W3C home > Mailing lists > Public > www-ws-desc@w3.org > March 2005

Re: Why do we have a component model?

From: Arthur Ryman <ryman@ca.ibm.com>
Date: Mon, 7 Mar 2005 09:25:04 -0500
To: Bijan Parsia <bparsia@isr.umd.edu>
Cc: Anish Karmarkar <Anish.Karmarkar@oracle.com>, Jean-Jacques Moreau <jean-jacques.moreau@crf.canon.fr>, www-ws-desc@w3.org, www-ws-desc-request@w3.org
Message-ID: <OF6DF4256F.EA0D130C-ON85256FBD.004D61B8-85256FBD.004F3301@ca.ibm.com>
Bijan et al.,

I think it is possible to eliminate the component model by talking about 
documents and collections of documents. The current component model text 
could be modified slightly so that what we now call components become 
"specializations" of infoset Element Information Items. Every component 
corresponds to an element, so rather that call it a component, we can call 
it a kind of element item, e.g. Interface Item, Binding Item, instead of 
Interface Component, Binding Component.  We augment the Infoset properties 
with the additional properties that are derived from the raw XML and that 
are currently described in our spec, e.g an Interface Item has an 
[operations] property that is the set of Operation Items that are its 

In addition to the Items, we need to describe the constraints on the 
collection of infosets that correspond to all the included and imported 

The result would be that we eliminate the concept of component. There 
would be some simplifications to the text since we inherit all the infoset 
constraints. However, the spec would still be roughly as complex.

Arthur Ryman,
Rational Desktop Tools Development

phone: +1-905-413-3077, TL 969-3077
assistant: +1-905-413-2411, TL 969-2411
fax: +1-905-413-4920, TL 969-4920
mobile: +1-416-939-5063, text: 4169395063@fido.ca
intranet: http://labweb.torolab.ibm.com/DRY6/

Bijan Parsia <bparsia@isr.umd.edu> 
Sent by: www-ws-desc-request@w3.org
03/07/2005 08:49 AM

Jean-Jacques Moreau <jean-jacques.moreau@crf.canon.fr>
www-ws-desc@w3.org, Anish Karmarkar <Anish.Karmarkar@oracle.com>
Re: Why do we have a component model?

On Mar 7, 2005, at 4:01 AM, Jean-Jacques Moreau wrote:

> The main justification for the component model I can remember is that 
> it cleanly deals with import/include, i.e. it extends the Infoset 
> across files. I'm sure they are others.

Yes, this was explained at the F2F. I kinda sorta knew this but forgot. 
One possible course of action is to document these justifications so 
that people coming to the spec cold understand *why* the component 
model is there.

> I too am quite worried by the complexity of the current spec, not just 
> as the editor.

It is unclear at the moment (to me) if it is possible to reduce the 
complexity of the spec without reducing the complexity of WSDL 2.0 

Well, that's overstrong. I'm sure there are tweaks and such. But it's 
quite open whether moving away from the component model would *really* 
simplify the presentation all that much.

Bijan Parsia.
Received on Monday, 7 March 2005 14:46:22 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:54:52 UTC