LINGUIST List 2.717

Sun 27 Oct 1991

Misc: PC Dictionaries, Prolog

Editor for this issue: <>

Directory

1. "Bruce E. Nevin", PC dictionary again
2. Pamela Munro, Re: 2.587 Queries
3. Michael Covington, Re: 2.701 Computational: Shoebox, Speech Database, Prolog

A followup on the "PC dictionary" note that came up a while back.
I read in the NYT yesterday (10/22) that a Federal jury had the previous
day awarded Merriam-Webster more than $2M, plus $500,000 in punitive
damages, after finding that Random House's copycat "Webster's College
Dictionary" intentionally infringed on their design.
The socalled "PC" Random House dictionary comes in a bright red jacket
with "Webster's" in white lettering down the spine--characteristics of
the Merriam-Webster's Collegiate dictionary since 1973.
Lawyers for Random House maintained that people had mixed up the words
"collegiate" and "college" long before the lawsuit, as though that were
germane, material, pertinent, or relevant to the issue.
	Bruce Nevin
	bnbbn.com

Message 2: Re: 2.587 Queries

RE: Machine readable dictionaries
I'm sorry that I've been so swamped I haven't written to report the
two responses I received to my request for information about machine
readable dictionaries a few weeks ago. Bert Peeters wrote to suggest
Le Robert electronique (on CD Rom), and Richard (celex.kun.nl) of the
Centre for Lexical Information in Nijmegen let me know about the work his
group is doing on Dutch, English, and German. When I conveyed all this to
my former student for whom I had made the request (as I will do with any
other responses I receive) he was most grateful. Thanks to all. Pam Munro

Message 3: Re: 2.701 Computational: Shoebox, Speech Database, Prolog

Re: Prolog
Of _course_ Prolog isn't a purely declarative language. It's a procedural
language which happens to be organized around a couple of very powerful
procedures (unification, and SLR-resolution). Most programming languages
are organized around a much simpler built-in procedure: "Do the first thing
in the list, then the second, then the third..."
This means that many Prolog clauses can be read _both_ procedurally
(To prove P, prove Q and then prove R) and non-procedurally (P if Q and R).
Back in 1957, Fortran was greeted the same way. It was the first program-
ming language that allowed mathematicians to write X=Y+Z instead of
LOAD Y, ADD Z, STORE X. In that sense, Fortran was "non-procedural."
But in fact it's impossible to use Fortran without thinking about the steps
that the computer goes through, and some Fortran assignment statements,
such as X=X+1, make sense only on a procedural interpretation.
Prolog is very useful for linguistic work. The two important things to
realize very early on are (1) it _is_ procedural and you have to know what
the computer is doing; (2) whenever possible you should process symbolic
structures rather than character strings.
- Michael A. Covington internet mcovingtuga.cc.uga.edu -
- Artificial Intelligence Programs bitnet MCOVINGTUGA -
- Graduate Studies Research Center phone 404 542-0359 -
- The University of Georgia fax 404 542-0349 -
- Athens, Georgia 30602 bix, mci mail MCOVINGTON -
- U.S.A. packet radio N4TMIWB4BSG -