Re: Frames are (almost) enough for EDI

• Mail folder: SRKB Mail
• Next message: Pat Hayes: "Re: ER schemas and ontologies"
• Previous message: Fritz Lehmann: "Re: ER schemas and ontologies"
• Maybe in reply to: Pat Hayes: "Re: Frames are (almost) enough for EDI"

Date: Mon, 26 Sep 94 07:44:32 -0400
From: sowa@ranch.poly.edu (John F. Sowa)
Message-id: <9409261144.AA06748@ranch.poly.edu>
To: cg@cs.umn.edu, edi-new@tegsun.harvard.edu, srkb@cs.umbc.edu
Subject: Re:  Frames are (almost) enough for EDI
Cc: fritz@rodin.wustl.edu, pdoudna@aol.com
Sender: owner-srkb@cs.umbc.edu
Precedence: bulk

The answer to the question "Are frames enough?" is very, very simple:

   NO!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

The point is that a frame can only represent two logical operators:
existence and conjunction.  The universal quantifier is only represented
by distinguishing frame definitions from frame instances:  the definitions
of new types have an implicit universal quantifier, but frames cannot
represent disjunction, implication, or negation.  And they only represent
the universal quantifier in a highly restricted special case.

A frame seems to be general because it can represent anything that can be
put into a relational database -- that means any data whatever.  But as
we have all been discovering all along is that just one more feature is
needed:  MEANING!  There is no way in a relational database, or a frame
system, to say what the frame or the tuple means.

Unless you can express the meaning of the data in the system itself,
all you can interchange through EDI or anything else is uninterpreted bits.
That is the fundamental problem that everybody has been facing.  The
ARPA-sponsored knowledge sharing effort and the ANSI & ISO conceptual
schema effort are two projects that are directly addressing the question
of how you represent meaning.  And both of them have been converging on
the same approach and have been working together to specify it.

John Sowa