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

["Bullard, Claude L (Len)" <clbullar@ingr.com>]

Umm... I think one can say it is a limited 
portion of validation.  I think it is reasonable 
to say that it is a part of validation that 
might be handled separately because it has 
to be handled often and means must exist that 
are more efficient given the frequency, so 
it is a part of validation that is useful 
to split away from all the other validation 
requirements.  That is arguable, I agree.

But IDness is not part of well-formedness 
either.  XML also tossed it into the XML DTD 
box and since we had a hand in that, we 
can't simply kick SGMLAsDemonOfEra.  But 
that is spilt milk.  

The middle ground here is frequency of 
application over volume of instances at runtime.  This 
argues for a more effective means for 
declaring and locating IDs.  Again, the medium is one 
in which location resolution is a primary 
requirement, so doctrine should be to make 
that as efficient as possible given the 
overall benefit to all system users.  Since 
it is a systemic requirement, we can also 
reasonably argue that a solution which 
adds to the system vocabulary is a sound 
approach.

len

-----Original Message-----
From: Tim Bray [mailto:tbray@textuality.com]
Sent: Friday, November 02, 2001 11:52 AM
To: xml-dev@lists.xml.org
Subject: RE: [xml-dev] Re: determining ID-ness in XML


At 09:31 AM 01/11/01 -0600, Bullard, Claude L (Len) wrote:
>  What we are actually 
>up against are the limits of well-formedness, so we are 
>starting to mix in validity requirements because we don't 
>like DTDs or schemas.  Nyet.

Giving something a unique address is arguably 100% 
orthogonal to validity.  The only reason we mix these two
up is that IDs were one of the many things that got 
thrown into the DTD basket by SGML. 

>  Propose the reason why 
>the standard means don't work 

The standard means - DTDs and schemas - are overwhelingly
concerned with validation.  This is a good and useful 
thing, but it in a considerable proportion of applications 
it is not applied at run-time.  And being able to find the
node to which a pointer applies is a function which can
exist entirely independent of validation.  QED. -Tim