Re: Schema provider/CA lookup status

Roché Compaan <roche-kb4xc5yObU3kQYj/0HfcvhBnub05S5/[email protected]> Mon, 31 Jul 2006 17:40:58 +0200
Newsgroups gmane.comp.web.zope.plone.archetypes.devel
Organization Upfront Systems
Message-ID <[email protected]>
On Mon, 2006-07-31 at 08:32 -0700, Martin Aspeli wrote:
> 
> whit morriss-2 wrote:
> > 
> > cool...will we still be able to register a schema by interface? wouldn't 
> > the classgen stuff go inside the adapter function/class?
> > 
> > so if you didn't need the slight overhead, you could avoid it? or will 
> > all schema require this?
> > 
> > I need to look at the code but anything you could do in Schema() should 
> > be able to be done in an adapter.  hopefully Schema() will go away 
> > eventually...
> > 
> 
> I don't quite see the win in this. Schema() (well, the aq-wrapping is a bit
> annoying) is the canonical way of getting the schema of an AT object, and is
> part of the IBaseObject contract.

An obvious win is that schemas can come from different "providers",
where providers are anything that returns a schema eg. a Python module,
ATSENG, another product, etc. You need to code a pretty perfect and very
accommodating Schema() otherwise. Or maybe you mean that Schema() should
be extended to accept schemas from different providers? One then has to
think what mechanism to use to let a particular class now where to get
its schema. Maybe something similar to how Z3 fields get their default
widgets?


-- 
Roché Compaan
Upfront Systems                 http://www.upfrontsystems.co.za


-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys -- and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV