🚀 go-pugleaf

In article <4055@goanna.cs.rmit.oz.au> ok@goanna.cs.rmit.oz.au (Richard A. O'Keefe) writes:

   [I hope someone on the ANSI or ISO committees is keeping a copy of
   this stuff.]

Yes, I am!  I also received a lot of email on the original question of
arg/3.  As you might expect, there are several different points of view:
some support the infinite table approach, others view 
arg(N, Term, Arg)  ("extra" spaces for Richard) with N greater than the
arity of Term as a grave error, similar to exceeding the bounds of an array,
and thus must be reported.

To answer Richard's original ``attack'' on my posting:  I agree that
it is not a good idea to have a separate convention for each predicate.
However, that is the way the standard has evolved (and I don't accept
the blame for this as I have only joined recently!)  I feel that very few
people involved  would agree to redo the predicates in accordance with
your (or someone else's) design methodology.  So, one viable possibility is
to try to replace the particular definitions that we do not like.

We decided to post the arg/3 question to get feedback to that specific issue,
and also to see if people were interested in commenting on standards
in general.   In the past, Richard has been accused the standards bodies
of not caring about existing practice, suggesting that they are isolated
from the real user community.  This was a small attempt to address such
concerns.  

I got lots of feedback on arg/3 and related issues, so if the rest of 
ANSI agrees, we may post other questions in future, hopefully in a 
more structured context.

Richard, if you can send me two free plane tickets to IJCAI, I'll be willing
to forget your attack (and I'll put spaces after commas).


--
				Bernard Silver
				GTE Laboratories
				bsilver@gte.com
				(617) 466-2663