ciphergoth: (Default)
[personal profile] ciphergoth
Linus sold Linux's soul to the devil, and now the devil is coming to collect.

Unrelated note: I am of course going to the pub tonight!

Kernel Traffic: BitKeeper Licensing Discussion

Larry, a simple question: does the BK license allow the Rational kernel developers to use BK (to eg. check out Linus' tree) when working on kernel support for ClearCase?

ie. is all kernel development activity against your license as long as the activity is a competitor of yours?

Larry avoided the question, saying the license answered it already. Ingo asked again, "so BK cannot be used to access the kernel tree in that case, correct? I'm just wondering where the boundary line is. Eg. if i started working on a versioned filesystem today, i'd not be allowed to use BK. I just have to keep stuff like that in mind when using BK." Larry again avoided the question. Rob Landley pointed out that Larry was avoiding the question, and tried a third time to get an answer. Larry replied, "What is with you anyway? Do you have nothing better to do than try and yank my chain and cause trouble?" Rob said, "Actually, I was just hoping to prod you into answering Ingo's question..." Larry did not reply.

Date: 2002-10-23 10:50 am (UTC)
From: [identity profile] cybermuppet.livejournal.com
Oh dear.. Clearly they didn't think about that one. A Large chumk of Clearcase (MVFS to be precise) is indeed kernel-level.
Incidentally, Clearcase is worth the money any day.

Date: 2002-10-23 11:01 am (UTC)
babysimon: (compile)
From: [personal profile] babysimon
I'll believe you about "worth the money" but why on earth does it need to be kernel-level?

Date: 2002-10-23 11:38 am (UTC)
From: [identity profile] cybermuppet.livejournal.com
Put simply, Clearcase's dynamic views (effectively the developer's sandbox) are implemented with a virtual filing system called MVFS. Files and directories appearing in MVFS are in fact objects from the repository database selected according to a set of rules (the config-spec) - for instance labeled versions, named branches etc.

Dynamic views give rise to another powerful tool: Build auditing. Clearcase associates with each target of a build a complete 'bill-of-materials' used in it's creation. If you're smart, you put the build tools under clearcase too and guarantee to get the exact same versions when you go and revisit a build (for instance when you need to bugfix).

Date: 2002-10-23 11:11 am (UTC)
babysimon: (compile)
From: [personal profile] babysimon
Shit. I thought you were being overdramatic (or had given vrms your LJ password). I like to prefer to believe in incompetence rather than conspiracy. Then I clicked through.

So, when does Linus get back from his boat?

Final comment. Promise.

Date: 2002-10-23 11:26 am (UTC)
babysimon: (Default)
From: [personal profile] babysimon
So how many kernel developers are going to descend on subversion now? Maybe it'll become distributed (or indeed reach 1.0) rather sooner than we thought...

Date: 2002-10-23 01:13 pm (UTC)
zz: (Default)
From: [personal profile] zz
erm, a quick explanation for those of us who are geeky enough to be interested, but not quite geeky enough to know what this is about? <:)
From: [identity profile] dennyd.livejournal.com
BitKeeper is a versioning system of some kind (cousin to CVS et al). It is available under two licencing schemes, one of which is free as in beer, but not as in speech. Part of the 'not as in speech' bit is that you can't use it to work on other versioning systems.

Linus Torvalds, leader of the Linux kernel project, chose (in the face of some strenuous objections at the time) to use BitKeeper for managing the kernel source.

People are now concerned that either the kernel project, or open source versioning systems, or both, will suffer from this decision.

The debate continues, but it's looking like the BitKeeper move might not have been a good one.

--

Note: corrections to the above are very very welcome! as a lot of it is guesswork... I run a LUG, but I have no involvement with kernel development at all and I don't read the lkml.

Regards,
Denny
From: [identity profile] pavlos.livejournal.com
If that is true, Linus must have been deeply asleep when he made the decision. It's extremely rare for any software tool to have "may not be used for..." license conditions. It is allowed to use Micrsoft's development tools to build open source software, windows clones, or commercial competitors to visual C++ and Office. It's also allowed to use Emacs, GCC, and linux as a development environment for commercial software, Windows software (if you feel so inclined), and evil defence software. By comparison, it's amazing that Linus would have accepted to use a product with such an unusually onerous license.

Pavlos
From: [identity profile] dennyd.livejournal.com
The licence change to include this clause was more recent than the decision to use BitKeeper despite it's non-free(speech) status.

Regards,
Denny
babysimon: (westham)
From: [personal profile] babysimon
Which is reason enough to abandon BK.

I think this will blow over, Linus gives the impression of someone who's able to admit it when they've been wrong (or misinformed).
lovingboth: (Default)
From: [personal profile] lovingboth
The deepest recesses of my memory are telling me that MS software did, at one point, have such a condition.

Most commercial libraries don't allow you to build competitors with them.

Profile

ciphergoth: (Default)
Paul Crowley

January 2025

S M T W T F S
   1234
5678 91011
12131415161718
19202122232425
262728293031 

Most Popular Tags

Style Credit

Expand Cut Tags

No cut tags
Page generated Jul. 7th, 2025 08:01 am
Powered by Dreamwidth Studios