Re: simple case of IRIs for Components in WSDL 2.0

Bijan,

Thx for the explanation of the parsing problem.

And concerning my explanation, Er, yes it was to you, unless I 
misunderstood your parenthetical comment: (though I'm 
still confused about Infoset items vs. components)

Arthur Ryman,
IBM Software Group, Rational Division

blog: http://ryman.eclipsedevelopersjournal.com/
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



Bijan Parsia <bparsia@isr.umd.edu> 
10/18/2005 10:24 AM

To
Arthur Ryman/Toronto/IBM@IBMCA
cc
public-ws-desc-comments@w3.org, "Henry S. Thompson" <ht@inf.ed.ac.uk>, 
public-ws-desc-comments-request@w3.org, Jonathan Marsh 
<jmarsh@microsoft.com>, Pat Hayes <phayes@ihmc.us>, David Orchard 
<dorchard@bea.com>
Subject
Re: simple case of IRIs for Components in WSDL 2.0






On Oct 18, 2005, at 10:13 AM, Arthur Ryman wrote:

>
> Bijan,
>
> Let me explain the XML Infoset versus Component thing.

Er...to me?

> A URI-Reference is a URI and a fragment.
>
> If the URI is used to retrieve a WSDL document (as an XML Infoset) 
> then the fragment refers to some element information item (EII). That 
> is why we register the fragment syntax with the IANA media 
> registration for application/wsdl+xml.
>
> If the URI is the namespace of a WSDL document then the fragment 
> identifies a component in the WSDL component model for that namespace. 
> We sometimes refer to this case as the component designator case.

Yes.

> The two cases are consistent when the namespace actually dereferences 
> to the WSDL document, i.e. the EII maps to the component.

My sole point was that XPointer fragments don't have to point to "XML 
subresources" in the narrow sense of an infoset items. Acutally, in 
this message, my point to Pat was that such URIs were the *use* of 
non-XML formats (as well as XML formats).

> BTW, one thing I still don't understand is why the parens cause a 
> parsing problem. The XPointer syntax results in balanced parens so it 
> should be easy to parse, e.g. when the URI-Reference is the argument 
> of some function call.

The problem is that Pat wants to use parentheses as token delimiters 
*and* he wants URIs not *not* be specially quoted. So there's an 
ambiguity from his perspective. You see this some times in mail clients 
that "recognize" uris. E.g.,
                 """Boris wrote it up (http://www.example.com), check it 
out.""""

If the program is sticking to the spec the uri highlighted will be:
                 http://www.example.com), ),
Which is broken :) "Smart" clients will do:
                 http://www.example.com

via an heuristic, which you clearly can't do for a formal grammar.

So it's not the balancing, it's the using, especially as a terminating 
character.

Cheers,
Bijan.