RE: [xml-dev] Tree v. Table - A relational XML object model...?

From
Bullard, Claude L (Len) <>
To
'David Megginson' <>,
Date
2003-03-03T14:19:09Z
ID
<>
Thread
RE: [xml-dev] Tree v. Table - A relational XML object model...?
We were told that the average web guy wasn't very 
bright and was very desperate.  Charles was 
being kind.

I was hoping to do away with parameter entities 
up front but was told HTML made too much use of 
them.  The not-to-be-violated requirement of SGML 
On The Web was HTML.  It not being a 
very relational-like language, and not much 
of a tree, what you see is what you got.

XML is limited to web stuff.  By design.  That 
helps to unlimit the web.

len

From: David Megginson [mailto:]

Tim Bray writes:

 > Charles Goldfarb (lead designer of SGML) actually suggested that we do 
 > this in XML, simply forbid mixed content.  

I'd be pretty suspicious of Charles on this point -- it would have
ensured that people kept using SGML for large documentation systems
and limited XML to Web stuff.

I do believe that it would be useful to have a middle-level data layer
(say, "XDL") on top of XML.  That layer could enforce both
restrictions (no mixed content) and more abstract, data-specific
conventions (such as typing, if it were actually a good idea).  Other
data oriented specs could be built on top of it.