[Zope-dev] Re: [Plone-developers] Re: Five and 2.9

[Martijn Faassen]

alan runyan wrote:
>>>Alan, does that address your concerns?
>>
>>Just understand that the Five approach is still being developed, so
>>there's nothing to "adopt" yet. :)  But I certainly encourage folks to
>>participate and help Martijn figure out what the approach should be.
> 
> 
> Raphael, I think its great that Jim responded. Now we (collectively)
> need to get some usage out of Five.
> 
> Jim, Thanks for explaining ZC's position re: FrankenZope or backporting
> CA into Zope2.  One of the big questions that I believe is open is how to
> use more of zope.app in Zope 2 - specifically Schema/Widgets.  I heard
> that Zope Corp. is using this in some projects.

Note that FrankenZope is going to go for the new way of Interface 
integration last I heard (Five's way); the FrankenZope style setup to 
make interfaces work is not needed anymore since the changes to 
zope.interface that I did. This new approach works in straight 2.7.

I also heard rumors that FrankenZope's use of schema/widgets is actually 
an evolved branch of an earlier version of Zope 3. Five's mission is to 
integrate Zope 3 into Zope 2 as much a possible without having to change 
either. Of course this is not always possible, but it's an important 
goal. That said, I'm sure we could learn more from FrankenZope's 
experiences in that area if someone would to speak up about this.

> So - I asked if people to do some CMF/Plone implementation of Views
> in certain aspects.  I am up for creating a Supplement to Plone to bolts on
> some of the technologies and uses it.  Maybe specifically using it for
> doing in-place versioning.  I urge the communities to use Plone or maybe a
> fork of it (or the CMF) to be a playground.  Like Idle forked to land some
> major features for Python.  Then it can be integrated back into subsequent
> projects.

Such experimentation is definitely very welcome, and I'd also like to 
see some in the context of Five. I'd very much like to avoid more 
dilution of efforts where Five is going one way and CMF/Plone is going 
somewhere else.

That said, I'm certainly not going to be able to spend time on forking 
CMF and Plone and hacking on this myself. I do not know what a CMF 
version of Views would look like or even what its requirements could 
even be. It also sounds to me that forking Plone and/or CMF would be a 
far higher risk approach to integrating Zope 3 into Zope 2 than the 
approach Five is currently taking. Additionally of course I am not using 
CMF myself. :)

I'd urge people against getting too ambitious with this; I know high 
risk strategies can have higher payoff, but the chance that they fail is 
also higher. Five's strategy is baby steps and using the technology 
*right now*.

Regards,

Martijn