LWN.net Logo

git

git

Posted Jan 26, 2006 18:50 UTC (Thu) by avr (guest, #27673)
In reply to: git by lm
Parent article: Tridge wins the 2005 Free Software Award

>And it's obviously not done because if it was then why are they working on it?
So BK development is stale ?

>actually uses these tools to do real work.
I think linux kernel development qualifies as "real work"

>Does anyone remember Linus about to burn out? Does anyone remember that BK prevented that by helping him scale?

You keep sayin that, and there might be some truth to it. But...
In my eyes what happened was that the workflow restructured itself and some kernel developers grew into positions that offloaded Linus and made the workflow more scaleable.
The most obvious example here is Andrew "the man-machine" Morton, whose mm tree, in the new role of mainline gatekeeper, has been a great positive influence on kernel development. The new 2.6 kernel-development model bears witness to this I think.
It was *people* that lightened Linus' workload and bitkeeper was the tool that may have helped to restructure the workflow.
As I see it BK was far from the is-all-end-all solution to Linus' impending burnout and claiming it was, just because it happened to be there at the time, is a very lopsided view.

>The point is that we gave something of value to the community, in good faith, and Tridge screwed that up

The point is bitkeeper introduced artificial walls inside the kernel developer community and some tried to undo that.

The whole affair was a most striking example of "dividing users and taking away their freedom." Usually we think of "users" as "end-users", but this reminds us that developers are themselves users as well. There were "haves" and the "have-nots" in the kernel developer community and the vendor used this dichotomy to exercize his influence.

>what you have is a bunch of very talented people wasting their time

I fail to see how advancing the current state of Free Software can be considered a waste of time.

>Linus made it pretty clear...
Linus is the one who made a mistake in the first place.
He assumed "The best tool for *the* job" was "The best tool for *my* job" and not "The best tool for *our* job".
As most mistakes regarding Freedom the implications are usually ill-considered and hard to undo.

>on something they had and could have continued to have for free.
The key word being "could". No hard guarantees were given that it would always be available for free regardless of the whims of the copyright holder, and look what happened.

>You tell me what's better: Linus working on the kernel or Linus working on an SCM tool.
Linus and peers working on and with Free Software.
Aside from ethical considerations, anything else would be counterproductive in the long run.

>Tridge was simply trying (and failing) to rip off our development efforts.
Or maybe he was just curious about interoperability. Apparently he even went out of his way to comply with the arbitrary terms of the license even though he had no need for one.

Wild speculation is hardly good grounds for slanderous remarks.

>You award ripoffs? What about the bum down the street stealing TVs for his drugs? Shouldn't you award that too? What's the difference?

?

>Tridge got awarded for ripping off commercial software in violation of the licensing terms.

Tridge probably got awarded for Samba and rsync and perhaps even a little for exposing the dangers of *proprietary* software by exercising his rights.
He also *allegedly* broke licensing terms regarding *proprietary* software according to the copyright holder who, instead of affirming their rights in an appropriate manner, resorted to sleazy business tactics and harassment.


(Log in to post comments)

git

Posted Jan 26, 2006 21:04 UTC (Thu) by Peter (guest, #1127) [Link]

It was people that lightened Linus' workload and bitkeeper was the tool that may have helped to restructure the workflow. As I see it BK was far from the is-all-end-all solution to Linus' impending burnout and claiming it was, just because it happened to be there at the time, is a very lopsided view.

No, he's right. Bitkeeper didn't "happen to be there at the time", that is the lopsided view. It was the enabler for Linus's new workflow. There are things you can villify Larry for, I know it's a popular pasttime, but you crossed the line into rewriting history.

Linus was a very picky customer, he hated CVS and similar systems, and Larry basically built BK around the Linus use cases. It has since matured and diversified, but originally it was to be that tool intended to allow Linus to use version control without hating it.

git

Posted Jan 26, 2006 23:01 UTC (Thu) by avr (guest, #27673) [Link]

>It was the enabler for Linus's new workflow.
An enabler is far from a complete solution.
Who knows what might have happened had the mm tree risen as the focal point of patches without bitkeeper?

>but you crossed the line into rewriting history.
A lot of credit for fixing Linus' workflow problem was lying around unclaimed, either because of modesty or because of disinterest, and bitkeeper scooped that all up and claimed it as its own.
Less pronounced dynamics in the background beside the adoption of bitkeeper were also important, just not as easily pinpointed.

I think it's not "rewriting history" but a parallax view on the whole affair. Sure bitkeeper had a role in it, but in my opinion bitkeeper is not the single silver bullet that saved kernel-development that it is made out to be.

Copyright © 2012, Eklektix, Inc.
Comments and public postings are copyrighted by their creators.
Linux is a registered trademark of Linus Torvalds