Re: Why do we have a component model?

On Mar 4, 2005, at 11:26 AM, Anish Karmarkar wrote:

> * Apologies for raising this issue at this stage *

I also apologize for being in cahoots with this.

> I would like to understand why we have the component model in the 
> spec? How and who does it help?

Me too. Thus far it hasn't helped me (qua RDF mapping editor or as user 
or as WSDL explicator to others or, as far as I can see, as 
implementor; the last is speculative since I've just planned, not 
started implementation).

> Currently we have the component model in section 2 (of part 1), 
> Infoset mapping, pseudo-schema, Z-notation and the type of the 
> properties are specified using XML Schema types. This makes it complex 
> and hard to understand. Is there any advantage to this added 
> complexity?

Plus, there are lots of tricky dependancies in the specs. The report by 
Roberto needing to touch 7 parts of the spec in order to tweak the 
model really worries me.

> Infoset is abstract and therefore can be mapped to different 
> serializations. Do we anticipate that the WSDL component model will be 
> mapped to things other than Infoset. IOW, isn't specifying things in 
> Infoset good enough?

I'd like to know how it helps/hurt creating extensions.

Cheers,
Bijan.

Received on Friday, 4 March 2005 16:41:32 UTC