LAWS@SRI-AI.ARPA (02/22/85)
From: AIList Moderator Kenneth Laws <AIList-REQUEST@SRI-AI.ARPA>
AIList Digest Friday, 22 Feb 1985 Volume 3 : Issue 25
Today's Topics:
AI Tools - Cheap Lisp Workstations & LISPM Tape Formats,
Literature - OPS4 Book,
Policy - Dubious Humor,
Linguistics - 2nd Person Plural,
Humor - Amirsardarism,
Seminars - The Epistle Project (CLSI) &
Use of Sound to Present Data (SU) &
A Logic of Knowledge and Belief (BBN) &
Computing Conversational Implicature (BBN),
Project Description - Cognitive Complexity (IBMSJ)
----------------------------------------------------------------------
Date: Wed, 20 Feb 85 16:46:45 est
From: fischer@ru-opal (Ron Fischer)
Subject: Cheap Lisp workstations
Has anyone heard of an under $10,000 Lisp workstation with an environment
comparable to Interlisp? If you reply with a mention of a diskless
workstation, please make it obvious that the machine needs a network.
(ron)
------------------------------
Date: 21 February 1985 11:13-EST
From: George J. Carrette <GJC @ MIT-MC>
Subject: LISPM Tape Formats
It seems this question comes up a lot. We don't have any set policy on
tape formats but LMI customers have been provided with the following
support when the need came up: (+ means comes with the system already).
Machine Format Supported
LMI-Backup TOPS-20-DUMPER ANSI-LABELED TAR FIXED-EBCD(IBM)
LMI-LAMBDA + * * * *
VAX/VMS * + *
UNIX * * + + (dd)
The LMI-Backup is similar to ANSI-LABELED in use of file and tape
marks, except lisp-like in its use of LABELS (disembodied plists,
parsed by READ). As I recall JIM at Tycho was able to read it easily
on his 3600. On the other hand the various Symbolics formats looked
considerably hairier and are probably covered by their trade-secret
policy, which is why we didnt try to reverse engineer it and provide
support for it on the LAMBDA, even though people are always asking us to
move code from 3600's to LAMBDAs. But either they usually have some
kind of VAX or Unix handy to make a tape or they have a
non-industry-standard 1/4-inch tape, in which case the easiest thing
is to find somebody with a 3600<->VAX settup to make a copy for you.
-gjc
------------------------------
Date: 18 February 1985 1131-EST
From: Lee Brownston@CMU-CS-A
Subject: book on OPS4 programming
[Forwarded from the CMU bboard by Laws@SRI-AI.]
"Programming Expert Systems With OPS5," by Lee Brownston, Elaine Kant,
Robert Farrell, and Nancy Martin, is scheduled to be published by
Addison-Wesley on April 15th. It is about 400 pages long, and contains
an introduction to production systems, a tutorial on the OPS5 language,
an extended example of program development, treatment of control, data
representation, and programming style, the RETE pattern-match algorithm
and how to exploit it for efficiency, a general discussion of production
system architectures, a survey of applications, a comparison of related
production-system languages, solved exercises, and much, much more.
The authors are now correcting page proofs and making an index. Things
seem to be on schedule. For those who can't wait until publication,
copies of the page proofs are available from Addison-Wesley Publishing
Company, Inc., Reading, MA, 01867.
------------------------------
Date: Wednesday, 20 Feb 1985 08:45-EST
From: mcc@Mitre-Bedford
Subject: Polly Nomial
I was both surprised and dismayed to find "The Adventures of Polly
Nomial," which is a story about a rape, in AIList. It saddens me to
realize that there are people who think there is something "humorous"
about rape, no matter how clever the description. That this "gem" is
still around proves that misogyny still is, too.
mcc@MITRE-BEDFORD
------------------------------
Date: Thu, 21 Feb 85 12:35 MST
From: May%pco@CISL-SERVICE-MULTICS.ARPA
Subject: Polly Nomial is Offensive
I am offended at the insensitivity of stories like this and ask the
chairman to deny publication in the future. Whether it is talked about
directly or disguised in a "cute" story, rape is a violent and cruel act
by one human being against another. When it happens to someone you
know, you begin to appreciate the horror of it all.
Sexual fantasies can be fun (and healthy) but not when they are at the
expense of another.
Bob May
------------------------------
Date: Thursday, 21 February 1985 13:18:22 EST
From: Purvis.Jackson@cmu-cs-cad.arpa
Subject: 2nd person plural
I have noticed several posts on the AI-list bboard recently that address the
use of alternative plural forms of the pronoun "you" in various
socio-geographical segments of America. Specifically, I recall references
to "y'all" and "youse". Growing up in South Carolina, I frequently heard
the former; while living for the past 5 years in Pittsburgh, I have
frequently heard the latter. For both of these forms, however, I have
noticed 2 pronunciations. "Y'all" is often pronounced "yaw-ul" through the
proper application of the South Carolina low country diphthong. "Youse" is
sometimes pronounced "yooze" around Pittsburgh. None of these forms,
however, seems to me to be as interesting as a form I often heard used in
South Carolina low country "geechie" English, a mix of low country standard
and gullah. In this strain, "you" becomes "yennuh" and is oftentimes barely
distinguishable as a single word because it can be imbedded in phrases that
are delivered with a rapidly rythmic tongue movement. Hence, the phrase
sounds somewhat like one long word. An example, one that I recall quite
clearly, was delivered by Mum Tweedie DeLee, a noted root doctor of
Dorchester County whose practice of midwivery yielded me and three of my
siblings.
On the day of the particular utterance, I and several of Mum Tweedie's
great-grand children were in the back dooryard of her shanty, poking sticks
through a chicken wire fence at a goose, which would peck at the sticks,
flap its wings, and hiss angrily, all of which provided for us a somewhat
frightening source of squeeling glee. Several times, Mum Tweedie came out
onto the stomp and warned us to stop picking at the goose. Each time we
dropped our sticks and pretended to take up a new form of intertainment. On
what must have been the fourth or fifth time on the stomp, Mum Tweedie
charged down the steps and grabbed Yockey, one of the older boys, by the ear
and damned near lifted him clear of the ground. With his attention focused
fully on her, she bent over and placed her mouth close to his firmly pulled
ear and shouted into it "Yennuhadduhbedduhmine" and then let loose his ear.
She then proceeded to use our goose sticks on our backsides to ensure we had
understood her point. After we had received our just punishment, cried for
a spell, and were standing around sniffling, Mum Tweedie called us into the
shanty. Once inside she gathered us around her where she could stroke our
heads and soothe us with whispered words of love laced into further warning.
"Yennuhadduhbedduhmine," she whispered, "Yennuhbedduh." Mum Tweedie lived
to be 117 years old, by all estimates, and I have tried to follow her advice.
------------------------------
Date: Mon 18 Feb 85 10:06:46-PST
From: Shawn Amirsardary <SHAWN@SU-SCORE.ARPA>
Subject: amirsardarism
[Forwarded from the Stanford bboard by Laws@SRI-AI.]
Wow! Thanks Evan for making my contribution to the English language
possible.
Actually the amirsardarism phenomenon comes from having a sufficiently
different set of 'cognitive' rules. Forward amirsardarism is the
application of these rules in a forward chaining paradigm--and hence the
necessary blurting out of irrelevent conclusions. While
reverse-amirsardarism is the application in a reverse or backwards chaining
paradigm. I do believe that reverse-amirsardarism is the worst of the two
since the aforementioned rule set forms logically consistant conclusions in
the reverse mode, which makes it hard to spot. While in the forward mode,
the conclusions are sufficiently illogical to make for good sarcasm.
--Shawn Forward-Amirsardary
(Another application of forward-amirsardarism:
All tykes on bikes should be offered 5 Dollars, then shot
)
------------------------------
Date: Wed 20 Feb 85 09:57:11-PST
From: Dikran Karagueuzian <DIKRAN@SU-CSLI.ARPA>
Subject: Seminar - The Epistle Project (CLSI)
[Forwarded from the Stanford CSLI bboard by Laws@SRI-AI.]
THE EPISTLE PROJECT
Martin Chodorow
Yael Ravin
IBM
T.J. Watson Research Center
Date: Friday, February 22
Time: 2 p.m.
Place: Reading Room,
Ventura Hall.
This talk will be an overview of the EPISTLE natural language processing
system, especially as applied to text-critiquing. The system has four major
components: a text pre-processor, a dictionary, a parser and a set of style
rules. We will describe two implementations that represent different
approaches to parsing and will discuss the style component in detail.
In the second part of the talk we will describe semi-automatic techniques
for enhancing the semantic information contained in the dictionary. The
results of this work will provide the foundation for additional applications
in other areas, such as document abstracting or machine translation.
------------------------------
Date: 19 Feb 85 2133 PST
From: EJS@SU-AI.ARPA
Subject: Seminar - Use of Sound to Present Data (SU)
[Forwarded from the Stanford bboard by Laws@SRI-AI.]
User System Ergonomics
A human interface journal club and discussion group
On Wednesday February 27th, 12:00 - 1:00 in Margaret Jacks Hall room 352
Stanford University
Dr. Sara Bly from Xerox will speak:
"Beyond Vision: Using Sounds in the Interface"
This talk will focus on the use of sound to present data information.
Multivariate, logarithmic, and time-varying data provide examples for
aural representation. Experiments have shown that sound does convey
information accurately and that sound can enhance graphic presentations.
Methods will be discussed and examples given.
Contact Ted Selker ejs@su-ai.arpa for information on USE.
------------------------------
Date: 20 Feb 1985 16:25-EST
From: Brad Goodman <BGOODMAN at BBNG>
Subject: Seminar - A Logic of Knowledge and Belief (BBN)
[Forwarded from the MIT bboard by SASW@MIT-MC.]
BBN Laboratories Artificial Intelligence Seminar Series
Speaker: Dr. David Israel
BBN Laboratories
SRI International & CSLI
Title: "A LOGIC OF KNOWLEDGE AND BELIEF
For Logically Omniscient Yuppies (and other extroverts)"
Date: Friday, March 1, 1985, 10:30 a.m.
Location: 3rd Floor Large Conference Room
BBN Laboratories Inc.
10 Moulton Street
Cambridge, MA
Having done extensive research on the subject--to wit, having read the
Newsweek cover story "The Year of the Yuppie"--I discovered that
orthodox epistemic/doxastic logics had it all wrong. (I assume, of
course, that such logics were meant to apply, inter alia, to Yuppies.)
But the crucial problems lie not in the ascription of logical
omniscience. Nay; it is in the attributions of (pale and sickly)
introspection. Yuppies who worry too much about about inner states,
their own or others', don't get to own BMW's. I shall offer this key
demographic cohort an epistemic/doxastic logic smartly tailored to suit
their needs.
------------------------------
Date: 20 Feb 1985 09:50-EST
From: Brad Goodman <BGOODMAN at BBNG>
Subject: Seminar - Computing Conversational Implicature (BBN)
[Forwarded from the MIT bboard by SASW@MIT-MC.]
BBN Laboratories Artificial Intelligence Seminar Series
Speaker: Julia Hirschberg
University of Pennsylvania
Title: "Computing Conversational Implicature"
Date: Tuesday, February 26, 1985, 10:30 a.m.
Location: 3rd Floor Large Conference Room
BBN Laboratories Inc.
10 Moulton Street
Cambridge, MA
Determining what an utterance conveys, beyond its semantic import, is an
important issue in Natural Language Processing. Such research seeks to
provide a principled basis for computational models of human behavior --
and so to support more natural computer-human interaction. To date,
however, results have been limited by the lack of formal representations
of nonconventional inferences and the consequent difficulty of
constructing algorithms for their calculation. The work discussed
examines one class of such inferences, scalar implicature, a type of
Gricean generalized conversational implicature. It proposes a theory of
scalar implicature based upon an analysis of naturally occurring data.
A formal representation of scalar implicature is described as well as
procedures for calculating licensed implicatures. An application to
computer-human question-answering - now being implemented in Prolog - is
discussed, as are other potential uses in Natural Language generation
and understanding.
------------------------------
Date: 20 Feb 85 1348 PST
From: Terry Winograd <TW@SU-AI.ARPA>
Subject: Project Description - Cognitive Complexity (IBMSJ)
[Forwarded from the Stanford bboard by Laws@SRI-AI.]
The Computer Science Department at IBM Research, San Jose,
has a project which is developing the "cognitive complexity"
theory of D. Kieras (formerly at U. of Arizona, now at U. of Michigan)
and P. Polson (U. of Colorado). The theory is being applied
to representation of the user "how to do it" knowledge implied
by the particular design of, for example, an interactive text
formatter. The technology K & P have developed is directed to
representing user "how to do it" knowledge in production rules,
representing the surface design of the application (with respect to
what is seen by the user) as a generalized transition network, and
then deriving some measures of complexity (related to the Card, Moran,
and Newell work) for the design with respect to typical tasks.
The idea is to be able to compare design proposals (with respect to
the ease of learning and ease of use measures indicated by the
theory) at early stages of the design process. [...]
Anyone interested should contact John Bennett, bennett%ibm-sj@csnet-relay
------------------------------
End of AIList Digest
********************