[Zope3-dev] Zope 3 release planning

[Martijn Faassen]

Stephan Richter wrote:
> On Thursday 17 February 2005 06:21, Martijn Faassen wrote:
[snip]
> The high-quality mandate of the Zope 3 development process gives our trunk 
> more stability than many other project's releases. So what are the advantages 
> of a release?
> 
> - Developers concentrate on bug fixing instead of features.
> 
> - The word of a release usually reaches beyond the Zope community, which means 
> exposure.

Don't underestimate exposure within the Zope community either.

> - There is a sense of a milestone that has a well-defined set of features and 
> API.

This is very important if you're going to develop against something -- 
see my remarks below.

> I think that all three of those are very important, but I am not sure they are 
> a reason to release prematurely. 

Zope 3 is definitely not in any risk anymore of being released 
prematurely. :)

> BTW, for Zope 3 we use the motto "Speed kills!"

That's fine, but being too slow also kills, as you'll have people drive 
past you or over you.

Anyway, one of the main advantages of a release is that *other* projects 
can do a release against those features too.

I think it's fairly common for people *not* to deploy against anything 
but releases. I typically also only use as a development platform 
something that's released, unless I control the other thing's release 
process myself. This means I don't develop against unreleased versions 
of Zope, or Python, or most other things.

I know some people like it wilder and don't mind deploying CVS versions 
of things... but it does add significant maintenance burden.

Regards,

Martijn