RE: [xml-dev] Re: determining ID-ness in XML

["Champion, Mike" <Mike.Champion@SoftwareAG-USA.com>]

> -----Original Message-----
> From: Jonathan Borden [mailto:jborden@mediaone.net]
> Sent: Friday, November 09, 2001 3:00 PM
> To: xml-dev@lists.xml.org; Christopher R. Maden
> Subject: Re: [xml-dev] Re: determining ID-ness in XML
> 
> 
> 
> So yes, you can shoot yourself in the head, but doing so is not
> recommended :-)

I completely agree.  BUT there is one use case where well-formed id-ness
declarations might be useful where this argument doesn't apply -- middleware
of some sort (SOAP-like processor, database, whatever) might find it useful
to declare the id-ness of unique key atttributes that it puts in to the XML
stream or infoset pipeline or whatever.  *IF* this were done with PIs, that
wouldn't break DTD validation.

I'm not convinced that this is a compelling reason to use PIs rather than
namespaces to declare id-nesss, but I do hope that it is taken into
consideration. 

What are the real, tangible downsides of using a PI for this?  "The W3C
doesn't like PIs because they looked ugly in the Level 3 browsers" is
presumably not a tangible downside anymore, IMHO.  I guess one is that XML
processors are not required to pass on PIs (right? or is it just that many
early implementations didn't).... and it's problematic to have them survive
an XSLT transformation ... any others?