Nice to see XML-DEV back after its outage. Perhaps a good time to pay 
tribute to Henry Rzepa and colleagues at Imperial college who ran the 
XML-DEV list in its first three years without a glitch. We had, I 
think, 1 spam.  Tempora mutantur...

At 07:02 23/08/2006, Rick Jelliffe wrote:
>>
>I think Juan needs to look at goal # 10 for XML "Terseness is of 
>minimal importance"
>and also the goal that there should be as few optional features as possible.

Many members of this list may not appreciate the enormous communal 
effort made by the XML community in 1997++. The 10 goals of XML were 
critical and have inspired Henry and me in our development of CML - 
we adopted those which were directly relevant and added a few more 
concerned with chemistry. Of course we took XML as a given.


>Any idiot can make up a better markup language than XML, and many 
>idiots in fact do so.
>But its value comes from its being a standard.

Exactly so. No one realised this better than Jon Bosak - father of 
XML - who at an XML meeting (?XML1999?) said at a plenary something like:
" be very careful what you do as you are setting the standard for the 
next 30 years." (If anyone has the exact quote and event I'd be 
grateful).  The XML family of languages shows the whole spectrum of 
quality - I won't comment individually but we are "stuck with" XSLT, 
XSL-FO, XSD, RELAX, SAX, Namespaces, MathML, SVG... Some I love and 
some I learn to live with, without complaint.

If you are going to innovate, build on top of this family. We've done 
this for CML. We were very conscious of Jon's dictum when we built 
the language. Some of the early constructs were built in the time of 
DTDs - before dataTypes in XSD - and are ugly. Luckily we can 
"deprecate" them without too much trouble. But whenever I get a CML 
WIBNI I tell the requester that if we put it in and get it wrong it 
can never be removed. I also tell them that it has to be implemented 
and either they have to write the code or they have to persuade me to 
lose even more sleep. It's very easy to come up with WIBNIs. People 
have to fight very hard for new features.

It is very hard to develop successful scalable distributable systems. 
XML was originally "SGML on the web". It was not promoted by the W3C. 
Most people expected it to wither and die stillborn. The key to 
success was the united enthusiasm and directed energy of about 100 
active developers. Tim Bray (?at Granada) said that he has lived XML 
for ?2 years. The success arose out of the cohesiveness of the SGML 
community, the fresh blood injected by the Web, and the discipline. I 
think I saw that XML was described as the "revenge of the  over 
40's".  Among the many contributions of this list (which I'd 
recommend new members to browse through)  was the adoption of an 
IETFlike insistence that any spec had to be shown to be 
implementable. Norbert Mikola, James Clark, Tim Bray (and others 
including even me) wrote parsers that tested whether XML was 
sufficiently carefully designed and documented to be workable.

Remember also that XML was a direct descendant of SGML. SGML was 
typical first version system - over-ambitious and (I believe) never 
fully implemented in a single piece of software. Henry and I 
developed CML as an SGML DTD. Without namespace it was hideous to try 
to integrate HTML and CML, for example. XML was an enormous relief 
when the only working free software for SGML was nsgmls.

"XML is the digital dial tone of the Web" - again I think that's Jon 
Bosak.  I assume that there are now (or soon will be) chips that are 
XML-aware. I love it.

P.


Peter Murray-Rust
Unilever Centre for Molecular Sciences Informatics
University of Cambridge,
Lensfield Road,  Cambridge CB2 1EW, UK
+44-1223-763069