Reviews and Standards

sowa <sowa@turing.pacss.binghamton.edu>
Date: Thu, 16 Dec 93 07:40:44 EST
From: sowa <sowa@turing.pacss.binghamton.edu>
Message-id: <9312161240.AA18479@turing.pacss.binghamton.edu>
To: cg@cs.umn.edu, interlingua@ISI.EDU
Subject: Reviews and Standards
I am thankful for the contribution by Mike Genesereth to this discussion,
since it relieves me of the responsibility for being the sole
defender of the KIF semantics, which, although I am very much in
sympathy with, I did not originally develop.  However, the
directions that I had been taking with the CG semantics were close
enough in many ways to the KIF approach that I was willing to
adopt it for the sake of harmony in the knowledge-sharing efforts.

However, I also support Matt Ginsberg's desire to have an open technical
discussion and analysis of the semantic foundations of all these
approaches.  In fact, I frequently find myself very much in sympathy
with Matt's technical points, although I believe that his anti-standards
flames are counterproductive.

Re standards:  Debating whether or not we should have standards for
knowledge interchange is irrelevant.  They are coming because they
are essential for computer applications.  The only question is whether
they are going to be defined by people who think that all we need to
do is "to add a few more features to SQL to support logic" or whether
they will take advantage of the best research that has been done
in the area.

The ANSI and ISO standards committees are about as open as you
can make them.  They are staffed by volunteers, and the quality of
the standards depends on the quality of the people who volunteer.

In principle, any US-based organization can join any ANSI committee
by paying $300 per year and sending a representative to the meetings
four times a year.  The committees are usually biased towards
companies that are making products or to large users like Boeing
and the IRS.  For a number of reasons, universities and academic
researchers are usually underrepresented:

 1. Professors get brownie points for publications, not for
    participating on standards committees.

 2. Universities do not have a major financial stake in commercial
    products that would give them an incentive to participate in
    defining standards for products.

 3. Because of reasons #1 and #2, university administrators do not see
    any reason for allocating funds to send professors to standards
    meetings and professors have no incentive to participate even if
    someone asked them to do so.

Therefore, all the issues that Matt G. is concerned about will be
ignored while the committees debate whether the SQL syntax can be
made "object-oriented".

I would very much like to see more of the academic debating have a
stronger, not a weaker influence on the standards process.  Many of
the people on those committees have a good background in computer
science and enough respect for theoretical issues that they are
willing to listen.  But if all they hear from the academic community
is bickering and flaming, they will dismiss the whole bunch and give
us another warmed over version of SQL.

Suggestion:  I think that Matt's desire for an open peer review
of the logical foundations of the semantics is well taken.  Whenever
Matt stops flaming, I have a high regard for his technical contributions
and I would very much like to see him and anyone else who may have
useful points of view to participate.  Instead of arguing about the
need for standards, let's have a serious workshop to address the
logical foundations in more depth than we can achieve in a few email
exchanges.

John Sowa