Re: failures

If you want to add to the list of industrial-strength applications
of semantics, look at:

  XQuery 1.0 and XPath 2.0 Formal Semantics 
  W3C Working Draft 15 November 2002

A summary of part of this work is contained in:

  The Essence of XML
  Jerome Simeon and Philip Wadler
  POPL 2003, New Orleans, January 2003. 
  [NB: Final POPL version will not be at this URL until tomorrow]

Here are two quotes from the paper:

  The act of preparing the formal semantics has uncovered a number of
  errors or omissions in the prose specification.  In particular,
  developing the material on the formal semantics of named typing led to
  the formulation of ten issues for consideration by the XQuery working
  group, each dealing with a point that was omitted in the prose
  specification of XQuery.


  In October 2002 the XQuery working group decided to adopt pure named
  typing.  The final step before adoption was a presentation describing
  how pure named typing simplifies the formal semantics.  (Pure named
  typing was adopted after this paper was accepted to POPL but before
  camera-ready copy was due, and these simplifications are incorporated
  in the version you hold in your hands.)  The presentation was followed
  by unanimous agreement to adopt pure named typing.  In the two-day
  meeting, this was the only decision adopted without dissent --- a
  resounding demonstration of the value of formal semantics!

By the way, there is a second XML standard with a formal specification
(partly inspired by our earlier work on XSLT and XML Schema).

  RELAX NG Specification
  Committee Specification 3 December 2001

All we need is a third to make it a pattern!  -- P