Re: Additional syntactic restrictions

[Gavin Thomas Nicol <gtn@xxxxxxxx>]

On Thursday 06 June 2002 11:10 am, Chris Lilley wrote:
> Well. "unlikely to be" is more accurate than "cannot", if some
> programmer figures out that the special IETF-UTF8-only-XML parser is
> re-entrant and tries to re-use it on the XML payload as well as the XML
> protocol used to send the payload because "adding two parsers to the
> product would be stupid".

This is assuming a design for an XML-based protocol that allows embedding 
arbitrary XML content, or embedding references to arbitrary XML content, in 
an XML subset, and makes assumptions about how the protocol is exposed to the 
programmer (fiddling with both protocol bits and payload bits directly). I'm 
not sure the assumptions are valid.

The whole point of a protocol is to constrain conversations/behaviours such 
that effective communication can occur...