Re: Clarification on proposed requirement

On Mon, Apr 15, 2002 at 11:33:48AM -0600, Uche Ogbuji wrote:
> > > As an interested outsider, I prefer the wording you suggest, David.
> > > I think that requiring the name attributes of WSDL elements to be
> > > unique would be sufficient.  I think it would be dangerous to fix on
> > > a particular system for ensuring this, because, unfortunately, ID
> > > attributes are in a bit of practical limbo these days.  Just making
> > > this a normative clause in the WSDL spec should do the trick.
> > 
> > I doubt that IDs are dead.
> 
> I didn't say they are dead, just that their status is in a practical limbo.  
> The many tortured threads on how to fix this limbo on XML-DEV are ample fuel 
> for this opinion.

IDs are dead! Long live IDs!

> An ID technically requires a DTD or XSD declaration, neither of which are so 
> far required by WSDL.
> 
> 
> > The notion of having a schema/DTD enforced
> > document-wide unique identifier is central to many popular
> > applications, including XHTML.
> 
> Yes.  And XHTML requires DTD for this many reason (among others), and Ann 
> Navarro of the XHTML WG recently sent a message on XML-DEV affirming that 
> XHTML will have this dependence for a while.  I don't know that you want to 
> require doctype declaration or schema location specification on WSDL 
> documents, which previously did not have this restriction.

On Mon, Apr 15, 2002 at 11:41:42AM -0600, Uche Ogbuji wrote:
> For completeness, I'll mention here what has been discussed ad nauseum on
> XML-DEV.
>
> The most effective way out of this limbo is a specification for an
> instance-level ID attrbibute type declaration, such as the
>
> <?xml-id-attr "foo bar"?>
>
> or
>
> xml:id-attr="foo bar"
>
> proposals that have been made by Tim Bray, John Cowan, James Clark,
> and others.

I was assuming that writing the specification in terms of DTD IDs
would allow us to leverage off whatever mechansims were developed to
reuse them in the schema-centric world. While it is possible to write
a namespace-friendly DTD using hardcoded namespaces (witness XHTML),
it may not be particularly useful.

I guess the existing general XML processors still rely on DTDs and
there will be a pain associated with modernizing.

I would like an identifier for an attribute with semantics like DTD
IDs and I would like XML Core to promise to describe the evolution
path to an emergent recommendation.

Whatever identifier we use, it would probably be practically
constrained to work with existing, deployed XML. I expect that a
document with xml-id-attr will be rejected by any conformant XML
processor.

Defining an ID attribute in terms of either XSD or DTDs appears to
require that WSDL processors implement that (XSD or DTD)
recommendation. However, I don't think that's the case. If we adopt
the fairly popular attribute name ID we get to leverage off a lot of
XML processors that are using that convention in lieu of a
schema-friendly XML ID specification.

I think that our best bet is to declare a WSDL:ID element and specify
it in terms of either or both the DTD and XSD ID elements. The
evolution path to whatever ID mechansism we end up with will almost
certainly be defined for us, and in a way that later generic
processors will be able to work with WSDL data.

> > Whatever recommendation eventually
> > addreses this functionality will need to address XML 1.1 IDs. I would
> > like to use IDs so that our definitions addressed by such a
> > recommendation.
> 
> But you have to say exactly what IDs you mean.  DTD?  XSD?

Here I meant DTD IDs.

-- 
-eric

(eric@w3.org)
Feel free to forward this message to any list for any purpose other than
email address distribution.

Received on Monday, 15 April 2002 14:31:20 UTC