[Prev][Next][Index]

12-1-83 Newsletter 11



                            CSLI Newsletter

December 1, 1983                 * * *                       Number 11

                         VISIT BY ROD BURSTALL

     Rod Burstall, the remaining member of our Advisory Panel, is here
to visit this week.  I hope that especially those interested in CSLI's
research area C  (theories of situated  computer languages) will  take
this opportunity to speak with him.
                                                         - Jon Barwise

        *        *        *        *        *        *        * 

                MEETING OF RESEARCHERS IN AREAS A AND B
     
     The meeting for the A and B Area people will be held from 1:00 to
2:00 p.m., Thursday, December 1, in the Ventura Conference Room  after
TINLunch.

                   MEETING OF RESEARCHERS IN AREA C

     The researchers associated with projects  in Area C  will have  a
general meeting  at 11:00  a.m. this  Friday, December  2, at  Ventura
Hall, to meet with  Rod Burstall.  It will be a chance for him to  get
to know us, and for us all to talk about general directions, projects,
interests, and problems that we see in this area.


                  JOINT MEETING FOR PROJECTS B3 AND B5

     At the joint  meeting of  B3 and  B5 on  Wednesday, November  30,
Geoff Nunberg discussed  his paper "Individuation  in Context."   Next
Wednesday, December  7, Phil  Cohen will  talk about  indirect  speech
acts.  His talk will be at 9 a.m. in Ventura Hall.


                       MEETING FOR PROJECTS C1-D1

     On November 29 and December 6, Yiannis Moschovakis will speak  to
the CSLI C1-D1 working group, held each Tuesday at 9:30 at PARC.   His
topic is: "On  the Foundations  of the Theory  of Algorithms."   These
talks will  present  in  outline an  abstract  (axiomatic)  theory  of
recursion, which aims to capture the basic properties of recursion and
recursive functions on the  integers, much like  the theory of  metric
spaces  captures  the  basic  properties  of  limits  and   continuous
functions on the reals. The basic notion of the theory is a (suitable,
mathematical representation of) algorithms.  In addition to  classical
recursion, the models of the theory include recursion in higher types,
positive elementary  induction, and  similar theories  constructed  by
logicians, but they also include pure Lisp, recursion schemes, and the
familiar   programming    languages   (as    algorithm    describers).
Technically, one  can view  this work  as the  theory of  many-sorted,
concurrent, and (more significantly) second-order recursion schemes.

        The first lecture concentrates on the pure theory of recursion
and describes some of the basic results and directions of this theory.
The second lecture looks at some of the less developed connections  of
this theory with the foundations of computer science, particularly the
relation between an algorithm and its implementations.

        *        *        *        *        *        *        *

                                Page 2

        *        *        *        *        *        *        *

          CSLI SCHEDULE FOR *THIS* THURSDAY, DECEMBER 1, 1983

10:00  Research Seminar on Natural Language

     Speaker:  Paul Kiparsky (MIT)
     Topic:	 On lexical phonology and morphology.
     Place:	 Redwood Hall, room G-19

12:00  TINLunch

     Discussion leader:  Paul Martin (SRI) 
     Paper for discussion:  "Planning English Referring Expressions"
                             by Douglas Appelt
     Place:  Ventura Hall

2:00  Research Seminar on Computer Languages

     Speaker:  Luca Cardelli (Bell Labs)
     Title:	 "Type Systems in Programming Languages"
     Place:	 Redwood Hall, room G-19

3:30  Tea

     Place:    Ventura Hall

4:15  Colloquium

     Speaker:  Charles Bigelow (CS, Stanford)
     Title:	 "Selected Problems in Visible Language"
     Place:	 Redwood Hall, room G-19


Note to visitors: 
 
     Redwood Hall is close to Ventura Hall on the Stanford Campus.  It
can be reached from Campus Drive or Panama Street.  From Campus  Drive
follow the sign for Jordan  Quad.  $0.75 all-day parking is  available
in a lot located just off  Campus Drive, across from the  construction
site.

        *        *        *        *        *        *        *

                               Page 3

        *        *        *        *        *        *        *

                           TINLUNCH SCHEDULE

     TINLunch is held at 12 noon each Thursday at Ventura Hall on  the
Stanford University campus as  a part of  CSLI activities.  Copies  of
TINLunch papers will be at SRI in EJ251 and at Stanford University  in
Ventura Hall.

     On Thursday, December  1, Paul Martin  will lead the  discussion.
The paper for discussion will be:

		"Planning English Referring Expressions"
			    by Douglas Appelt

This  paper  describes  a  theory  of  language  generation  based  on
planning.  The theory is illustrated through a detailed examination of
the problem of planning referring expressions.  This theory provides a
framework in which one can account for noun phrases used to refer,  to
supply additional  information, and  to clarify  communicative  intent
through coordination with  the speaker's  nonlinguistic actions.   The
theory is embodied in a computer system called KAMP, which plans  both
linguistic  and  nonlinguistic   actions  when   given  a   high-level
description of the speaker's goals.

NEXT WEEK (DEC. 8):

Robert Moore will  be leading the  TINLunch discussion on  a paper  by
Daniel Dennett entitled "Cognitive Wheels: The Frame Problem of AI."

        *        *        *        *        *        *        *

                       WHY CONTEXT WON'T GO AWAY

     On Tuesday, November 29, Peter  Gardenfors, who is visiting  CSLI
this year from  Lund University in  Sweden, gave a  talk entitled  "An
Epistemic Semantics for Conditionals."  The abstract is given below.

     A semantics  for  different  kinds of  conditional  sentences  is
outlined.  The ontological basis  is states of  belief and changes  of
belief rather than possible worlds and similarities between worlds. It
is shown how the semantic analysis can account for some of the context
dependence of the interpretation of conditionals.

Next week's speaker: Ivan Sag
                     December 6, 1983, 3:15 p.m. 
                     Ventura Hall

        *        *        *        *        *        *        *

                               Page 4

        *        *        *        *        *        *        *

        COMPUTER SCIENCE COLLOQUIUM NOTICE WEEK OF NOV 28-DEC 2

11/28/1983	      Robotics Seminar
Monday		      Clyde Coombs
4:15 p.m.	      Hewlett Packard
MJ252		      Manufacturing Strategy for Information and Automation

11/29/1983	      Knowledge	Representation Group Seminar
Tuesday		      Bob Blum
2:30-3:30	      Stanford CSD
TC-135 (Med School)   Representing Clinical Causal Relations in	the RX
		      Knowledge	Base

11/29/1983	      CS Colloquium
Tuesday		      John Seely Brown
4:15		      Cognitive	Sciences, Xerox	PARC
Terman Aud.	      A	Computational Framework	for a Qualitative
		      Physics--Giving computers	`common-sense' knowledge
		      about physical mechanisms

11/30/1983	      Special Tutorial
Wednesday	      Dr. Adrian Walker
2:00 - 4:00	      IBM Research Lab,	San Jose
MJH 252		      Introduction to PROLOG and Its Applications

11/30/1983	      Talkware Seminar
Wednesday	      Amy Lansky
2:15-4:00	      Stanford U./SRI
380Y (Math Corner)    GEM: A Methodology for Specifying	Concurrent Systems

12/02/1983	      Database Research	Seminar
Friday		      David Dewitt
3:15 p.m.	      University of Wisconsin
MJH 352		      Benchmarking Database Management Systems and Machines

        *        *        *        *        *        *        *

                               Page 5

        *        *        *        *        *        *        *

                       TALKWARE SEMINAR - CS 377

Date: November 30
Speaker: Amy Lansky (Stanford / SRI)
Topic: Specification of Concurrent Systems
Time: 2:15 - 4
Place: 380Y (Math corner)
 
     This talk will describe the  use of GEM, an event-oriented  model
for specifying and  verifying properties of  concurrent systems.   The
GEM model  may be  broken  up into  two components:   computations and
specifications.  A  GEM  computation  is a  formal  representation  of
concurrent execution.   Program executions,  as  well as  activity  in
other domains, may be modeled.  A GEM specification is a set of  logic
formulas that may be applied to GEM computations.  These formulas  are
used  to  restrict  computations  in   such  a  way  that  they   form
characterizations of  specific  problems or  represent  executions  of
specific languages.
 
     A primary result of my research with GEM has been a demonstration
of the power and breadth  of an event-oriented approach to  concurrent
activity.  The model  has been used  successfully to describe  various
language primitives (the Monitor,  CSP, ADA tasks), several  problems,
including two  distributed algorithms,  and for  verifying  concurrent
programs.

     In this seminar, I will introduce some of the important  features
of  GEM  and   demonstrate  their  use   in  modeling  many   familiar
computational    behavior    patterns,    including     sequentiality,
nondeterminism,   priority,    liveness,    fairness,    and    scope.
Specification  of  language   concepts  such   as  data   abstraction,
primitives such  as  CSP's  synchronous  I/O,  and  familiar  problems
(Readers/Writers) will  be  included.   This talk  will  also  discuss
directions for further research based on GEM.  One possibility is  the
use  of  graphics   for  the  construction   and  simulation  of   GEM
specifications.

Date:     December 7
Speaker:  Donald Knuth (Stanford CS)
Topic:    On the Design of Programming Languages
Time:     2:15 - 4
Place:    380Y (Math Corner)

Date:     December 14
Speaker:  Everyone
Topic:    Summary and discussion
Time:     2:15 - 4
Place:    380Y (Math Corner)

        *        *        *        *        *        *        *

                               Page 6

        *        *        *        *        *        *        *

             SEMINAR IN LOGIC AND FOUNDATIONS OF MATHEMATICS  

SPEAKER:  Professor J. E. Fenstad, University of Oslo

TITLE:    Peanos's existence theorem for ordinary differential equations
          in reverse mathematics and non standard analysis.

TIME:     Wednesday, Nov. 30, 4:15-5:30 PM

PLACE:    Stanford Mathematics Dept. Faculty Lounge (383-N) 

Abstract:  We continue  the  exposition of  Steve  Simpson's  work  on
reverse mathematics, locating the  exact position for the  provability
of Peanos's theorem.  It  follows that the  nonstandard proof is  more
constructive than the standard textbook proof.

        *        *        *        *        *        *        *

           1984 INTERNATIONAL SYMPOSIUM ON LOGIC PROGRAMMING
                        Atlantic City, New Jersey
                            February 6-9, 1984
 
                     Sponsored by the IEEE Computer Society
 
Registration details from:

     Registration - 1984 ISLP
     Doug DeGroot, Program Chairman
     IBM Thomas J. Watson Research Center
     P.O. Box 218
     Yorktown Heights, NY 10598

or from (ARPANET):  PEREIRA@SRI-AI

     The opening address  will be  given by Professor  J.  A.   (Alan)
Robinson of Syracuse University.  The guest speaker will be  Professor
Alain Colmerauer  of the  University of  Aix-Marseille II,  Marseille,
France.  The keynote speaker  will be Dr. Ralph  E.  Gomory, IBM  Vice
President &  Director  of Research,  IBM  Thomas J.   Watson  Research
Center.  On February 6, Ken Bowen of Syracuse University will  present
"Tutorial: An Introduction to Prolog."  Finally, during the  remaining
three days, February 7-9, 35 papers will be presented in 11  sessions.
 
        *        *        *        *        *        *        *

-------