W3C home > Mailing lists > Public > www-tag@w3.org > June 2002

Re: Two messages from James Clark that ought to be read

From: <noah_mendelsohn@us.ibm.com>
Date: Mon, 10 Jun 2002 13:49:30 -0400
To: Tim Bray <tbray@textuality.com>
Cc: www-tag@w3.org, jjc@jclark.com, murata@hokkaido.email.ne.jp
Message-ID: <OFBE0AE024.E4857985-ON85256BD4.00625978@lotus.com>

James Clark writes in the first of the notes referenced by Tim [1]:

>> Certainly there is a need to be able to associate a 
>> schema with a document. We discussed this during 
>> the development of RELAX NG, but we reached the 
>> conclusion that it was better not to include this 
>> functionality as part of the schema language. 

...and then he goes on to list some of the 
good reasons...

As it turns out, we considered many of the same factors in desiging W3C 
XML schemas.  We came to very similar conclusions and for similar reasons. 
 Indeed, we took a lot of heat for not more tightly tying schemas to 
documents and/or namespaces. 

Anyway, the base architectural principle in W3C XML schema is that there 
is no fixed association between instance and schema documents.  Processors 
and applications are free to validate any document (indeed any element 
information item) against any schema appropriate to the needs of the 
application.  I infer that RELAX NG does the same.  W3C XML Schema does 
offer an xsi:schemaLocation attribute for inclusion in instance documents. 
 Its use is optional, and even if present it is only a hint.  So a 
document author can say, "here's a schema you might consider", but he or 
she cannot mandate its use. 

Regarding the need to support multiple schema languages:  the fact that 
xsi:schemaLocation is in a schema-defined namespace suggests that it is 
particular to the W3C schema language.  If RELAX NG had its own such 
attribute, you could use that in the same docuemnt to locate RELAX schemas 
(of course, there would be the issue of making each schema accept the 
others' attributes.) 
So:  I think those who worked on W3C schemas and those who worked on RELAX 
NG have arrived at somewhat similar conclusions, which is reassuring. 
(BTW:  I claim no expertise in RELAX NG details).  Thank you.

[1] http://lists.xml.org/archives/xml-dev/200206/msg00274.html

------------------------------------------------------------------
Noah Mendelsohn                              Voice: 1-617-693-4036
IBM Corporation                                Fax: 1-617-693-8676
One Rogers Street
Cambridge, MA 02142
------------------------------------------------------------------







Tim Bray <tbray@textuality.com>
Sent by: www-tag-request@w3.org
06/09/2002 04:05 PM

 
        To:     www-tag@w3.org
        cc:     (bcc: Noah Mendelsohn/Cambridge/IBM)
        Subject:        Two messages from James Clark that ought to be read



from xml-dev:

  http://lists.xml.org/archives/xml-dev/200206/msg00274.html

  http://lists.xml.org/archives/xml-dev/200206/msg00275.html

  -Tim
Received on Monday, 10 June 2002 14:08:18 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 15:32:32 UTC