Re: [xml-dev] James Clark: XML versus the Web

[Dave Pawson <davep@dpawson.co.uk>]

When the elders sat down to think about SGML for the web in 1996
whatever, they were staring at a man-mountain, and were looking for an
athlete. Compare XML with SGML and it is an efficient subset.  Since
then, the athlete XML has been asked to do so many jobs that it has
had to develop add-ons that weren't needed in 1997 and are making it
resemble SGML again in today's world of the web. 

What we have today is a  technology that can do most of what we
ask, but is completing the circle and becoming over-sized due to all the
add-ons, sticking plasters etc., just as was SGML when viewed through
the eyes of the web. So why can't we just take the existing ideas and
mould them into another athlete fit for today's tasks?

We can, but need the motivation to do it. That motivation is referred
to in James post.  HTML5 and JSON don't need XML. It's industry and
corner cases that need the extra muscle that XML picked up from
SGML. For 80% of its usage it simply presents baggage which isn't
needed. S-XML (speedy/slim/standard) could build on James ideas,
incorporate the useful aspects of JSON (see Ricks post) and be reborn
as a 2010 athlete, again fit for purpose. 

As did XML in 1997, leave XML and its family in the corner for those
who need its bulk, and see if we can create SXML as the younger
sibling, fit and well for today's tasks.



-- 

regards 

-- 
Dave Pawson
XSLT XSL-FO FAQ.
http://www.dpawson.co.uk