[Zope-dev] removing zope.app from the ZTK

[Martijn Faassen]

Fred Drake wrote:
> On Tue, Dec 29, 2009 at 4:04 PM, Martijn Faassen <faassen at startifact.com> wrote:
>> But right now we need to provide some guidance for how people can move
>> away from these packages in a sane manner. And we should make sure we
>> continue to test the zope.app.* packages when we make ZTK changes, for
>> the time being.
>>
>> Let's work out a plan and a timeline.
> 
> I think we disagree as to the scale of what's needed still.

I'm looking at this from the perspective of the discontinuity we will 
introduce for existing users of the libraries that are now in the Zope 
Toolkit but were formerly presented as Zope 3, and the guidance we offer 
for people to move onto the ZTK. When we release ZTK 1.0 we need to have 
some story for this.

For this we need to have some level of commitment (as a community, and I 
think also a transition obligation of the ZTK maintainers) to maintain 
zopeapp as a backwards compatibility KGS for a period of time. We need 
to offer people and projects using the ZTK some guidance as to how to 
shift to the new way recommend and maintain (the ZTK).

In part this can be done on a per-project basis, such as for Zope 2 and 
Grok. But zopeapp is something we can maintain centrally.

I'm also worried about the large amount of Zope 3 code that's out there, 
which doesn't seem to have anyone watching out for it, possibly as 
people figured that'd be us, who are dropping that responsibility.

[snip]
> If there's anyone who wants to maintain the new bastard-stepchild,
> they're free to step forward.  There's no obligation on the part of
> the ZTK maintainers to do that, nor should there be.
> 
> That's the whole point.

On the longer term there shouldn't be any more obligation for the ZTK 
maintainers than to anyone else. That isn't *no* obligation, just like 
the Python developers feel they have some obligation not to break Python 
software even though they do not maintain this software.

In the immediate future I think the ZTK maintainers would do well not to 
forget about the historical background and compatibility constraints of 
the ZTK.

But yes, that's the point of splitting it up, indeed.

Regards,

Martijn