[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Release 1.0 and checklist: this deserves your attention.

From: Alex Perez
Subject: Re: Release 1.0 and checklist: this deserves your attention.
Date: Wed, 23 Feb 2005 13:02:36 -0800
User-agent: Mozilla Thunderbird 1.0 (Windows/20041206)

Stefan Urbanek wrote:

There were many talks about 1.0 release of GNUstep. Many of participants
argued, that the GNUstep is not ready for the 1.0 release. Can someone
who know the internals say, what exactly is needed for the release?

C'mon, guys! Stefan didn't write this e-mail just so it could be ignored :)

Please, create a check-list here:

After finished, we can discuss it and make a compromise about features
for 1.0 and 1.1.

Without concrete goal the project will be only wandering around. Also
the 1.0 is a magic version number, for some people it is the "i can use
it" point. Therefore it is more psychological.
I agree. That said, if a 1.0 release doesn't meet their expectations, then we have a serious problem.

Also the project needs its own versioning as a whole. This was discussed
too. What is gnustep now? base-1.1.0, gui-0.9.4? Are those libraries
compatible for sure? Should not I use base-0.9.4 instead to match the
gui version? Yes, you know what those versions mean, as you work with it
almost every day. But there are too many version numbers that can make
larger confusion.
I think having a version number for -core as a whole would be good. The only question is how to derive said version number.

Again, instead of saying "gnustep is not ready", fill few concrete
points into the check list. "Fix bugs", "fully follow cocoa" or "make
full implementation of OpenStep" is NOT concrete point. Not saying that
the "OpenStep" goal for GNUstep 1.0 is not relevant, as we can always
say, for example, that "GNUstep is fully OpenStep compliant from version
1.3" or "GNUstep is comliant with Cocoa in following classees: .. from
version 1.2". You get the idea.
IMHO, Cocoa class compatibility should not be a 1.0-blocker. An exception could perhaps be made for XML archives, but I'm not even sure about that.

Therefore I would suggest:
- create a very concrete checklist
- make a feature freeze
And adhere to it :)
- release 1.0 candidate
- (get publicity)
- fix bugs
- release 1.0
- (get publicity again)
- start working on 1.1
- repeat
Sounds good!

Conclusion: on one hand have long-term goal that is grand, original,
ours, perhaps achievable and call it "dream", if you like. And on the
other hand have concrete goals that can be completed in near future.

This has been what's bogged the project down in the past. Not enough "concrete milestones". The actual GNUstep devs need to weigh in on this, and not just ignore it. These decisions take time, so your first-draft doesn't need to be anything more than a draft...

Alex Perez

reply via email to

[Prev in Thread] Current Thread [Next in Thread]