Re: tough nuts, with primitives

• Mail folder: Interlingua Mail
• Next message: Richard Fikes: "On frames and logic (forwarding a message from Gruber)"
• Previous message: schubert@cs.rochester.edu: "tough nuts, with primitives"
• In-reply-to: schubert@cs.rochester.edu: "tough nuts, with primitives"
• References: schubert@cs.rochester.edu: "tough nuts, with primitives"

Message-id: <oaR7wwcB0KGgE3Rb5k@nero.parc.xerox.com>
Date: Mon, 11 Jun 90 22:35:56 -0700 (PDT)
From: Danny Bobrow <bobrow@parc.xerox.com>
To: interlingua@venera.isi.edu, schubert@cs.rochester.edu
Subject: Re: tough nuts, with primitives
Cc: kr-standards@vaxa.isi.edu
In-reply-to: <9006082141.AA01533@ash.cs.rochester.edu>
References: <9006082141.AA01533@ash.cs.rochester.edu>

Len has provided us with some interesting examples.  Mike noted one
interesting missing part 
-- the specification of what primitives would be allowed.  
I would like to suggest that there is another thing that is missing
 -- a specification of the use of the information.  To be more specific,
I think it would help to know:

1) what kinds of questions should it now be possible to deduce an answer
to after
 having been given the statement (typical queries would be OK, as
opposed to a real specification):

2) what kinds of questions should be EASY to answer

Example:

 a. Company ABC has been manufacturing product X for three months.

      [ABC, be-manufacturing-product, X, three, month]

Possible:  Could ABC have made this instance of product Q containing X1 
(an instance of X), that I bought four months ago?  

Easy: Was ABC producing anything last month?

The first may be a knowledge level question, and the second a symbol
level question (with an 
implied implemetnation), but perhaps not.