[XML-DEV Mailing List Archive Home] [By Thread] [By Date] [Recent Entries] [Reply To This Message]

Re: Re: Syntax + object model

• To: <xml-dev@l...>
• Subject: Re: Re: Syntax + object model
• From: "Rick Jelliffe" <ricko@a...>
• Date: Mon, 26 May 2003 14:06:01 +1000
• References: <r01050400-1026-089EF5DB8D9611D7AA160003937A08C2@[192.168.124.11]>

From: "Simon St.Laurent" <simonstl@s...>

> Maybe they're just trolls, maybe they
> genuinely believe that syntax isn't interesting, but damn they're
> annoying.

Its not the people who believe that syntax is uninteresting who hinder
XML. It is the people who think syntax is unimportant, for example
see any spec which makes its own BNF for "XML". 

Actually, I would include element/attribute syntax (validity) as well
as delimiter syntax (WF) in those comments.  I think this all relates to the
software engineering principle that the code writers should not be its testers: 
because the things they missed in their code will also be the things they 
overlook in their tests. For specifications (in house or from standards bodies) 
such as data interchange or markup specifications, if you adopt the
disipline of following someone else's standard syntax (e.g. XML, Schematron,
WSDL, etc) it blocks off one class of problems, compared to using home-made
syntaxes. (Of course, it opens the opposite problem, that the standard syntax
may not express what you want, but that is just another tradeoff.)

XML may easily (if it has not already) paved the way for a pretty standard
data model for exchange (AVT with IDs with optional datatyping) and that
model may indeed turn out to have a better natural serialization syntax (if
it has not already with ASN.1). I hope people attempt such things. But that 
will not impact XML's usefullness as something distinct (or logically prior to)
a data model (future FUD notwithstanding.)

Cheers
Rick Jelliffe

• References:
  • RE: Re: Syntax + object model
    • From: "Simon St.Laurent" <simonstl@s...>

• Prev by Date: Re: XPath Data Model proposal
• Next by Date: Re: ANN: owl.dtd
• Previous by thread: Re: Syntax + object model
• Next by thread: Re: Syntax + object model
• Index(es):
  • Date
  • Thread