RE: [xml-dev] IDs without DTD/Schema , Is there a way ?

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

I agree this is an architectural problem and won't 
go away.  It is also one we have encountered here.

1.  Reserved string: (hijack ID) feels klugy.  There 
was a long debate in the original XML design phase about 
trying to avoid magic strings.  I don't like the idea 
of hijecking a string that is very likely to show 
up elsewhere often.

2.  Reserved namespace:  I like this better. It seems 
to be more in line with what namespaces can do.  It 
does feel a bit like a PSVI or privileged properties 
approach.  However, this also appears to be a case 
where that is warranted.  I am not suggesting a 
Schema is needed, but that the association of reserved 
properties vis a vis a semantic is there.

len

-----Original Message-----
From: Tim Bray [mailto:tbray@textuality.com]
Sent: Thursday, October 25, 2001 1:16 PM
To: xml-dev@lists.xml.org
Subject: Re: [xml-dev] IDs without DTD/Schema , Is there a way ?


At 01:34 AM 25/10/01 -0700, Ronald Bourret wrote:
>There are only two ways to determine if an attribute is an ID attribute:
>
>1) From a DTD or XML Schema
>2) The attribute name is hard-coded in your application
>
>You cannot just look at an attribute and decide if it is an ID
>attribute. 

Yes, and this is one of our really big outstanding serious
architectural problems.  It's really important for the 
workings of the web that an address such as 

http://example.com/foo#Chapter12

have well-defined semantics.  If foo turns out to be XML, 
this is hopelessly underdefined.  At various times James Clark
and I have both suggested that we just brutally hijack the 
attribute name "id" and assert that it is of DTD type ID.

Other ideas have included using xml:id or having a reserved
namespace http://w3.org/xmlid or some such; any attribute 
associated with it is of type ID.

This one isn't going to go away.  -Tim