Re: SYNTAX: Lexical-Functional Grammar approach

I attach two PDF files, the A3 version has the desired 2D layout, the A4
version is presumably more printable.

In a chat before the teleconference with Dave Beckett he argued, fairly
convincingly:
+ we shouldn't impose another formalism on our target developer
community.
  It may be more realistic to assume XSLT for example. 

+ he also found my working from the EBNF rules as somewhat disingeneous.
  If we are saying that they are syntactically flawed (which is the
motivation for an infoset approach), it then seems strange to motivate
our triple production on top of them.

(I hope I haven't misrepresented DaveB).


I am trying to argue in favour of a formal declarative spec for triple
production; without going to extremes (e.g. xml:lang can be added using
English text)

Jeremy