W3C home > Mailing lists > Public > w3c-sgml-wg@w3.org > May 1997

RE: Data types again

From: Andrew Layman <andrewl@microsoft.com>
Date: Wed, 21 May 1997 13:25:31 -0700
Message-ID: <7BB61B44F197D011892800805FD4F792A4BFAE@RED-03-MSG.dns.microsoft.com>
To: "'Paul Prescod'" <papresco@calum.csclub.uwaterloo.ca>
Cc: "'w3c-sgml-wg@w3.org'" <w3c-sgml-wg@w3.org>
You wrote:

"I think that we should start to unify and rationalize element content
and
attributes. SGML/XML has enough special cases and un-unified concepts as
it 
is. 

"I personally feel uncomfortable with the notion that attribute are only
in the language because they are "convenient". If that were the case
then
why do we treat them so differnt from content in the grove model, in our
stylesheet languages, in our query languages, in our SGML editors, etc.
In
my mind, attributes are *named data roles*, like "members" in an OOP
language, or "properties" in the grove model. As in the grove model, 
content is a special attribute, the attribute that describes the
spanning
tree that we call the "document hierarchy". Thus I consider moves to
make
attributes and elements even more different a step backwards."

I agree completely.

Similarly, I agree with your point that our thinking about datatypes is
dependent on what we do with namespaces.  Datatypes live within some
namespace.
Received on Wednesday, 21 May 1997 16:25:51 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:25:26 UTC