BitMover Releases Open Source BitKeeper Client

diegocgteleline.es writes "Larry McVoy, the owner of BitKeeper (also one of the guys behind LMbench) has posted a message to linux-kernel where he announces a open source client of BitKeeper, which would only allow synching against BK trees. It looks like it's licensed under the NWL (No Whining License) that will force you to 'not whine about this product or any other products from BitMover, Inc.'"

Indeed...

[Svartalf]

Larry's entitled to license things under any license he wants to. It's HIS product. However, having said this, it's still quite understandable for people to not want Linux development being tied to a closed-source product with nasty gotchas in it's free license. That's not whining in the least.

The only thing resembling "whining" seems to be coming from Larry himself with this silly license. All it's going to do is make the acrimony WORSE, not better. Kind of childish, in my not so humble opinion.

Re:Strange

[tupshin]

Actually he said in the email that the whining license was a joke and he's actually licensing it as BSD (and later said it could even be considered public domain), though until the source code is re-released with proper license headers, I doubt his statement to lkml is legally binding in any way.

-Tupshin

Re:Strange

[M1FCJ]

Thinking of his track record, I wouldn't bet on his software. First you couldn't do anything, then you couldn't even use it if you were messing with other source control systems, now he is saying it is free for anything. If someone reverse engineers a GPL/BSDL BitKeeper server clone using the client will he tolerate it or will he try to crush it? That's the crux of it.

Re:Bazaar-NG

[Per Abrahamsen]

The original implementation of GNU Arch was done in bourne shell. Pyhton is a big step up from that.

In any case, I think it is a fine combination when the core functionality of a program is written in a statically typed language, and UI binding it together is written in a dynamically typed language.

Re:Bazaar-NG

[shish]

How do I use bazaar, arch or subversion to check out the kernel's bitkeeper repositories?

The point of this article is that you no longer need to use the "we own your soul" closed source BK client just to download the kernel; you can use the open source client instead.

Re:FTFA...

[stor]

Argh. Yes I should have RTFA and not posted somehting so inflammatory.

I actually meant it in the nicest possible way ;)

I like reading LKML and I like Larry and appreciate his gift but he seems to get sucked into the relentless BK flamewars and catalyse them sometimes which I think is unfortunate. He'd do himself a great service avoiding getting too involved imnsho.

I guess I didn't help just then. Ahh well. Sorry Larry et al.

Cheers
Stor

Re:Strange

[Anonymous Coward]

oi no whining!

Re:Open alternatives

[Anonymous Coward]

You never saw any problem in a one-user project.
Unfortunately you can't just multiply "zero problems times ten thousand users" and end up with 0.

- Peder

Re:Open alternatives

[koko775]

Because BitMover will itself be a "real open alternative". Why bother with subversion, arch, and (your personal favorite) darcs? Surely it's because you have the power of choice, and the freedom of source?

bitkeeper is not on my radar

[Anonymous Coward]

I don't keep my beer in my neighbor's fridge, I don't keep my money in my accountant's saving account, and I don't keep my source code in closed-source revision control systems.

Re:Open alternatives

[winchester]

There are so much real open alternatives like subversion, arch and (my personal favourite) darcs - just to name a few. Why bother with bitkeeper?

Gosh, get a clue, will you! Or read the lkml archives. Linus chose bitkeeper precisely because all the alternatives you mention don't cut it.

This is exactlythe attitude that keeps holding open source back. It's not about whether the source is open or not, it is about choosing the right tool for the right job. More people should understand this...

Re:Open alternatives

[AuMatar]

No, your attitude is the one with the problem. No current open source solution adequate? Then help make one that is- either by improving an existing alternative, or starting your own. If you don't have the skills/time to do so, encourage others who do to take it up. Just criticizing without doing anything about it helps noone.

Re:Open alternatives

[ebyrob]

Gosh, get a clue, will you! Or read the lkml archives. Linus chose bitkeeper precisely because all the alternatives you mention don't cut it.

Ya, somewhat smaller projects than the Linux kernel like Apache, Mozilla, the GCC and Debian just can't get off the ground since they don't use BitKeeper. Maybe if they switched they'd have better luck...

I mean really, it's Mr. Torvald's perogative to choose a source control tool he likes and sure when you're on someone else's court you play by their rules. But that hardly makes BitKeeper the holy grail of all source control tools.

Which is nice...

[Kjella]

No, your attitude is the one with the problem. No current open source solution adequate? Then help make one that is- either by improving an existing alternative, or starting your own.

...if you're trying to do the community a service. If you're looking to put food on the table doing something completely unrelated (where this is simply a support function), it is mindnumbingly stupid. Most likely you're long out of business by the time it is working.

Sure, if all you need is some minor customization, maybe. But if it clearly isn't anywhere near being up to the task, pick something non-OSS. Earn some money, help out the projects where it is feasible to replace proprietary with OSS.

That is the way OSS projects prospers. I make a 98% solution a 100% one. That makes it a 98% solution for someone else, who'll make it a 100% solution for them. And the snowball is rolling. Not by one company breaking its back trying to bring it from 40% to 100%.

Kjella

Re:I wonder how this bitkeeper thing compares

[leshert]

To quote The Princess Bride, "I do not think it means what you think it means."

"State of the art" doesn't mean "best overall implementation". It means that it implements the most recent advances in the field. Perforce is actually quite conventional (being originally based on either RCS or SCCS--I can't recall which). It uses the "single authoritative repository" model of version control.

The "state of the art" in version control is exploring the model of distributed and decentralized repositories. BK, darcs, arch, etc. are implementations of this model.

That being said, I like Perforce--a lot. In fact, just this year I helped successfully push for its adoption at work (beating out StarTeam and ClearCase). Perforce is fast, reliable, and not exceedingly complex for end users.

I'm not yet convinced that the distributed repository model is the best model for all purposes, but it's certainly closer to the usual meaning of "state of the art" than Perforce.

Re:Indeed...

[RedWizzard]

However, having said this, it's still quite understandable for people to not want Linux development being tied to a closed-source product with nasty gotchas in it's free license. That's not whining in the least.

Larry's view is that it's whinning because they don't have to use BK and it not using it doesn't put them in a worse position than they were in before BK was adopted. He's kind of got a point there, though it's by no means black and white. Still the "whinning" could have been a lot worse. Just imagine if Linus had of adopted a commercial system (it's not like he's religous about using open-source tools).

The only thing resembling "whining" seems to be coming from Larry himself with this silly license. All it's going to do is make the acrimony WORSE, not better. Kind of childish, in my not so humble opinion.

You might have a point, except that license is a joke. It's under the BSD license.

Re:Indeed...

[Anonymous Coward]

However, having said this, it's still quite understandable for people to not want Linux development being tied to a closed-source product with nasty gotchas in it's free license. That's not whining in the least.

The problem is that people who whined about BK being propietary should have shut their mouth up, but they didn't. When Linus switched to BK, he made clear that he would NOT force to anyone to use BK, and that's how it has been: Linux kernel.org releases are released in GNU diff format, so everybody can code and contribute. The one reason why all^Wmost of the kernel developers use BK is because they aren't stupid, BK is a great tool and can save hours of work, and it lets them to work easier between those who use BK. When someone wants to get a patch to get merged they also made the GNU diff format patch available, and even if they didn't, bkbits.net provides you a link to get a GNU diff patch for every changeset out there. Those who claim [gnu.org] that "anyone who wants to closely track patches to Linux can only do it by installing that non-free program" to develop the kernel are just wrong because you have access to the latest [bkbits.net] kernel changes without installing a non-free tool. -mm and -ac tree are maintained using open tools, so I don't see where is BK being "required".

The one reason why people whine is because they want to have the advantages of BK, but without using a propietary tool. That's not possible, there's not a OSS tool comparable to BK, subversion arch and friends are not even close. Everybody agrees that having such tool would be great (Linus even tried to convince Larry to release BK under a open license) but there's not one.

IMHO is just like when RMS had to use propietary tools to start developing GNU - Linux developers just use BK because using a OSS SCM would mean the linux kernel development would slow down a lot, and that's not good (and again, if you are going to propose to use subversion, arch, etc, it probably means that you do not understand the frenetic kernel development needs and the power of BK)

Re:bitkeeper is not on my radar

[dave420]

But you keep your beer in your own fridge, which you don't have the schematics for, so your analogy is somewhat lacking ;)

I admire your ideology-before-productivity attitude, though... inspiring.

Rule of Thumb

[R.Caley]

Always worry about a company who won't give any idea about pricing unless you get in contact with a salesdroid.

So far as I can see on their website, BitMover fall under that heading.

You got the schematics to your fridge

[SmallFurryCreature]

It is the fridge itself. From looking at a fridge any skilled craftsman can use it is as a blueprint/schematic to build another.

This is true for most real world objects. Only software is radically different.

You can also hack your own fridge all you want without dmca style rules coming into play.

So his anology works for a skilled craftsman anyway.

Re:Rule of Thumb

[base3]

The conversation, simplified, is something like this:

Customer: How much is it?

Salesman: How much do you have?

This would never be approved by OSI

[Russ Nelson]

This license would never be approved for the "Open Source" logo by OSI. If necessary, I would suggest that we change the OSD to make sure that a license does not impose restrictions on freedom of speech. Sheesh.
-russ

Unfair characterization

[FreeUser]

I admire your ideology-before-productivity attitude, though... inspiring.

That is a very unfair (and inaccurate) characterization of the grandparent post.

Productivity is only one factor, and often not the most critical one. Just as any liability lawyer, security consultant, or sysadmin whose had to recover using an offsite backup.

Your data is your most valuable possession. The cost (in time, energy, money, resources, you name it) of creating your data far outweighs the value of the hardware it resides on, the software you paid for, and probably even the office in which it resides. It is the one thing insurance can't replace, and the one thing you (or your business) probably can't live without.

Having your data (e.g. the Linux kernel) beholden to a proprietary product, managed in a proprietary format, is over the long term quite foolhardy. Imagine, for example, if Microsoft were to buy Bitkeeper (this is hardly unimaginable, and arguably not so unlikely). It isn't an "end of the world" scenerio by any means, but it is damn inconvinient to move the kernel sources to another revision control system, and unfortunately for the kernel developers, there is unlikely to be a libre one that suits their purposes available because they haven't been contributing feedback, criticisms, and suggestions for improvement to any of the free projects by virtue of the fact that they aren't using any of them and so aren't in a position to make said suggestions, etc.

It is generally a mistake to have one's data beholden to a proprietary product. Sometimes it can't be avoided, and sometimes the cost is worthwhile. And sometimes, the results are absolutely catastrophic. Unfortunately, in the case of the Linux kernel, if the results should be catastrophic in some manner, it will be catastrophci for the millions upon millions of Linux users around the world. OTOH there are enough tarballs and parallel CVS repositories around that such a scenerio is very unlikely. What isn't so unlikely is the "OMFG this is painful, we'll have to move to $free-rcs and its going to cost us at least a couple of months productivity."

Now, in the case of the Kernel, Linus has judged these risks to be small enough, and his productivity improvements to be great enough, for the potential tradeoff to be worthwhile. The grandparent post has judged the opposite. Both may be correct for their respective problem domains, but to characterize the one as "ideology-before-productivity" is very disingenuous, and ignores a whole slew of real-world issues that proprietary management schemes, formats, and restrictions bring to the table.

BitKeeper compared to Darcs.

[Anonymous Coward]

With Bitkeeper it's very easy for every organization - and even every developer - to have his own "fork" of the tree which acts as a "master repository" for others to create branches off of.

Darcs does this. http://abridgegame.org/darcs/ [abridgegame.org]

For example, within RedHat, they can have one (or many) child branches from Linus's branch (or any other developer's branches); and "reparent" the branches as needed to merge in the various pieces they need. Other employes' repositories may point to one inside RedHat; or they may point to Linus's; and of course they can "reparent" their repository to switch between the two as needed.

Darcs can do this, although the mechanics may be slightly different. I'm not sure how easy it is to do, as I have never needed to do this.

Similarly, any company or group of developers can have similar structures.

Darcs does this, too.

Also; it's important to note that not everyone needs access to a "master repository"; and that indeed no-one needs access to a "master repository" except when they're merging with that master.

The same is true of Darcs. In fact, with Darcs, you don't even need access to the master repository to send changes to the master repositroy as merges can be sent, via email, to someone who has write access to the repository. Darcs even sends the email for you.

Bitkeeper works perfectly on my laptop in disconnected mode - and I have the full power of the source control system on my laptop even with no net access - I can create branches, merge branches, etc. If I'm traveling with someone else from the company I can merge my branches with his merely with a cable between the laptops - no connection to the home office is needed.

Wow! Darcs works perfectly in disconnected mode, too! I sense a pattern!

http://abridgegame.org/darcs/ [abridgegame.org]

Consider darcs. Darcs is free, open source software.

Re:BitKeeper compared to Darcs.

[dozer]

How about full disclosure? Darcs is also really slow, an absolute pig with memory (pretty much requires 1G even for small trees) and falls over on projects even half the size of the Linux kernel. It's a a very good start, but Darcs has a long way to go before it is useful in the real world.

Re:This would never be approved by OSI

[Brandybuck]

The purpose of the license isn't to get OSI approval, but to wag appendages at uptight licensing legalists.

Re:Going too far

[Brandybuck]

Actually I am a libertarian, but like everyone else, I do have to live in the real world. Actions have consequences. If you stick your hand in a running garbage disposal, you'll pull out a stump. If you use Michael Jackson as a babysitter, you child will get buggered. If you incessantly whine about your right to incessantly whine, someone's going to beat the crap out of you.

As a liberatarian, it won't be me who will be pounding your face into a pulp, but as someone who has to live in the real world, don't expect me to be standing at the head of the line to come to your defense.

Yes, you have the right to be annoying. No, you don't have the right to prevent other people from getting annoyed.