• To: "'John Cowan'" <jcowan@r...>,"'Elliotte Rusty Harold'" <elharo@m...>
• Subject: RE: XML 1.1 and other specs
• From: "Michael Kay" <michael.h.kay@n...>
• Date: Fri, 25 Oct 2002 11:05:37 +0100
• Cc: <xml-dev@l...>
• Importance: Normal
• In-reply-to: <200210221435.KAA09852@m...>
• Reply-to: <michael.h.kay@n...>

John Cowan:
> If the Infoset were an API, I would absolutely agree with 
> you.  The DOM
> *is* an API, and has the same lack of rules, which is one 
> reason why it's a poor API, as we both know.  But the Infoset 
> is only an abstract data model, and it doesn't need to 
> enforce well-formedness rules.  It's just a way of talking 
> about what it is that concrete APIs and other processes have 
> or make available.

A good abstract data model would certainly define a set of integrity
constraints, and I would expect a good abstract data model for XML to
include constraints such as "an element cannot have two attributes with
the same name".

Michael Kay
Software AG
home: Michael.H.Kay@n...
work: Michael.Kay@s... 

• References:
  • Re: XML 1.1 and other specs
    • From: John Cowan <jcowan@r...>

• Prev by Date: RE: ANN: xpath1() scheme for XPointer
• Next by Date: flush a DOM tree to a text file
• Previous by thread: XML Introduction
• Next by thread: How to print multiple spaces in my document with XSLT
• Index(es):
  • Date
  • Thread