 to http://kerneltrap.org/node/4966

When you have finished reviewing my changes, please set the checked parameter below to true to let others know.

Cheers.—cyberbot II  Talk to my owner :Online 00:16, 28 February 2016 (UTC)

External links modified
Hello fellow Wikipedians,

I have just modified 1 one external link on BitKeeper. Please take a moment to review my edit. If you have any questions, or need the bot to ignore the links, or the page altogether, please visit this simple FaQ for additional information. I made the following changes:
 * Added archive https://web.archive.org/web/20060815061854/http://better-scm.berlios.de:80/bk/ to http://better-scm.berlios.de/bk/

When you have finished reviewing my changes, please set the checked parameter below to true or failed to let others know (documentation at ).

Cheers.— InternetArchiveBot  (Report bug) 08:54, 3 November 2016 (UTC)

Evidence for "plus one year" non-compete?
I just added a "citation needed" to the following sentence in the article: "The license for the 'community' version of BitKeeper [required that] those developers did not participate in the development of a competing tool [...] for the duration of their usage of BitKeeper plus one year."

This claim surprised me because Linus Torvalds created Git almost immediately after switching off BitKeeper, with neither a one-year delay nor threats of legal action. I thought maybe I had misunderstood what "plus one year" meant (Was it the prior year? Was he forbidden from using BitKeeper for one year? etc.), so to find out, I tried to track down the original license text mentioning the one-year duration. I came up empty-handed.

In fact, I can't find any evidence for a one-year period of any sort. None of the non-compete clauses I saw mentioned it. Nobody contemporary to the events wrote about it. The only other mentions of it on the web all point back to this Wikipedia article.

Does anyone happen to have any further details? Chris Mounce (talk) 00:45, 26 May 2024 (UTC)