Plans to port CPS on Zope 2.12 or ZTK

Hi everyone,

this is a follow up to a discussion that started on cps-users and also  
to a few direct contacts I've had in the past few weeks.

While it's rather clear that CPS is really not up to the current  
standards of the Python and Zope worlds, it's amazing that it still  
has a good reputation among  those that have experienced developing  
with it.

By « current standards », I mean at least Zope 2.12 or ZTK on python  
2.6, setuptools/distribute packaging, zc.buildout or virtualenv  
support, and of course the WSGI goodness that comes with that.
This is a bit sad, especially if one remembers a few typical Zope 3  
concepts have actually been inspired by ancestors from CPS.

Porting to a recent Zope or ZTK version would certainly be a big work.  
I'm most probably the only one around working almost full time around  
CPS, but gathering half a dozen contributors with more limited time  
dedication looks feasible from my point of view, and that'd make the  
project more than a wish.

There are lots of open questions, technical and organisational. Maybe  
it's the right time to start discussing them in the open.

Sooo… who's interested ?
In particular, what would bother you with the idea ? Stating the other  
way round : what would be the breakthrough to get you in ? No  
commitment asked at this point.

Below's my comment on the CMF dependency by the way.

Cheers,

On Aug 20, 2010, at 9:57 AM, Encolpe Degoute wrote:
>
>>> Have you plan to port CPS on Zope 2.12/2.13 or on pur Zope 3 ?
>>
>> Ahah, the big question, and you're not the first to ask. I'm all in
>> favor of a move to Zope 2.12 for CPS 3.6, but certainly  not within  
>> the
>> next three months.
>> Pure ZTK (Bluebream maybe?) would be probably even better, but I just
>> don't know if that would be possible without sacrificing too much.
>
> As CMF is not available in pure bluebream version (and no work is done
> in that way as far I know) it would be difficult to port CPS on
> bluebream without removing CMF.

You know that doesn't sound so bad actually ? Almost joking here, but  
I'd more concerned about  CPS being deeply a Zope 2 application than  
about the CMF dependency. I actually think any porting project would  
start by reducing that latter dependency.

For instance, I'm pretty sure that rewriting CPSWorkflow from scratch  
with almost no CMF dependency would really be a good thing and would  
mean less work than trying to make it work on current CMF. And it  
would save everyone, especially custom project developers, countless  
hours of painful debugging (try and follow with pdb why a transition  
does not pass if you haven't done that already).

Also, I believe that lots of those CMF concepts that are in use in CPS  
almost as-is have found their way into ZTK.

>> Now the thing is that I personally lack real experience on Zope >  
>> 2.9,
>> and probably won't get any before such works could begin. Not to say
>> that'd be a real problem for implementation, but that's a big one  
>> in the
>> foreseeing department (we can start discussing possibilities on
>> cps-devel if you wish). Also, beforehand advice wouldn't be enough:  
>> we
>> obviously need more active developers.
>
> Yes. It's a community reboot.
>
>> I know there are interested and knowledgeable people around, but no  
>> one
>> (including me) will move until it starts looking doable. If everyone
>> showed up in a discussion with no required commitment, things could  
>> very
>> well be different.
>>
>> As for the immediate future. I'm a bit tired of big jumps. (...)  I  
>> guess any real implementation work
>> would have to wait until 2011.
>
> I'm understanding that take this big project on your shoulders is  
> hard.

> I'm asking for this to know if I can motive my boos to come on this  
> project.

--
Georges Racinet, http://www.racinet.fr
Zope/CPS expertise, assistance & development
GPG: 0x4862FFF7

_______________________________________________
cps-devel mailing list
http://lists.nuxeo.com/mailman/listinfo/cps-devel