Re: Schema provider/CA lookup status

Martin Aspeli wrote:
> 
> Rob Miller-6 wrote:
>> i'm all for concentrating on z3 schemas.  they're still a ways away from
>> being a viable substitute, though.  there are a lot of AT fields and
>> widgets out there, and a very few z3 fields and widgets that provide
>> comparable functionality.
>>
>> while i'm stoked that we get to improve things considerably in this
>> transition (b/c the z3 approach is better in many ways, and b/c we can
>> AJAX-ify our schema interfaces in the process), i think we need to
>> realistically assess how far away that is, and adequately support the
>> existing paradigms until we've got a viable replacement.
>>
>> note that i'm not making a statement about which way this should go, just
>> that we need to have a discussion about it.  i'm not convinced either way,
>> myself.
>>
>>> The functionality you describe here (schema composition through
>>> components) is probably something that you'd need to add in Z3-land as
>>> well.  And Martin says he needs it today, so maybe it should go in. The
>>> only thing I'm afraid of is that we end up with too many ways of doing
>>> things (think AT/CA vs. formlib/Z3 schema).  I'd like to hear your
>>> opinions.
>> i think that the changes should happen.  even if the Z3 schema story was
>> ready today (which it isn't), we're still going to have to interoperate w/
>> AT-schema based objects for a while.  i think we need to develop methods
>> that allow the two to live together, to ease the transition.  i want to be
>> able to extend an existing AT object by adding a Z3 field, but w/o having
>> to (necessarily) rewrite the existing schema as well.

Yeah, I can see how Whit's schema provider thing can help here.  This
really beats me though.  I have to meditate over the formlib way first
to see how those two can fit together.  Maybe Rocky or Alec have some
ideas?  I hope that 'plone.form' ripens real soon.  Then, I'd like to
see a proposal for this.

> 
> Amen, brother :)
> 
> The bigger point for me is that CMF doesn't yet have a good story for how it
> talks to Zope 3 schema based content types. For example, how does it appear
> in the 'add' menu (and you can't juse the add menu in Z3 directly, either,
> because that's Zope-instance global, not site-local as most things in CMF,
> meaning you couldn't have it installed on one CMF site and uninstalled on
> another)? How do such objects get properly cataloged? How do they interact
> with workflow?
> 
> It may not be terribly hard, but we won't get the "full" zope 3 story until
> those kinds of things are adressed. For example, the plone_schemas examples
> I've seen still derive from DynamicType (I think) and construct a full FTI
> in a dict (of course, GS makes that easier now). 

I don't think anything is wrong with using DynamicType or
CMFCatalogAware (or our own base).  When I said Zope 3 schema, I meant
the *schema*, not the way that the content type is implemented.
Cataloging and workflow aren't really part of this.  (Look at Rocky's
tutorial for an example of how to do cataloging elegantly.)

> On the other hand, the use case of listen is perfect for when Z3 schemas
> make sense right now - lightweight objects, not full CMF-ish content
> objects, that are managed with a specific UI rather than the generic CMF
> actions-and-menus system.
> 
> At least that's my perception of the state of thing right now - and I hope
> we can nudge them in the right direction. People like Rocky, Alec, Whit and
> Rob are doing great work on pushing the envelope here, and I hope we can
> start to document and push the word out a bit more as these examples mature
> (the zope.formlib tutorial, for example, is great).

While we're at this, let's also push some proposals. :-)


Daniel



-------------------------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642