for review: EDCG syntax: READ ME INSTEAD OF PREVIOUS EDCG EMAIL

[Peter Schachte]

Hi Peter,

I find some of this syntax rather confusing.  I considered most of these
issues last year and concluded that the concept of global variables pretty
well captured what I wanted (and you want) to do.  It allows a minimum of
extra syntax, and, I think, fits in better with concepts people already
have.  I suggest you take a look at my ICLP paper.  It's available from my
publications page (http://www.cs.mu.oz.au/~pets/publications.html). 

I think this approach is pretty applicable to Mercury, though you'll
probably want to add some declarations.  Don't let the name "global
variables" scare you off:  I think it does pretty much the same thing as
EDCGs.  I just think the syntax and "mental model"  are a bit nicer. 


-Peter Schachte			| We don't know who discovered water, but we
pets at cs.mu.OZ.AU		| are certain it wasn't a fish. -- John Culkin 
http://www.cs.mu.oz.au/~pets/	| 
PGP key available on request	|