Email message sent to Gerard Ellis on Tue 4 December, 1990.

Gerard,
	Just a few notes on the appendix I sent you yesterday.
Firstly you will have probably guessed that CoGNO is the name
of the program I wrote for my honours thesis (It stands for
Conceptual Graph Node Organiser - well I liked the acronym and that's
the best I could come up with).

+ I have used the CONGRES "subtype of <id> is <id>" statement
  to construct the concept type hierarchy.
  (this is not exactly part of the linear form).

+ As in CONGRES, I have used the dollar sign ($) to represent the
  universal quantifier.

+ Oops, I have just picked up a typographical error
  last line of CoGNO's grammar should be:
  <varlist> ::= id ["," <varlist>]
  (I have fixed this)


Comparing CoGNO's grammar to the other two presented:

+ The grammar for CONGRES is quite good though it seems that
  a few short-cuts have been taken, eg. "type" construct
  is a monadic lambda abstraction and therefore should accept
  only one parameter yet CONGRES allows <paramlist>.

+ CONGRES does not allow sets (though to be fair, CoGNO
  does not deal with them either even though I have included
  them in my YACC parser - they are neglected)

+ CoGNO allows referent [id] "#" [number] (e.g. Gerard #12345)
  while CONGRES does not (as far as I could tell from the code - 
  didn't have enough time to test this).

+ I think KRE has the simplest grammar.
  I have made newlines significant in my grammar so that
  it follows the linear form presented in the Sowa's book.
  This presents a number of implementation problems though
  since most computer languages treat newlines as white space.
  KRE gets past this problem by using a colon (:)
  separator - I didn't like this idea.
  It also allows fewer types of referent and does not
  provide for abstraction and numbering of arcs (neither
  does CONGRES - CoGNO does but the numbers are ignored
  I should have more to say about this when I send you my grammar).

That's about all I can think of at the moment - if I come up with more
I'll get back to you.

BTW, I have started modifying my YACC grammar. I hope to write
a small README file for you also to outline a few of the decisions
made, problems etc. I hope to have it finished in the next few days
(are you in a hurry? - At the latest I hope to have it done by the end of the
week ).

Maurice.
__
email: morri@csis.dit.csiro.au