bazaar/mercurial meeting

[James Blackwell]

On Thu, May 25, 2006 at 11:27:45PM -0400, Kevin Smith wrote:
> Something along those lines would help the world understand why the two
> projects shouldn't merge. It would also help folks decide which system
> is more likely to meet their specific needs. When I move to a DVCS, it
> will probably be bzr specifically because of those first two bullet
> points. Obviously the third bullet point is critical to many folks (but
> not to me).

I can do so indirectly.  It so happens that the other day a nice person
asked me why there's so many decentralized revision control systems.
The point holds even though I get a little snippy in the last paragraph.

I have a theory (probably stated long ago in some dusty business book)
that growing fields such as 3rd gen RCS results in a set of experiments.
Each entry tests an overlapping set of assumptions such as whether dumb
transports layers are up to the task for RCS, which type of merges work
best, which storage semantics work best. So forth and so on.

Mercurial does have a widely overlapping set of tests. The stuff that
does not overlap, though, is nontrivial.

Even worse would be the control issues that would surface.  The mercurial
team is pretty smart. They would likely be to look carefully at the
previous examples of what happens when projects/teams like gnuarch, cscvs,
debian, plone, the supermirror, school tools, the ubuntu board... 

Would Bazaar-NG's primary sponsor be capable merging bazaar-ng  on even
terms with the mercurial project? Would they be able to work in a
relationship in which it didn't hold a majority stake and could dictate
terms?

Maybe they'd just try and buy out the lead mercurial members.....


>> From: "Tabitha [REDACTED]
>> To: jblack at canonical.com
>> 
>> Bazaar-NG = YAVR (Yet Another Versioning Repository System)
>> 
>> Why? Why does the world need yet another one? CVS was infantile and
>> needed to be deprecated and surely subversion is central, but gosh
>> damnit, when I look at the multitudes of versioning systems out there
>> now (svk to mirror stuff in a decentralized manner, Bazaar, Bazaar-NG,
>> darcs, mercurial, etc., etc., etc.) --> I have only so many days left
>> on this planet to live and all of this noise uncessarily caused me to
>> lose minutes of my life looking at the noise (and enough to irk me to
>> write this email). Please -- stop the noise. As Tina Turner sang "We
>> Don't Need Another Hero" ... "We Don't Need Another Versioning
>> Repository System".
>> 
>> Ugh!


> From: jblack Thu May 11 09:24:06 2006
> Date: Thu, 11 May 2006 09:24:07 -0400
> To: Tabitha McNerney
> 
> Yeah, there's a need. There's actually a need for more
> implementations.  The third gen RCS world is still in an experimental
> phase. Each implementation counts as an experiment for a set of what
> things work and what things don't.=20
> 
> There are a variety of open questions being tested by each 3rd Gen
> RCS.  Mercurial is asking whether binary blob storage is necessary.
> Monotone is asking whether or not building backwards helps. GnuArch
> asks whether changesets provide benefits. Git is asking whether or not
> three way merging is necessary. Bazaar-NG is asking if centralized and
> decentralized methodology can be combined. So forth and so on.
> 
> These questions get answered over time and the field will eventually
> stabilize. This will result, in the long term, of a shaking out of
> inferior implementations. Then everyone starts over on the 4rth gen
> systems. Some of the systems are already starting to shake out now.
> Then again, you'll probably see a couple more 3rd gen systems enter
> over the next three years.=20
> 
> I can understand if the amount of choice is confusing to you. You can
> avoid the confusion by sticking with the second generation revision
> control systems for about five years. By then the third generation
> will have stabilized and the experiment will happen in the fourth
> generation.
> 
> But lets throw fact out the window. Lets pretend that there are too
> many.  Your question still doesn't make any sense on the individual
> level -- it wold be like me deciding that we can partially solve the
> problem of there being too many people on the planet by me demanding
> that you stop breathing.



-- 
My home page:   <a href="http://jblack.linuxguru.net">James Blackwell</a>
Gnupg 06357400  F-print AAE4 8C76 58DA 5902 761D  247A 8A55 DA73 0635 7400
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 191 bytes
Desc: Digital signature
Url : https://lists.ubuntu.com/archives/bazaar/attachments/20060526/a8c44e6a/attachment.pgp