W3C home > Mailing lists > Public > public-xml-core-wg@w3.org > September 2004

Re: Moving towards "strict" xml:id checking

From: John Cowan <jcowan@reutershealth.com>
Date: Wed, 1 Sep 2004 12:38:25 -0400
To: Norman Walsh <Norman.Walsh@Sun.COM>
Cc: public-xml-core-wg@w3.org
Message-ID: <20040901163825.GJ15630@skunk.reutershealth.com>

Norman Walsh scripsit:

> The proposal I was trying to make when we ran out of time is the following:
>   It is an error for an ID value to be specified more than once.
>   A processor which detects this error MUST report it. Processors
>   SHOULD attempt to detect all such errors.

My objection to this is from the viewpoint of SAXish or other stream
processors, which are being asked to keep an ID table requiring O(n)
space, where n is the document length.  Currently they need only a
stack requiring space proportional to the document's nesting depth.

I'd rather see this:

   It is an error for an ID value to be specified more than once.
   Processors MAY detect and report an error and MAY recover from it.

(That's from the XML definition of "error".)

Not to perambulate                 John Cowan <jcowan@reutershealth.com>    
    the corridors                  http://www.reutershealth.com
during the hours of repose         http://www.ccil.org/~cowan
    in the boots of ascension.       --Sign in Austrian ski-resort hotel  
Received on Wednesday, 1 September 2004 16:38:26 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:16:33 UTC