[svn.haxx.se] · SVN Dev · SVN Users · SVN Org · TSVN Dev · TSVN Users · Subclipse Dev · Subclipse Users · this month's index

Re: svn 1.0 in 2006 or now?

From: Julian Fitzell <julian_at_beta4.com>
Date: 2002-12-30 12:02:37 CET

solo turn wrote:
> --- Karl Fogel <kfogel@newton.ch.collab.net> wrote:
>>solo turn <soloturn99@yahoo.com> writes:
>>>what are your plans for the 1.0?
>>>my impression after checking the task list for the milestones
>>>1.0 two months ago and now is that continuing like this will end
>>>v1.0 in year 2006 or eternity ... as the list is growing, not
>>Periodically we do go over the whole bug list and push what we can
>>past 1.0.
> may i suggest to postpone the following issues past 1.0:
> 1042,1044,1019,1034,1035,1036,1037,1046,1031,630,1015,1025,
> 510,639,1004,1008,913,1006,724,
> 495,992,837,699,919,990,933,787,869,971,977,751,774,957,951,
> 952,950,885,882,620,571,443,746,558,738,
> 838,749,773,852,959,850,650,667,997,854,986,752,991,777,
> 946,987,
> and go beta if the remaining "critical" ones are done
> (730,668,1003,891,940)?
>>>i know too many people now NOT using svn, cause it is not
>>>but this is just not true. it is already more useable than cvs,
>>>that is sufficient for most projects and use cases.
>>Feature-wise, maybe, but not in terms of stability.
> there is two kinds of stability:
> - data is safe and accessible:
> i consider it as stable, as i never heard of
> somebody who lost a repository recently.
> and berkely db is definitely known as very
> reliable. so the data store is safe.

It's probably safe, but until 1.0 there may still be times when you have
to dump the db and reload it, etc. For a user to have to keep up with
this in point releases (which they would if we went to 1.0 before these
foreseeable changes are made) seems arduous.

> - the interfaces are stable
> you can just make it stable
> but svn 2.0 does NOT need to have
> the same interfaces, otherwise

no, but svn 1.1 should, and 1.2, etc. By releasing a 1.0, you make a
certain commitment to keeping a relatively stable API in the short term.

> it will end up as one of the unusable
> unix-dinosaurs where one wrong decision
> gets dragged along for years.
> don't have the illusion of doing
> it perfect for 1.0.
>>Remember that Linux had many users before 1.0, as people gradually
>>realized that it was getting stable and that Linus was just being
> granted.
> but i still think svn has less crashes than linux. and the majorities
> of issues have nothing to do with crashes.
>>interfaces, we've missed the point. And it's not as if we're
>>suffering from a shortage of users or testers right now. Let's
>>about bugs and features, not labels :-).
> this is the only point we disagree.
> labes are important. and features is not any more important for 1.0.

I have to agree with Karl here. I mean, you're saying subversion is
basically stable enough to use right now; so why not use it? The only
reason people need to wait for 1.0 is if they want a stable interface
since (as you point out) the reliability of the data seems to be pretty
good. And if people are holding off until 1.0, waiting for a stable
interface that they won't have to put a lot of effort into updating for
each new version, then it would be misleading to release 1.0 now when we
know there are significant changes that still need to be made to achieve
the base set of features.

Anyway, that's my 2 cents. Personally, I tinker with subversion but I'm
waiting for it to stablilize in terms of API, in terms of storage
format, in terms of config files, etc. before doing any kind of major
work with it. And while it would be fabulous if 1.0 came sooner, I'm
not waiting for the code to have the 1.0 /label/, I'm waiting for it to
have a /state/ with a 1.0 level of stability.


Beta4 Productions (http://www.beta4.com)
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Mon Dec 30 12:03:25 2002

This is an archived mail posted to the Subversion Dev mailing list.