some days I'm wrong

I was just looking over Sean McGrath's fine blog, and noted this:
http://seanmcgrath.blogspot.com/2003_09_14_seanmcgrath_archive.html#
10639589639410483

Compact non-XML syntaxes seem like a good - and growing idea.

Way back in 1998 I posted this, at the start of a process that turned
into XSchema/DDML:
http://www.simonstl.com/articles/xmldtd2doc.txt

I really liked the idea of using XML documents to describe XML
structures, but at this point, I'm thinking it was a serious wrong turn
in the development of schemas.  It's not necessarily a terrible idea -
XML versions of schemas are sometimes quite useful for transformations
and some kinds of processing - but compact syntaxes make a lot more
sense if you're creating schemas.

I spent the morning working with a friend of mine who called with a
"Help!  How do I write an XML Schema?" question.  We worked through the
information model he had, used RELAX NG Compact Syntax to describe it,
and everything was convenient, human-interpretable, and quite lovely.
Then, of course, we used Trang to turn it into a W3C XML Schema, but the
good part was the RELAX NG Compact Syntax.