ANNUAL REPORTS
presented to the Executive
of the Association
for Computational Linguistics
on
Tuesday, June 28, 1994
Las Cruces, NM, USA
ACL Executive Committee Officers, 1994-1995
President: Karen Sparck Jones
Vice President: Doug Appelt
Secretary-Treasurer: Judith Klavans
ACL European Chapter chair: Ewan Klein
CL journal editor: Julia Hirschberg
member, 94-92: Bente Maegaard
member, 93-95: Stuart Shieber
member, 94-96: Eduard Hovy
Report collated by Eduard Hovy, hovy@isi.edu
=============================================================================
TABLE OF CONTENTS
1. ACL Management
President's report Karen Sparck Jones, ACL President
Secretary-Treasurer's report Judith Klavans, ACL Secretary-Treasurer
Summary of ACL accounts for 1993 Fernando Pereira and Karen Sparck Jones
ACL European Chapter Susan Armstrong
ACL Executive Nominating Committee Ralph Grishman
The Walker Fund -- status Stuart Shieber
The Walker Fund -- financial status Judith Klavans
2. ACL-94
ACL-94 Program Committee James Pustejovsky
ACL-94 registration Judith Klavans and Betty Walker
ACL-94 local arrangements Janyce Wiebe
ACL-94 tutorials Lynette Hirschman
ACL-94 student session Beryl Hoffman
3. Computational Linguistics Journal
CL Journal, Editor's report Julia Hirschberg, CL Editor
CL Journal, Finite String Ralph Weischedel
CL Journal, Book Review Graeme Hirst
CL Journal, Squibs and Comments James Pustejovsky and Bob Ingria
4. Future ACL and Associated Conferences
COLING-94 Program Chair's report Yorick Wilks
Applied NLP-94 Program Committee Paul Jacobs
Applied NLP-94 local arrangements Christian Rohrer and Uwe Reyle
ACL-95 local arrangements Robert Berwick
ACL-96 site selection Doug Appelt, ACL Vice President
5. Organizations of Interest
Text Encoding Initiative Judith Klavans
European Corpus Initiative Susan Armstrong
Consortium for Lexical Research Louise Guthrie and Katherine Mitchell
Linguistic Data Consortium Jack Godfrey
6. Information Dissemination, Directories, and Surveys
Electronic distribution of CL papers Stuart Shieber
Survey of CL courses Bonnie Dorr
Graduate Directory Martha Evens
NLP Software Registry Elizabeth Hinkelman
7. Special Interest Groups
SIG general matters Doug Appelt, ACL Vice President
SIGGEN David McDonald
SIGLEX James Pustejovsky
SIGMEDIA Kent Wittenburg
SIGMOL Robert Berwick
SIGPARSE Harry Bunt and Masaru Tomita
SIGPHON Steven Bird
=============================================================================
PRESIDENT'S REPORT
Karen Sparck Jones, President
kn11@phx.cam.ac.uk
STAFF AND ORGANISATIONAL MATTERS
This has necessarily been a difficult transitional year for ACL,
since we have not been able to make the intended change from
old Secretary/Office organisation to a new in the smooth way planned.
As Judith Klavans indicated she did not want to become Secretary-
Treasurer, a Search Committee consisting of the Nominating Committee
plus myself have been looking for a successor to Don Walker. At
the same time, negotiations are underway to engage an Office-Manager
to replace Betty Walker, who wishes to take her well-earned retirement
from ACL's administration. These two new people will be working
within the framework of a revised allocation of responsibilities,
with the Secretary-Tresurer in a senior position with much of the
work Don did delegated to the Office-Manager: our negotiations with
people are based on formal job descriptions for both positions.
Betty has very kindly supported the office since the autumn, and indeed
ACL could not have survived without her knowledge and dedication; but
after ACL 94 the transfer to the new office will begin, which we hope
will be up and running by January 1995.
FINANCES
These organisational changes will necessarily imply heavier central
costs, and the transition period has also had its own costs. Fernando
Pereira and I have been engaged with the necessary financial analysis
of our present operations. We are preparing the ground for the revised
arrangements that will be required with the office reorganisation.
The Executive recognises that the ACL, through Don Walker's personal
situation, was able to operate in a very economical fashion; and they
accept that in the future the cost of running ACL as an organisation with
some 2000 individual members round the world and many and varied activities,
notably publications and conferences, associated with its leading position
in the field, will lead to increased dues, fees and other charges.
The Executive has already approved higher dues and fees for this year.
It must, however, be emphasised that the ACL's charges, given what it
offers, are very modest compared with many other similar organisations.
FORMAL STRUCTURE
It is clear that the increasing range of ACL's activities, and widespread
membership, should be better reflected at the formal level by, for instance,
a larger Executive, more definition of the relationship between the central
ACL and its Executive and the European Chapter. I intend to start work
on these matters so as to bring proposals to next year's Annual Business
Meeting. At the same time, the ACL is now well-established, with a recognised
standing and image, and it is therefore possible to envisage more devolution,
for example in conference management, as has already been the case with the
SIGs that are a growing part of ACL's activities.
PUBLICATIONS ETC
During the year an important administrative change was made, by moving
the sales of back Proceedings to Morgan Kaufmann. A major Graduate Course
Survey appeared during the year, and the Computational Linguistics Journal
flourishes. The ACL has continued to endorse the various electronic databases
and initiatives with which it has been concerned in the past, and there
are new efforts in Europe. The Executive has also been actively engaged in
considering ACL's relationship with the new trends and opportunities that
are occurring in electronic publication, both for their own value to the field
and benefit for ACL, as well as for their potential implications for ACL's
existing publications.
MAJOR EVENTS
This year ACL has two Conferences, with the European Conference early
next year as well. The Executive has also taken initial steps towards
occasional joint meetings with COLING, if suitable opportunities arise.
ACL's conferences are increasingly important as a venue for the whole field,
to offset the limited participation that applies to the ARPA Workshops.
THANKS
Having perforce during the year to undertake many Secretary-Treasurer
tasks, I want to acknowledge the value of the many illustrations I have
had of Don Walker's prudence in financial matters, friendliness in
personal ones, and openness in intellectual ones. I want to thank Judith
Klavans for handling many administrative chores, the members of the
Executive for undertaking many adhoc tasks required of them by our
organisational emergency, and also Sandra Carberry for stepping in to
help with ACL94 local arrangements. The Executive is particularly
grateful to Betty Walker for keeping the Morristown office show on the
road, as without her ACL would certainly be foundering in every way;
and I personally owe a large debt to Fernando Pereira for continuing to
help when he should have been able to retire from ACL duties.
=============================================================================
SECRETARY-TREASURER'S REPORT
Judith Klavans
klavans@cs.columbia.edu
Many changes have affected the S/T position of the ACL this year.
First and foremost, Don's health continued to falter through the
summer. I took on a significant amount of responsibility during this
period, and started setting up some procedures for the eventual
transfer of the position. This was an extremely trying transitional
period, since we wanted to maintain continuity while at the same time
respecting Don's wish to remain at the helm. The latter wish was
honored until Don's death at the end of November.
At that point, there were a host of items needing attention, some of
which had slipped and some of which arose at the time. At the same
time, due to the change in my professional situation, it became
increasingly evident that I would be unable to devote the enormous
amount of time that Don was able to give to the organization, given
the support of Bellcore and of course of Betty.
Given this situation, in consultation with Karen Sparck Jones, current
president, and Fernando Pereira, past president, several of the
functions I was serving were delegated to others to help with the
overwhelming amount of responsibility entailed in the position. Among
those involved are: Sandra Carberry, who has taken primary
responsibility for 1994 arrangements with Jan Wiebe and of helping
with financial matters for the 1994 annual meeting; Kathleen McKeown,
who has taken responsibility for negotiating with publishers for a new
ACL book series; Fernando Pereira, who has taken the responsibility
for the 1994 program announcement and for putting the financial
matters into shape for 1993 taxes, Stu Shieber, who is helping to
establish guidelines for distribution of the Walker Fund, Martha
Palmer, who took responsibility for the Don Walker Festschrift, Ed
Hovy, who has gathered the 1994 Annual Reports, and Karen Sparck Jones
who has taken the responsibility to oversee all matters of dues, and
of establishing guidelines in many other areas. I continued to be
involved and responsible for many other aspects of the position.
My primary activity from June to December was to help with the
transition as much as possible. From December until now, I also have been
involved in other activities. These include contributing to decisions on:
1994 meeting
-- tutorial chair, tutorials, pc committee
-- reviewing financial matters for the annual meeting, and other
activities
1994 Applied meeting
-- choice of chair and consulting on ongoing decisions
1995 meeting
-- scheduling, choice of site, financial matters
1995 EACL
-- scheduling, choice of site, financial matters
Walker Fund
-- establishing fund
-- responding to each donator
-- keeping bookwork up to date
-- informing exec committee of status of the Fund
Electronic Information
-- maintaining the listserv for ACL at Columbia
-- setting up the ACL Home page at Columbia through www
Text Encoding Initiative
-- attended February meeting in Poughkeepsie
-- work with TEI Steering Committee as ACL representative
General ACL Business
-- negotiating new contract with Mike Morgan for distribution
of proceedings
-- fielding requests for support for meetings and individuals
-- Survey of Programs (Bonnie Dorr)
-- Helping with new SIGs (e.g. sigphon, sigdata)
-- Finite String
-- mailings to the membership
-- and many other matters too detailed to list and which are part
of ongoing ACL business
I have written up a short manual which will serve as guidelines to help
the next incoming Secretary/Treasurer in fulfilling the many responsibilities
associated with the ACL.
=============================================================================
SUMMARY OF ACL ACCOUNTS for 1993
Fernando Pereira (former President) and Karen Sparck Jones (President)
kn11@phx.cam.ac.uk
US DOLLARS
ASSETS
End 92 End 93 Change
Total 323,204.75 310,418.40 ( 12,786.35)
INCOME
Line totals Category totals Totals
Membership :
Personal Dues 33,542.00
Student 4,074.29
Japanese 4,500.00
Others 1,770.00
Unclass CC 4,096.00
Total Dues 47,982.29
Journal Mailing 2,470.00
Publication Sales :
CL Back Issues 3,320.00
Others 455.00
Total Back Issues 3,775.00
Proceedings 34,456.00
Other Mail 1,729.00
Coling Sales :
Pre 90 4,455.00
COLING 90 6,312.00
COLING 92 12,990.00
Total COLING 23,757.00
Funds :
Int Fund 1,505.00
Walker Fund 1,657.50
DCI CDROM 2,076.50
Labels 1,751.00
Meetings :
92 Meeting 6,948.83
Workshops 1,165.00
93 Meeting 66,025.00
Total Meetings 74,138.83
Royalties 11,956.21
Miscellaneous 560.00
Investment Income 13,702.31
Foreign Exchange ( 1,720.65)
TOTAL INCOME 219,515.99
EXPENDITURE
Services 36,204.18
Mailing 18,119.63
Journal 81,559.82
Meeting 47,341.11
Proceedings 14,030.00
Travel 9,481.61
Supplies 937.11
Equipment 2,007.43
Telephone 2,251.77
Refunds 2,303.34
Ling Inst 12,000.00
Workshops 4,392.07
Grad Dir 4,909.43
SIGs 333.92
TOTAL EXPENDITURE 235,871.42
=============================================================================
ACL EUROPEAN CHAPTER
Susan Armstrong
susan@divsun.unige.ch
The past year has witnessed a period of transition for EACL, with the
goal of expanding the range of activities undertaken by the Chapter,
augmenting the administrative structure to support such activities,
and developing links with other organisations in Europe.
ADMINISTRATION
There have been two noteworthy changes on the administrative front.
The European Chapter's proposal that the Chair of EACL should be
ex officio a member of the Association's Executive Board has
been implemented on an informal basis. This has significantly
helped Ewan Klein gain a better picture of ACL activities and
working procedures.
In addition, the Chapter has found the need for a more formal
decision-making body which would parallel the Association's Executive
Board, and has proposed that this be composed of the Advisory
Committee (currently, Boguslavsky, Sagvall Hein and Somers) plus the
Executive Officers (currently Armstrong, Klein and Rosner). This
arrangement has been adopted as a practical step for the interim,
prior to any constitutional change.
The importance of these steps has been underlined by Don Walker's
untimely death, and the consequent redistribution of tasks among ACL
Executive Board members. There were many European concerns that Don was
sensitive to and actively involved in, and for which new procedures
must be established. We are working on addressing these issues, many
of which are only slowly becoming apparent. Amongst the tasks so far
identified are:
- ensuring that the membership list is properly updated and that
clearer communication protocols are followed between the offices of
the Association and the Chapter;
- overseeing the organization of the EACL conference and in particular,
approving the budget arrangements;
- ensuring that Chapter officials are adequately informed about
Association discussions and decisions;
- handling requests for subsidies to support European initiatives.
Inevitably, much of the extra workload has fallen on the shoulders of
the Chapter secretary.
One important outstanding question yet to be clarified is which
decisions can be taken at the discretion of the Chapter, as opposed to
those which require final ratification by the Association's Executive
Committee.
Most of the EACL business is conducted over email and by telephone.
Meetings have been held on an ad hoc basis whenever they could
be combined with other events (i.e. at minimal cost to the Association).
In future, however, it would be preferable if Association funding
were available for some face-to-face communication in the years where
no Chapter meeting is held.
ELECTIONS
At the end of 1994, a number of European Chapter Officials will be
up for replacement and/or reelection. The Chair of the Nominating
Committee, Peter Sgall, has initiated action to nominate a slate of
official candidates. Notices of these nominations will be sent to the
membership for approval or additional nominations.
MEMBERSHIP
Membership Statistics
It is difficult to give precise indications of European memberships for a
number of reasons discussed below. Our current estimates are as follows:
MEMBERS 1989 1990 1991 1992 1993
Student Members 54 53 66 47 88
Total Members (inc. students) 612 577 703 685 773
Continuing Members -- 461 505 521 534
Note: a "continuing member" for a given year is defined as someone who
was also a member the previous year.
Membership Lists
The EACL offices (secretary and treasurer) hold membership lists
that are updated regularly from the United States office. However, we
find that the datafiles currently in our possession are incomplete and
have been compiled according to different criteria. Changes and
deletions carried out at irregular intervals have caused distortions
in the statistics we have tried to compile over the past few years. We
suggest that, in order to streamline this process, the updating tasks
and list distribution previously handled by Don receive special
attention as soon as the new ACL secretary has been appointed
We offer some practical suggestions for improving the quality of the
data.
- Decide what information is necessary;
- adopt a standard format for data entries (currently there are at
least three); and
- decide on a consistent policy and procedure for deleting
obsolete entries, to ensure completeness of the entries for a given
year.
CONFERENCES
Language Engineering Convention
Susan Armstrong represented EACL interests as a member of the
Scientific Advisory Committee for the Language Engineering Convention,
co-sponsored by DG XIII's Language Research Engineering (LRE)
programme and the French ministries of Research (Ministere de
l'Enseignement Superieur et de la Recherche) and Industry (Ministere
de l'Industrie, des Postes et Telecommunications et du Commerce
Exterieur). ELSNET (largely in the person of Ewan Klein as
coordinator) is under contract with the EC to provide
scientific input to the organisation of the Convention.
4th Conference on Applied Natural Language Processing
We have been following with interest the applied ACL conference to
be held in Stuttgart this year. The program chair has shown a good
awareness of the importance of actively pursuing industrial
participation. The coordination of academic and industrial interests
has recently become a crucial issue for all European funded Language
Engineering programs. In order to reinforce the role of the
asscociation in Europe, it would be desirable to install some
institutionalized communication flow between EACL and organizers of
any ACL activity held in Europe.
Guidelines for Conference Organization
There are currently a number of documents circulating that address
conference organization. The Rappaport guidelines present a very good
starting point but the detailed comments are very specific to
organizational issues for ACL meetings held in the United States. As
the conferences begin to move around the world we suggest it would be
useful to have a more global document to help organizers and ensure
that ACL policy is followed. In preparation for such a document, we
have asked past EACL organizers to provide input on their
experiences which can be incorporated within the framework of
Rappaport's document.
7th EACL 95, Dublin
Previous EACL conferences were closely monitored by Don Walker. His
work covered essentially all aspects of the conference from local
organization, address lists and mailings to program committee matters.
The EACL Executive has been in close communication with Allan Ramsay,
organizer of the conference, in an attempt to assume those tasks
previously covered by Don.
As a departure from previous practice for European Chapter meetings,
we have appointed a program chair. This task will be undertaken by
Erhard Hinrichs in collaboration with Steven Abney (both of Tuebingen
University). We have delegated other tasks such as Tutorial and
Student Chair to other European centres. In lieu of any formal
procedures, we will submit a full slate of proposed program committee
members and the current budget proposal to the ACL Executive for
approval. In light of our experiences with this conference we will
work on developing clearer procedures for tasks that in the past were
personally taken care of by Don and Betty Walker as organizers,
advisors and formal liaison to ACL. Immediate issues we have
identified include:
- choice of EACL conference site;
- choice of program chair;
- acceptance of program committee;
- acceptance of budget proposal;
- releasing ACL funds for working expenses; and
- question on extension of student support to EACL conferences.
Meetings in 1997 and 1998
According to the current, tentative, schedule, the main ACL meeting
in 1997 will take place somewhere in Europe, and it would therefore
make sense to combine this with the 8th European Chapter meeting.
1994 New Mex -- Kyoto Europe
1995 Cam,Mass Dublin
1996 NAm West -- Copenhgn
1997 Europe NAm (Philadelphia?)
1998 Montreal Europe Montreal
ACCOUNTS
A separate finacial statement has been prepared by Mike Rosner
(cf. European Treasurer's Report -- 1993).
PUBLICITY
A WWW document for EACL was written and linked to the ELSNET WWW
page at the University of Edinburgh in December 1993. Some steps
should be taken to set up links between this and the new ACL WWW
server, and to ensure that the EACL page is properly maintained in
future.
=============================================================================
ACL EXECUTIVE NOMINATING COMMITTEE
Ralph Grishman
grishman@grimm.cs.nyu.edu
The nominating committee places the following names in nomination:
Doug Appelt for President
Oliviero Stock for Vice President
Kathleen McCoy for Executive Committee
Karen Sparck Jones for the Nominating Committee
=============================================================================
THE WALKER FUND -- STATUS
Stuart Shieber
shieber@das.harvard.edu
The Executive Committee of the Association for Computational
Linguistics unanimously voted in 1993 to establish a fund to honor Don
and Betty Walker and their years of dedication and work for the ACL.
They have helped establish ACL as a significant national and
international organization, and have made extensive contributions
inside and outside the field. The fund is intended to be used to help
defer travel costs for worthy students to attend meetings, in
recognition of the priority that the Walkers have always placed on
making the ACL accessible and welcoming to students and others who for
geographical or financial reasons would otherwise not be able to
participate as fully in ACL activities.
Currently, the Fund has some $14,000. This is a substantial amount,
one which is an honor to the Walkers and which gives us a way to
express our appreciation via student scholarships.
A proposal has been developed by Stuart Shieber and Becky Passoneau in
consultation with various colleagues (especially Judith Klavans) and
several students as to an appropriate mechanism for administering the
Walker fund. The proposal addresses the issues of
o Eligibility
o Application method
o Selection method
o Selection criteria
o Dispersal method
o Amount of grant
o Transition
Primary among our goals were flexibility, simplicity, and
adherence to the fund's aim. We thought it especially important that
the mechanism used in the first round of granting funds be thought of
as experimental, and that the executive committee should review these
methods and allow them to evolve as we gain experience with the Walker
Fund granting process.
We expect that the first dispersal of funds will be for the upcoming
1995 meeting of the ACL.
=============================================================================
THE WALKER FUND -- FINANCIAL STATUS
Judith Klavans
klavans@cs.columbia.edu
At the 1993 annual meeting of the Association, the executive
committee unanimously voted to establish a Fund to honor the
many years of service to the ACL by both Don and Betty Walker.
I am very pleased to inform the membership that, as of
June 15, 1994, there is $15169 in the Fund. This amount far
exceeded the estimates given to us by other professional
organizations with similar Funds, and is a visible tribute to
the appreciate the membership feels towards the Walker's. I
anticipate another round of donations at the 1994 meeting. Over
150 people contributed to the Fund, with contributions ranging
from $5.00 to $5000.00.
The following accounts show the submissions to the Fund by
month, and the current location of the Funds. The amount in the
Klavans account will be transferred to the ACL Paine-Webber
account just after the 1994 annual meeting.
June 1993 $1812.00 (ACL-Walker)
July 1993 717.50 (ACL-Walker)
August to November none
December 1993 4660.00 (ACL-Klavans)
January 840.00 (ACL-Klavans)
February 575.00 (ACL-Klavans)
March 5395.00 (ACL-Klavans)
April 750.00 (ACL-Klavans)
May-June 420.00 (ACL-Klavans)
Total to date $15169.50
($ 12640 deposited in Klavans account, $2529 in Walker account)
=============================================================================
ACL-94 PROGRAM COMMITTEE
James Pustejovsky
jamesp@cs.brandeis.edu
=============================================================================
ACL-94 REGISTRATION
Judith Klavans and Betty Walker
klavans@cs.columbia.edu
=============================================================================
ACL-94 LOCAL ARRANGEMENTS
Janyce Wiebe
wiebe@cs.nmsu.edu
Exhibitors
We have exhibition commitments from the following organizations: CLR,
CSLI, Cambridge University Press, Kluwer Academic Publishers, Lawrence
Erlbaum Associates, MIT Press, Morgan Kaufmann Publishers, Oxford
University Press, University Of Chicago Press, and Walter De Gruyter,
Inc.
Sponsorship
We have raised a total of $3,500 to help offset the cost of the banquet and
receptions from the following organizations: Apple, Las Cruces Convention
and Visitors Bureau, Microsoft, and Sun Microsystems. In addition, Sun is
generously donating equipment for our World Wide Web Visitor's kiosk.
Software Demonstrations
We have been contacted by a number of people and are still working out
the final schedule.
=============================================================================
ACL-94 TUTORIALS
Lynette Hirschman
lynette@linus.mitre.org
There will be four tutorials presented at the 1994 ACL Meeting,
spanning several emerging areas of computational linguistics. Two
tutorials focus on sub-sentence level processing, namely computational
phonology (which will complement the workshop following the ACL
meeting on computational phonology), and tools derived from
text-to-speech synthesis work. These two were developed from
proposals submitted by the presenters. The second two tutorials
(which were solicited to add breadth to the program) focus on emerging
application areas, namely spoken language systems and computational
linguistics applied to biology.
Steven Bird (University of Edinburgh) will present a tutorial on
computational phonology, which will begin with an introduction to the
principles of phonology, including distinctive feature theory,
underspecification and markedness, subsegmentals and suprasegmentals,
and non-linear phonological representation. The tutorial will then
present the principles of computational phonology, including neural
nets, finite state devices and declarative phonology approaches.
Richard Sproat and Michael Riley (AT&T) will present a tutorial on
methods derived from text-to-speech synthesis work. These tools, which
use corpus-based methods, address problems such as text normalization
(including word boundary detection for e.g., Chinese), grapheme to
phoneme conversion, and intonational phrase-boundary and prominence
prediction.
Stephanie Seneff (MIT) will present a tutorial on spoken language
systems, combining speech recognition technology and language
understanding. Spoken language interfaces can provide intuitive,
easy-to-use interfaces for the rapidly increasing volume of information
available on the National Information Superhighway. The tutorial will
discuss techniques for robust understanding noisy input (due
to recognition errors and speaker disfluencies), and for carrying on
two-way conversations to aid the user in accomplishing a task.
David Searls (University of Pennsylvania) will present the fourth
tutorial on computational linguistics applied to genetics, covering
techniques from mathematical and computational linguistics used to
analyze biological structures (DNA, RNA, proteins) as strings over
fixed alphabets. The tutorial will discuss formal language-theoretic
aspects of DNA, linguistic aspects of evolution, and syntactic
ambiguity in the structure and function of genetic material. These
techniques will be applied to biological problems such as the use of
syntactic pattern recognition techniques for gene finding.
The tutorials will be held on Monday, June 27, with two tutorials
running in parallel in both the morning and the afternoon:
Track I Track II
************************** **************************
AM Bird (phonology) Seneff (spoken language)
PM Sproat/Riley (text-to-speech) Searls (biology)
=============================================================================
ACL-94 STUDENT SESSION
Beryl Hoffman
hoffman@linc.cis.upenn.edu
1. The Committee
The ACL-94 Student Sessions program committee includes 10 students,
and 11 nonstudents (4 from industry, and 7 faculty members). To
represent the diversity of the computational linguistics community,
we strove to balance the committee according to areas of expertise,
gender, and geographical location.
The student members of the committee are:
Jennifer Chu (U Delaware), Jason Frank (Ohio State U),
Steve Green (U Toronto), Vasileios Hatzivassiloglou (Columbia U),
Beryl Hoffman (U Penn), Peter Heeman (U Rochester),
Chris Manning (Stanford U), Gaelle Recource (U Paris 7),
Sheila Rock (U Edinburgh), Suzanne Stevenson (U Toronto/U Maryland).
The nonstudent members are:
Chinatsu Aone (SRA), Alan Black (ATR Japan), Ken Church (AT&T),
Robert Frank (U Delaware), Megumi Kameyama (SRI),
Robert Kasper (Ohio-State U), Chris Mellish (U Edinburgh),
Gord McCalla (U Saskatchewan), John Nerbonne (U Groningen),
Rebecca Passonneau (Columbia U), and Ingrid Zukerman (Monash U Australia).
The co-chairs of the ACL-94 Student Sessions, Beryl Hoffman (U
Pennsylvania) and Rebecca Passonneau (Columbia U), were nominated by
last year's student coordinator, Linda Suri, and approved by the ACL
executive committee. The ACL exec committee also chose Eric Iverson
(New Mexico State U) for student local arrangements. The students on
the committee were chosen by the co-chairs from among the volunteers
at the 1993 ACL conference and from personal recommendations by
various faculty members. We appointed Sheila Rock, who was a holdover
member from the 1993 committee, as the Committee Organizer, and she
helped us to choose the non-students on the committee which was then
approved by the ACL exec committee.
In past years, the student planning committee was formed first, and
then they together chose the non-student reviewers for the program
committee. This year, the non-student reviewers were chosen by the
co-chairs and the committee organizer, after consultation with
individual student members in their areas of expertise. This was
partly due to not having enough student volunteers to form the student
committee right away, and partly to try this new and perhaps more
efficient method sugested by last year's student chair, Linda Suri.
We feel that this method was successful in producing a well-balanced
committee under the stringent time constraints. Most of the student
members of the committee were also happy with the level of input they
had in planning the student sessions this year.
The committee worked well together and did a great job. We are also
grateful to Sandra Carberry, Judith Klavans, Fernando Pereira,
James Pustejovsky, Linda Suri, and Don Walker for their help.
2. The Submissions
This year, there were 41 submissions to the student sessions, and 10
were accepted for presentation at the conference. The average number
of submissions to the ACL Student Sessions for the past four years is
39.3. A survey sent to all of the authors who submitted papers this
year indicated that 17 of the 25 students who replied (i.e. 71%) had
not presented or submitted a paper to ACL before this year. The
students replied that they had heard about the ACL Student Session
through mail lists or newgroups (7 students, mostly through the
Linguist List), through their advisors (6 students), through ACL mail
sent to members (6 students), or through various other channels (5
students). Based on this information, I advise next year's committee
to widely circulate the call for papers, and to urge faculty members
to encourage their students in submitting papers to conferences.
2.1. The Format of the Submissions
When the ACL student sessions were created in 1991, many decisions had
to be made regarding the purpose of the student sessions and their
format. After three years of experimenting, the format of the student
sessions has become standardized. For the last two years, our
requirements for the submissions have been three-page papers
describing work in progress by student authors; the accepted papers
are presented at the conference and published in the proceedings. Last
year, the ACL main and student sessions adopted blind reviewing, and
this has continued this year as well. A survey sent to all authors who
submitted a paper this year indicated that the majority agreed with
the submission and conference formats, although many student authors
found the 3 page restriction (which guarantees space in the
proceedings for the student session papers) very limiting.
We did have some problems with the format for the submissions this
year. Our first call for papers had the incorrect name for the style
file required for LaTeX submissions. In addition, some confusion was
caused by the main session and student session using different style
files for the submissions, and by the requirement that a title page as
well as an identification page be provided. Based on this experience,
we will provide an improved model for submissions next year.
We allowed authors to submit hardcopy only or both e-mail and hardcopy
versions of their papers. Last year's committee allowed email-only
submissions, but also experienced e-mail problems at the university on
the day of the submission deadline. Following the main session's
directions, we required that all authors, even those submitting
hardcopy only, send their identification page through e-mail. I found
this very useful because it allowed me to start a dialogue with
authors whose submissions had formatting problems. I also used the
opportunity to urge all authors to send in an e-mail version if at all
possible, which covered us when some of the next-day airmail packages
arrived a week late, due to problems with university mail. Since
neither e-mail nor air mail is completely reliable, the both e-mail and
hardcopy option is the safest format; however, this is costly for the
student authors. Another option that next year's chair may want to try
is waiving the hardcopy requirement if authors e-mail a printable
form of their paper the day before the deadline.
After much debate, we allowed the e-mail format to be ASCII, LaTeX, or
PostScript files (the main session did not allow PS files). We
encouraged LaTeX submissions, however the authors overwhelmingly
preferred PS submissions. The following table shows the format for
e-mail submissions in 1992 and 1994. The increase in PostScript files
and decrease in hardcopy only submissions this year is probably
due to my personal requests for e-mail submissions.
---------------------------------------------------------------------
FORMAT FOR E-MAIL SUBMISSIONS
1992 1994
Hardcopy Only 13 4
LaTeX 17 7
ASCII 3 4 (1 was missing figures)
PostScript: 17 (1 unprintable) 26 (3 had unprintable figures)
----------------------------------------------------------------------
About 7 (5 PS and 2 LaTeX) of the 41 submissions this year had
formatting or printing problems which were either fixed by me, or by
the author, or unresolved in the case of the 3 PS files. I spent about
the same time to process the LaTeX and the PS submissions. Advantages
of the LaTeX files are that I could fix minor formatting errors myself
rather than ask the authors to do it; LaTeX files take up less space than
PS files and are easier to e-mail; and they allow reviewers to change
the papers to whatever format they prefer to make them easier to read.
However, over 50% of the 24 authors who replied to the survey
indicated that they would prefer submitting PS files, and only 8%
preferred LaTeX. Some authors can only provide PS or ASCII files
because they do not use LaTeX; some had to transfer files from PCs to
e-mail them to me. Authors also found the LaTeX requirement too
stringent because they wanted to be able to include their own style
files, macros, and PS figures. Perhaps making packages such as fig2dev
(which can translate fig code to LaTeX code) available may help to
make LaTeX submissions more attractive.
In the past 2 years, the chairs have reported that e-mail submissions
did not help to cut down air mailing costs to the reviewers because
almost all of the reviewers had at least one hardcopy only submission.
This year, I only had to fedex papers to 5 (about 1/4) of the
reviewers since we reduced the number of hardcopy only submissions.
2.2. The Distribution of Submissions
The following tables show the statistics on the distribution of
submissions and success rates according to geographical region,
gender, and subject areas.
Submissions showed an increase in Europe and Australia, but a decrease
in Asia and Canada when compared with last year. Perhaps more effort
is needed in circulating the call for papers outside of the U.S. to
improve the submission and acceptance rates in these regions.
-------------------------------------------------------------------
Region (by school) Submitted Accepted Success Rate
Australia 2 (5%) 1 (10%) 50%
Europe 14 (34%) 1 (10%) 7%
U.S.A. 25 (61%) 8 (80%) 32%
TOTAL 41 10 24%
------------------------------------------------------------------
Although it is unfortunate that the success rate was not equal among
female and male authors this year, the numbers are too small to be
significant. Last year, the success rate of female authors was higher
than the male authors; the average of the two years places the success
rate of women at 33% and men at 32%. Note that in both years the
reviewing was blind.
-------------------------------------------------------------------
Gender(of 1st author) Submitted Accepted Success Rate
Female 7 (17%) 1 (10%) 14%
Male 32 (78%) 9 (80%) 28%
Unsure 2 ( 5%) 0
TOTAL 41 10 24%
-------------------------------------------------------------------
Parsing and statistical approaches were the subject areas with the
highest submission and acceptance rates this year. Fractions in the
following table are a result of counting papers in up to 3 subject
areas.
----------------------------------------------------------------------
SUBJECT AREAS Submitted Accepted Success Rate
Connectionist Approaches 0.8 ( 2%) 0.3 ( 3%) 38%
Discourse(planning) 4.0 (10%) 0.5 ( 5%) 13%
Discourse(reference,temporal) 3.5 ( 9%) 1.0 (10%) 29%
Generation 1.5 ( 4%) 0.5 ( 5%) 33%
Information Retrieval 2.0 ( 5%) 0 0%
Knowledge Representation 0.5 ( 1%) 0 0%
Lexicon 2.5 ( 6%) 0.3 ( 3%) 12%
Machine Translation,Lang.learning 3.3 ( 8%) 0.8 ( 8%) 24%
Morphology 1.5 ( 4%) 0 0%
Parsing theory/practice 7.2 (18%) 2.3 (23%) 32%
Psycholing.and sentence proc. 0.5 ( 1%) 0.5 ( 5%) 100%
Semantic formalisms/theories 4.2 (10%) 0.8 ( 8%) 19%
Statistical Approaches 6.5 (16%) 2.5 (25%) 39%
Syntactic formalisms/theories 3.0 ( 7%) 0.3 ( 3%) 10%
TOTAL 41 10 24%
--------------------------------------------------------------------
3. The Reviewing Process
The co-chairs tried to assign each paper to the most appropriate
reviewers in the committee for blind reviewing. Each submission was
reviewed by a student and a nonstudent reviewer. The reviewers were
asked to rate the papers on a five-point scale and provide a
recommendation to the co-chairs and constructive comments to the
authors. Student authors were asked in a survey to rate the reviewers
and the reviews as well. 75% (18/24) of those who responded to the
survey rated the reviewers as qualified or very qualified in their
fields; the same percentage felt the reviews were at the right level,
and that the reviews were very or somewhat helpful to their research.
Once the reviews came back, the co-chairs evaluated and compared the
reviews and ratings. The final decisions were made by taking the
average of the ratings; if one reviewer was easier or harsher than the
average reviewer, we took this into account before making the final
decisions. If two reviewers strongly disagreed about a paper, I either
mediated a discussion between the two reviewers (for four papers) or
requested a third review (for four papers). Ten papers on the
borderline of acceptance were additionally read and reviewed by one or
both of the co-chairs.
A third of the student authors felt that more reviews would have been
helpful. I also feel that having 3 reviews (perhaps from 1 nonstudent
and 2 student reviewers) would remove the stress of getting
tie-breaking reviews at the last minute, as well as providing more
feedback for the authors. In addition, I feel the reviewing process can be
improved to elicit more standard and more constructive feedback from
the reviewers. Although the student reviewers liked the free format on
the reviewing forms and did not want to give it up, most agreed that a
list of questions to keep in mind, for instance the following, would
be useful.
a. Please give a one or two line summary of the paper.
b. Is the paper at the right stage to be presented (i.e. does it
present a problem, a possible solution, preliminary results, and
discuss future work)?
c. What are the strengths of this paper?
d. Are there any references which the author needs to be aware of?
e. Are there problems in the presentation of the paper (e.g. typos,
lack of clarity, organization of the material)?
f. What does this paper need to strengthen its argument?
4. The Conference
The average submission and acceptance rates for the past 4 years has
varied according to the time and format allotted to the student
sessions as seen below.
---------------------------------------------------------------------
Year Submissions Acceptances Success Rate Format, Number of Hours
---- ---------- ---------- ---------- ------------------------
1994 41 10 24.4% parallel sessions, 2 hrs
1993 30 11 36.7% single session, 4 hrs
1992 48 20 41.6% parallel sessions, 4 hrs
1991 38 15 39.5% parallel sess. during lunch
Average 39.3 14 35.6%
----------------------------------------------------------------------
The number of papers we accepted was strongly influenced by the time
constraints at the conference this year. The main session allotted 2
hours of parallel sessions at the conference for the student sessions.
We scheduled 20 minutes for each of the 10 papers (15 minutes for each
talk followed by 5 minutes of questions). We also scheduled two 10
minute breaks to allow for informal discussions with the authors
without losing the audience for the next papers.
In the surveys of the students on the committee and students who
submitted papers, 40% of the students preferred single sessions during
conference, 40% preferred mini student sessions mixed in with the ACL
main sessions, and 20% preferred parallel sessions; lunch time
sessions were rejected by all. Many students stated that they would
not mind parallel sessions if it allowed more papers to be accepted,
and some (30%) felt that the acceptance rate of papers this year was too
low. I would advise next year's student chair to negotiate with the
main session chair in advance to reserve at least 2.5 hours in
parallel sessions, which would allow 12 papers to be accepted (with 20
minutes per paper and some time for breaks). Some students (20%) also
expressed an interest in poster sessions, in addition to the paper
presentations, to maximize student participation and feedback at the
conference; next year's committee may want to investigate this option
as well.
Pending approval from the ACL exec committee, we would like to place
information about the ACL Student Sessions (including an archive of
the reports, surveys, and perhaps an ACL student member directory) on
the WWW. As a result of the surveys, 13 students volunteered to serve
on next year's student sessions committee; in addition, students
attending the conference will be asked if they want to volunteer.
The planning for the 1995 ACL Student Sessions will begin once this year's
student committee and the ACL exec committee chooses a student chair
from among the volunteers.
=============================================================================
COMPUTATIONAL LINGUISTICS JOURNAL, EDITOR'S REPORT
Julia Hirschberg
julia@research.att.com
In May 1993 James Allen began to transfer reviewing of CL submissions
to me, while retaining responsibility for pre-1993 submissions. James
continued to supervise the publication of journal issues, through
20:1. The transfer of all records and editorial responsibilities was
completed by the end of 1993, and very smoothly, thanks to James and
to Peggy Frantz.
Eighty-one manuscripts were submitted to CL in 1993, 22 of them to the
special issue on Computational Phonology (20:3). Four of the 22
submissions to the special issue were accepted, and another six asked
to resubmit revised manuscripts to a regular issue of the journal.
Overall, 15 manuscripts submitted in 1993 have been accepted, 37
recommended to be revised and resubmitted (of which six have been
resubmitted and are currently under re-review) or submitted to the
squibs section, 25 rejected, and no (first) decision yet on 4. Five
of the accepted manuscripts have been published (in 20:2 and 20:3),
four papers are in hand for 20:4 and three for 21:1. We are awaiting
delivery of final manuscripts for the remainder.
Submission and reviewing for the special issue on Computational
Phonology (20:3) was primarily electronic. This has been valuable
experience for plans to move CL to (primarily) electronic submission
and reviewing. Thanks to Steven Bird are in order, for orchestrating
this experiment, as well as for his exemplary work as a special
editor. Thanks also to Stuart Shieber, who is preparing new versions
of the CL LaTeX style file and a bibliography style format, which will
be needed before electronic submission can be instituted.
SUMMARY STATISTICS
Disposition of Manuscripts:
1993* 1992 1991 1990
Submitted 81 56 83 50
Accepted 15 10 15 16
Rejected 25 39 50 27
Resubmission recommended:
Squib 6
Regular 31
No decision 4
*Type of first decision
Mean time from receipt of manuscript to first decision: 125 days
1993 Submissions by Country (of First Author):
Argentine 1
Canada 4
Czech Republic 1
France 4
Germany 4
Greece 4
India 2
Israel 1
Korea 1
Japan 3
Netherlands 4
New Zealand 1
Saudi Arabia 1
Spain 1
Switzerland 3
Taiwan 5
UAE 1
UK 4
USA 36
=============================================================================
FINITE STRING
Ralph Weischedel
weischedel@bbn.com
Regarding publishing reports to the Executive Committee in The FINITE
STRING:
- Please try to limit your report to two pages or under in ASCII,
with the exception of the Secretary-Treasurer's report.
- Please review your report for appropriateness for publication
and submit your final version via email by Monday, July 26.
(Please submit plain text, not latex, troff, ..., since it must
be reformatted in MS Word for the Macintosh.)
Alysa Stark had been providing all the administrative support at BBN,
including formatting the announcements received and providing camera-ready
copy to MIT Press. With Alysa's increased responsibilities at BBN, we
transitioned this to Diane Bass, also of BBN. This caused some delays in
preparing issues 19-4 and 20-1.
The content during the last twelve months was the following:
Issue Length Calls for Announcements(#) Miscellaneous
(in pages) Papers (#) Items (#)
19-3 32 3 0 16*
19-4 24 11 1 2
20-1 20 10 0 3
20-2 16 7 0 2
*These 16 items were the annual reports to the ACL, which comprised the
overwhelming majority of the issue.
=============================================================================
COMPUTATIONAL LINGUISTICS JOURNAL, BOOK REVIEW REPORT
Graeme Hirst, Book Review Editor
gh@cs.toronto.edu
TIMELINESS OF PUBLICATION
We are continuing to get most reviews published in a timely manner --
or, at least, get them sent off to The MIT Press in a timely manner.
In general, reviewers are responding to the specific deadlines that I
set. The distribution of response time is bi-modal: those who are
overdue tend to be many months overdue.
VOLUME OF REVIEWS PUBLISHED
I am continuing to be fairly strict in deciding if a book is to be
reviewed. The ``Briefly noted'' section is a compromise between a full
review and total neglect.
In volume 19 (1993), we published 18 reviews and 18 brief notes, about
the same as in 1992.
PRODUCTION MATTERS
It would save me (and The MIT Press staff) lots of time if the complete,
documented LaTeX style file for book reviews were made available to
reviewers and editors, and if there were a consistent, published style
guide for references.
I continue to be indebted to Chrysanne DiMarco for long hours of
reading out loud with me to check the galleys.
=============================================================================
CL SQUIBS AND COMMENTS
James Pustejovsky and Bob Ingria
jamesp@cs.brandeis.edu
=============================================================================
COLING-94 PROGRAM CHAIR'S PROGRESS REPORT
Yorick Wilks
yorick@crl.nmsu.edu
The COLING94 program committee operated as a set of more or less
autonomous members, differing in area of expertise, and making final
recommendations to me based on the forms filled by the reviewers
themselves. In only one case, I think, did I interfere in that
recommendation: all other exercises of discretion by me were where
the committee member remained uncertain and, more generally, where
the final recommendations, as a whole, exceeded the slots available,
given their type (long or short). We did change the names of the
types between the first announcement and the final outcome, after
it became clear how hard it would be to provide platforms for
demonstrations in Kyoto with the resources available.
The acceptance/rejection ratio was almost exactly the same as in
1992, namely same as the last COLING, namely 1:2. This would be
insufficiently macho for more abrasive conferences, but I feel
sure the quality of this conference is good and will compare
favorably with its predecessors. It is hard to gauge the changes
in area of interest in the last two years since we have used a
different taxonomy (without separating off Applications) from
what was used at Nantes, but I suspect that the proportions are
very similar, perhaps with syntax having a small resurgence, and
empirical methods (broadly defined and not excluding syntax) still
running very strong.
A very pleasing feature is the wide spread of countries
represented (28), which has always been an important matter in
at COLINGs. They are, I believe, the most international
meetings in our field, more so than ACL or AAAI, which
remain fundamentally American meetings, a feature which makes the
whole review process much simpler, a matter I shall return to below.
As always, the host country is well represented, though
if you add together the papers from Germany and the UK they
total the same as those from Japan, and the summed population
of the two countries is also roughly that of Japan. By that
standard, Japan is not over represented at all, and
perhaps the US and other parts of Europe, as well as many
other countries, are under represented, with respect both to
their populations and the amount of relevant work being done
in them.
US under-representation abroad is always a result of
the no-repeat-papers rule imposed by US conferences like the
ACL, and we lost a number of accepted papers in that way this
year, which is why we took twenty "reserve" papers into the
printed proceedings. However, the US was well-represented on
the program committee and the ratio of papers submitted to those
accepted was pretty invariant across countries. So, this
under-representation remains a pity, especially in view of the
important role that US Government research programs, particularly
the ARPA Speech and Language Program, have had on the field
world-wide in recent years. That Program has done more than
any other to reestablish the importance of evaluation to
progress and the primacy of empiricism in CL/NLP; even though some
of the growth points came from elsewhere, particularly example-based
MT from Japan, and statistical corpus tagging from the UK. Hopefully,
the International Projects Day, chaired by Antonio Zampolli, will
give a platform for what has been done in the US under that program.
It must be said here that aspects of the program construction
process have made some people unhappy, including me: especially
(i) the fact that the date given for return of verdicts on
papers was missed; (ii) that referees' comments on REJECTED
papers were not returned to authors immediately, and (iii) there
were mailing errors of various sorts.
No author has complained specifically about (iii), but at least
one of the errors happened at Sheffield and may have contributed
to problem (i). I apologize for all the above, since I am responsible
though, as adults know, that is not the same as being to blame. As
to (ii), I decided initially that, as this was not done uniformly,
and COLING has little in the way of a formal budget, I would not
return the comments on rejected papers, even though I realised and
appreciated that much work had gone into them. However, the
volume of protests caused me to reconsider and these have
now been sent out, but I very much take the view that this was
done as a courtesy to colleagues and not because there is
some mysterious right to them, and especially not because
the Committee has to prove to any individual that its decisions
are transparent and not open to question. The last is a very
important point, and I will return to it below, but I am
certain that the reviewing process world wide cannot
continue to function if our community comes to believe that
decisions of this kind are justiciable, and can be
judged and questioned by legal criteria. That is a deep
cultural error, I am certain. The reliability of the process
as a whole rests on the collective reputations of the
committee as individuals in our research community, and nowhere else.
I have a particular perspective on this because I was also
the COLING84 Program Chair. Then as now, there was a slightly
amateurish chaos about arrangements, if anything it may have
been slightly worse that year because I was on sabbatical leave
in Costa Rica at the time, and running things with a worse
communication system than was generally available ten years
ago. There were glitches as always, but the conference was
as great success, as COLINGs normally are, and all was right
on the night.
Costa Rica was, in that sense, a symbolic center from which
to organize a COLING, in that the ICCL has always taken a
more world-wide view of the field than have the professional
societies which has meant, since global communications
are not evenly distributed, that the most up-to-date
methods have not been used: we have stuck to paper and mail, perhaps
wrongly, as the basis of reviews, and have had a wide-flung
committee, all of whom must report before decisions can be
taken. All that inevitably leads to a slower, bumpier
review process and we shall perhaps be able to change it in the
near future, given the rate at which the Internet is spreading.
However, in that decade, there have been other real changes in
this process that I have noticed and would like to comment on,
particularly in: (a) the expectation level from authors and
participants and, in particular, the belief of some of them that
the client-provider model is an appropriate one for the
author-program-official relationship and (b) the way the immediacy
and density of email can produce damaging side effects in procedures
like conference reviewing.
There have been, in recent times, four main modes of conference
organization:
a. The old ICCL way of organizing COLING, where the ICCL itself
reviewed abstracts and made quick decisions with no reports to
authors. A magnificent version of that was Hans Karlgren's
extraordinary performance as program chair of COLING90 in Helsinki,
where he read and commented on virtually every paper himself.
b. The fully funded style US meeting with a professional organization
(AAAI) or secretariat (ACL) and a paid committee meeting where all
issues MUST be settled to a schedule. This is a high cost method, but
deadlines are kept because any late sub-reviewer will have their papers
reviewed at the meeting by some part of the Committee. This requires
strong geographical proximity in the Committee which COLING does not
have, even if it had funds.
c. The conventional low-cost format that we have operated at this
COLING, with no committee meeting, and mailing paper backwards and
forwards across the world. There can be fully electronic forms of
that process, of course, but many COLING papers still come from
email-poor areas and groups which would penalize exactly the people
we have always favored, in some sense.
d. The streamlined conference arrangement that I see more frequently,
particularly coming from Japan, and now being used elsewhere: rapid
off-the-cuff comments and decisions from committee members by email and
with little in the way of formal requirements, all done in full
knowledge that a huge favor is being asked of busy reviewers, for no
reward whatever, and with an assumption that it probably does not
make much difference to the outcome, which is pretty robust across
different reviewing procedures.
As claims increase on the time of all of us, we will move, I
am sure, towards form (d) which is also, paradoxically,
close to the original COLING form (a). The ICCL might want to
move towards the "full accountability model" (b) as used
by most American professional bodies. It is high-cost,
efficient, and can probably only work well in a tight-knit
national community. Moreover, in many cases it can become
bogged down in issues of political correctness that the
international community usually finds less pressing that real
political issues, on which COLING has had an excellent
record: e.g. defending and supporting CL in Eastern Europe
during very difficult periods.
COLING remains the most amateurish of such meetings, without budget
or treasurer, and without fixed rules or a membership. It is not,
in a strict sense, accountable to anyone and, however much that may
be unacceptable to some, it does tend to produce nicer conferences.
These reflections, and their relationship to accountablity, have
been provoked by the changes over the decade I referred to above.
While doing this job in 1994, I have received an enormous amount
of email, demanding responses on everything from paper receipt to
margin sizes to hotel reservations to serious complaints about my
"defective consciousness of public duty" as one correspondent put
it. He was one of two I know of who chose to go straight to the
Internet with these complaints in very strong terms indeed, though
without any detailed knowledge of what was happening in the review
system and why.
I am sure all this criticism was richly deserved, but the
danger I see, beyond the purely personal, is that it will
become very hard to recruit colleagues to serve in any
of these roles (in journals as well as conferences) for
no reward, and certainly outside the fully-supported conference
of type (b), if the density and unpleasantness of much
of this email were to continue. The very immediacy of the
email medium, just banging out a message world-wide or to
individuals to relieve one's feelings and without thought
or knowledge, encourages this situation.
The source of much of it, and it comes, shall we say, from
the more litigious CL zones of the world in general, is
a belief that an author's relationship to a program chair
or committee member is one of client-to-provider,
to one who is accountable, for a fee, and can be sued. Only that
belief could explain the vehemence, the stridency, the
demandingness for accountablity of decisions, that I have
recently encountered. Such beliefs are, of course, very much
culture specific, and represent a transfer of the
attitudes associated with the fully-funded
responsible-to-a-membership professional meetings in the US
to the quite different situation of COLING.
The ICCL will probably have to ask itself, at some stage, if
it wants to follow model (b), though without a membership
that would be hard to fund, or perhaps to state a firm position
that COLING is not a conference of that sort and that older
attitudes, of a sort that used to prevail among academic
colleagues, where one was careful what one said because
one's roles would be reversed next time, may continue to
prevail. None of that would be incompatible with greater
efficiency in the organization of the program, and perhaps
with a fully automated form at the next COLING if it is
thought that the Internet has now penetrated all parts of
the known CL universe. Personally, I am sure COLINGs will
continue to be a little different from other meetings, to
the benefit of all of us and to the field we represent.
=============================================================================
APPLIED NLP 94 PROGRAM COMMITTEE
Paul Jacobs
pjacobs@unagi.cis.upenn.edu
=============================================================================
APPLIED NLP 94 LOCAL ARRANGEMENTS
Christian Rohrer and Uwe Reyle
rohrer@adler.ims.uni-stuttgart.de and reyle@adler.ims.uni-stuttgart.de
ANLP-94 will be held at the University of Stuttgart from Thursday October
13 to Saturday October 15, with tutorials on Tuesday October 11 and
Wednesday October 12.
Stuttgart is the state capital of Baden-W"urttemberg in south-west
Germany. It was founded about 1000 years ago as "Stuotgarden", a stud
farm, and today it is the cultural and commercial centre of the state, with
almost 600 000 inhabitants. For centuries the city was the residence of the
dukes and kings of W"urttemberg, and this epoch is much in evidence as you
stroll through the city centre.
One of the buildings of this epoch is the "Haus der Wirtschaft", the
primary ANLP-94 site. It is located just on the edge of the campus and is
just 5-minutes' walk to the city centre. Exhibitions, software
demonstrations and book exhibits will also be located at the "Haus der
Wirtschaft". Tutorials will be held on campus at Keplerstrasse 17. Net
access for participants will also be available there.
The Holiday Inn Garden Court will be our primary conference hotel. It is
located in the middle of the Weilimdorf Business Park. There is a tube
connection to downtown Stuttgart (about 15 minutes). Further rooms have
been reserved in hotels within walking distance of the "Haus der
Wirtschaft".
Location Rooms (single) Rate/night (DM)
Holiday Inn 200 100 .-
Maritim 50 259.-
Others 110 115.- to 140.-
A refectory meal service will also be available on campus.
We look forward to welcoming ANLP to Stuttgart.
=============================================================================
ACL-95 LOCAL ARRANGEMENTS
Robert Berwick
berwick@ai.mit.edu
I. GENERAL
ACL-95 will be held at the Massachusetts Institute of Technology
in Cambridge, MA from Tuesday, June 27 through Friday June 30,
with tutorials on Monday June 26 (following workshops on
Saturday, July 1). MIT and Cambridge are easily accessible by
land, air, and even sea: the usual NE routes, Logan
International airport and then subway, on the Kendall Station
Redline stop (for a much finer entrance reminiscent of Venice,
take the water taxi from the airport to Rowe's Wharf landing,
and then take the Red Line subway directly from nearby South
Station to Kendall Square). It is 2 more stops from the
conference site to Harvard Square. (Warning: by next year,
major construction will be underway on the roads leading to/from
the airport, and may be banned to car traffic, so renting a car
from the airport is not recommended. Taxis will still run, of
course.)
The primary ACL-95 conference is MIT's 900+ seat Kresge
Auditorium and Kresge Little Theatre, directly on campus and a
close walk to the MIT AI Lab and the Laboratory for Computer
Science and LCS. Additional lecture halls seating 400+ include
34-101, the main computer science lecture hall, and 26-100 (site
of the last ACL conference at MIT). It is a few minutes' walk
from any of these conference rooms to the main conference hotel,
the Cambridge Marriott, and 0.5km to the backup hotel, the
Cambridge Hyatt Regency, and the student dorm rooms.
Kresge Auditorium has substantial space for registration in
front of the main lecture hall. Exhibit space for software
demonstrations and book exhibits is available next door, in the
main student center (e.g., Sala de Puerto Rico). Network
service is available here. Security is available for these
rooms. Tutorial space is available in building 34.
All of these MIT rooms are available free of charge.
We look forward to welcoming ACL once again to MIT.
II. ACCOMODATIONS:
The following blocks of rooms have been reserved:
Location Rooms Rate/night
Cambridge Marriott 400 $129/night
single or double (ACL rate)
(on site)
Contact: Terry Hooper
Cambridge Hyatt Regency 100 $135/night
double $155/night
(300m away)
(deletable if Marriott
suffices)
Stephanie Krigman
MIT Dormitories 400 $41.00/single
(Eastgate suites) $58.00/double
(300m away)
Total Rooms 900
(No, this isn't New Mexico for prices)
III. RECEPTION:
The tutorial and conference receptions will be held at the MIT
Faculty Club, overlooking the Charles River. Note that the
rental cost of this room is free; we are charged only for drinks
and appetizers.
IV. BANQUET:
We have reserved the entire Boston Computer Museum, including
the Walk-Through Computer (complete with a huge PICT display of
our choosing); Tools and Toys; Robots and Other Smart Machines;
People and Computers; the auditorium. The selected caterer is
Creative Gourmet, which has done numerous wonderful company
functions for me in the past.
TRANSPORT to and from the banquet will be provided by shuttle
buses, either Crystal Transport or Boston Mini Coach
(negotiations underway for best price).
MUSIC for the banquet is tentatively planned to be INCA SON,
which plays music from the Andes of Peru and all of Latin
America; the group is widely known, and has appeared on PBS.
A contract to be signed is in hand.
V. OTHER MEETINGS
The Executive Meetings (Exec. dinner, ACL Journal, ACL book
series, etc.) will be held at the MIT Faculty Club function
rooms; the exception is the Exec. breakfast meeting, to be held
at the Mariott due to the early time. The Faculty Club room
rental is also free.
VI. STUDENT SESSIONS
Student Sessions will be held in 34-101 or in smaller rooms in
building 34 or adjacent. Box lunches for the student sessions
will be provided by Rebecca's.
VII. STAFF SUPPORT
A. Secretary support: we intend to hire a student, and make use
of two extremely capable staff assistants, Ms. Jacqui Taylor and
Ms. Mary Pat Fitzgerald. (Ms. Fitzgerald has has extensive
experience running conferences.) In addition we will receive
support from MIT Conference services, under the direction of Ms.
Gail Fitzgerald.
=============================================================================
ACL-96 ANNUAL MEETING SITE SELECTION
Doug Appelt, Vice President
appelt@ai.sri.com
It had been decided by the executive committee that the 1996 annual
meeting of the ACL will be held in western North America. I had attempted
to solicit bids from potential conference sites that would be consistent
with this predetermined constraint. Of all potential hosts that were
asked, two sites expressed particular interest in hosting the 1996 meeting:
the University of California at Los Angeles (Ed Stabler and Ed Hovy
volunteering to handle local arrangements) and the University of California
at Santa Cruz (Geoff Pullum volunteering to handle local arrangements).
Geoff Pullum withdrew UCSC from consideration before the submission of
a final bid because he felt that time pressures on himself, fellow faculty,
and students were intense enough that he did not feel that he could
guarantee that all possible contingencies were covered.
However, I had visited the UCSC campus myself, and I had fairly extensive
discussions with both Geoff Pullum and the UCSC conference office. We had
concluded that the facilities provided by UCSC would be adequate for the
annual meeting, and that the surroundings provided an interesting, pleasant,
and attractive atmosphere that would be conducive to a good conference.
Therefore Geoff Pullum said that UCSC is still interested in being considered
as a potential ACL host in the future, and would like to ask the executive
committee to consider UCSC as a possible site for ACL meetings.
UCLA appears to be an excellent conference site with an experienced
conference-center staff to provide support for attendees. Both UCLA and
USC-ISI have a large computational linguistics community that can
serve as a source for people to take on local arrangements tasks.
Facilities include a 600 person auditorium for plenary sessions, as well as
a 200 person meeting room for parallel and student sessions.
Seminar rooms and auditorimus are also available for tutorial and
workshop sessions. The conference center staff can provide all needed
catering and banquet services. On-campus housing is available in various
levels of luxury at reasonable rates, including dormitory rooms, and
hotel-like rooms at the conference center.
More extensive information regarding local arrangements can be
obtained by contacting Ed Stabler (stabler@biology.ucla.edu).
=============================================================================
TEXT ENCODING INITIATIVE
Judith Klavans
klavans@cs.columbia.edu
The Text Encoding Initiative has as its goal the setting of standards
for the encoding of various kinds of text, and of establishing
standard DTDs (Data Type Definitions) in SGML for the text processing
community. Don was a founding member of the Steering Committee. The
development of a comprehensive encoding scheme for text is an
important task, particularly as more text become available on-line and
through the Internet.
I attended the February meeting of the TEI Steering Committee. The
focus of this meeting was to help release the final version of the
latest standards before the ACH-ALLC conference in Paris, and to seek
future funding for additional and ongoing work of the TEI.
The editors, CM Sperberg McQueen and Lou Burnard, of TEI P3: the
Guidelines for Text Encoding for Interchange, did release the final
version of the guidelines on time, just before the Paris meeting.
This was an impressive achievement, considering that the hard-copy
version is over three volumes large, an important contribution to the
field. The Guidelines represent a virtually unparalleled spirit of
collaboration and a summoning of collective energy within the
scholarly and research communities to work toward a common goal. This
publication marks the end of the first development phase of the TEI
Guidelines.
Experts in Humanities computing have been awaiting the release of the
guidelines. Several tutorials are being given to explain its contents
over the next year. Funding is being sought to aid in the ongoing
work of the TEI.
>From a personal point of view, I had not understood the important
mission of the TEI to the extent that I did after the meeting and
after witnessing the impact of these standards. This work is of
utmost importance to the ACL, and should be supported in every way
possible.
=============================================================================
EUROPEAN CORPUS INITIATIVE
Susan Armstrong
susan@divsun.unige.ch
Status: The data has been prepared and a CD-ROM (ECI/MC1) is being pressed
by LDC.
DATA: 97 million words in 27 (mainly European languages).
MARKUP: Most of the data fully validated SGML document type description
based on TEI guidelines.
DISTRIBUTION: In the United States by LDC, in Europe by ELSNET.
RESTRICTIONS: Signed User Agreement stipulating non-redistribution and
agreement to respect particular copyright notices and special restrictions
as specified by the data providers.
DOCUMENTATION: Papers to be presented at pre- and post-Coling workshops.
STEERING COMMITTEE: Susan Armstrong (ISSCO, Geneva), Nicoletta
Calzolari (CNR, Pisa), Robert Dale (ELSNET, Edinburgh), Lori Lamel (LIMSI,
Paris), Mark Liberman (LDC, Phildelphia), Wolf Paprotte (Univ. Muenster),
Henry Thompson (Univ. Edinburgh).
LARGER CORPORA:
GER03 German Newspaper texts from the Frankfurter Rundschau
July 1992 - March 1993
Provided by Universitaet Gesamthochschule Paderborn Germany
Approximately 34 million words.
FRE01 French Newspaper texts from Le Monde
September, October 1989, and January 1990
Provided by LIMSI CNRS, France
Approximately 4.1 million words
DUT02 Extracts from the Leiden Corpus of Dutch
(newspapers, transcribed speech, etc)
Provided by Instituut voor Nederlandse Lexicologie, Leiden, Holland
Approximately 5.5 million words
MUL05 International Labour Organisation reports of the
Committee on Freedom of Association 1984-1989.
Parallel texts in English, French and Spanish
Approximately 1.7 million words per language
DATA COLLECTION AND PREPARATION: Carried out mainly at HCRC by D. McKelvie
and H. Thompson and at ISSCO by S. Armstrong and D. Petitpierre.
SUPPORT: Work done on a purely voluntary basis under the guidance of
the steering committee. In addition to the infrastructure support provided
by these institutions, modest financial contributions provided by ELSNET,
the LDC and NERC. The ACL and the LAnguage Technology Group of HCRC
provided loans.
=============================================================================
CONSORTIUM FOR LEXICAL RESEARCH
Louise Guthrie and Katherine Mitchell
louise@crl.nmsu.edu, lexical@nmsu.edu
Sharable Resources in Natural Language Processing:
1. Introduction
Sharable natural language resources are a timely topic that has
received a great deal of attention lately. Based on the ACL's
recommendation, the Consortium for Lexical Research (CLR) was
established with funding from ARPA in order to facilitate the sharing
of lexical data and tools. The CLR was set up in July of 1991 with an
Advisory Board consisting of Roy Byrd, Ralph Grishman, and Mark
Liberman, and is sited at the Computing Research Laboratory, New
Mexico State University,
The objective of CLR is to serve as a clearinghouse for software and
resources of importance to the natural language processing research
community. Sharable resources, and the task of centralizing lexical
data and tools, are of foremost concern in lexical research and
computational linguistics. Our goal has been to create and maintain an
archive of nlp materials which will help alleviate the repeated
re-creation of basic software tools, and assist in making essential
data sources more generally accessible. With a repository of over 275
items, ftp accesses 1993 averaging 8,000 per month, and a
membership roster which includes most of the major natural language
processing centers in the U.S., CLR has fulfilled many of the goals
envisioned for it at its inception.
Now, with almost 3 years experience behind us, we would like to share
our insights into sharable natural language resources, the obstacles
we have encountered in expanding the resource base available to the
nlp research community, and future needs and trends we foresee.
Additionally, we would like to mention the Mosaic WWW (World Wide Web)
interface to CLR resources.
2. Recent Results
Collection: The CLR repository has grown significantly in the last
year, becoming especially strong in tools for linguistic analysis. Our
holdings encompass lexicons, parsers, morphological analyzers,
dictionary tools, and a wide variety of other materials. We maintain a
descriptive catalog of all major holdings. This catalog has paragraph
length descriptions of over 140 items, for a variety of platforms,
which are all housed at our central ftp site repository.
Contracts: Together with our university lawyers, we have developed
contracts for members and for providers of materials. Negotiations
with dictionary publishers have been slow, but we now have
arrangements with Longmans and Harper-Collins publishers to facilitate
the purchase of their machine readable dictionaries by members.
Membership: We now have 60 members in CLR: 33 universities, 19
companies (including Apple, BBN, and Fuji Xerox) and 8 government
organizations. Membership growth is lead by acquisitions development;
as our repository becomes more substantial we have witnessed a
concurrent rise in membership inquiries.
3. Obstacles to Acquisitions
Our initial vision of CLR was a tripartite consortium of academia,
corporate entities, and dictionary publishers, which would allow
researchers access to Machine Readable Dictionaries without violating
copyright or intellectual property laws. CLR's goal of making MRD's
more widely available has been far more lengthy, complex and difficult
to achieve than we ever anticipated. We have succeeded in making these
crucial resources available for a small fee to university researchers,
but the cost for corporate researchers is still often prohibitive.
Universities and corporate members also have disincentives to
depositing material in the CLR. For universities, the time burden of
documenting and finalizing software so that other researchers will
find it usable often precludes their ability to make these resources
more widely available. Corporate research groups have both
competitiveness and profit motivated concerns. We suggest that a
structural solution to these obstacles would be to develop agreements
between federal funding agencies and grant or contract recipients
assuring that certain data or software would be made available through
CLR.
3. Future Trends for CLR
The value and usefulness of CLR to its members is determined almost
exclusively by the quality of its holdings. Members' needs have
defined two tasks for CLR in the upcoming year. The first is a
scouring of the networks to collect materials from around the world
that are freely available but not easily tractable. Centralizing these
resources and creating a "one stop" source for current lexical
information has proven to be a very valuable service for our members.
Our second task is to continue negotiating acquisitions of hard to
obtain materials from software developers, publishing houses, etc.
CLR and LDC (the Linguistic Data Consortium) are continuing to work
together, and we hope to soon offer CLR resources as an added service
for all LDC members. We believe our parallel tasks, LDC's emphasis on
corpora and creating lexicons, and CLR's emphasis on gathering
existing lexicons and tools, will be more efficiently served through
our combined efforts. Just as CLR's alliance with LDC will broaden
our acquisition efforts in the U.S., it is necessary to create
strategic international alliances to expand our acquisition avenues
abroad. European and Asian organizations with missions similar to
CLR's could exchange their archive holdings with ours; this symbiotic
relationship is being explored this year.
In the future CLR would like to provide administrative archival
services to decentralized research groups needing a focal location for
their shared resources. This service was provided in 1993 for the
Fifth Message Understanding Conference. It is anticipated that this
trend will grow and become a substantial activity for CLR.
As our membership grows to include linguists and lexicographers who
are not primarily computing researchers, it is imperative that we
facilitate their access to the internet system. In March of this year,
we set up a WWW Mosaic site that enables anyone to browse our catalog,
download a copy of it, read recent newsletters, and ftp those files
which are not secured for members only. Additionally, we offer links
to other Mosaic sites offering resources for nlp research. The WWW
address is: HTTP://CRL.NMSU.EDU/.
4. Summary
CLR has built a strong momentum, and is well positioned to capitalize
on the growing demand for sharable resources. Its nature as a mature
repository will include alliances with LDC and international archive
organizations, a new role as a shared resource facilitator for
decentralized research groups, and a lobbying effort to persuade both
funding agencies and research sites that certain outcomes of federally
funded projects be deposited in a centrally accessible location such
as CLR.
=============================================================================
LINGUISTIC DATA CONSORTIUM
Jack Godfrey
=============================================================================
ELECTRONIC DISTRIBUTION OF COMPUTATIONAL LINGUISTICS PAPERS
Stuart Shieber
shieber@das.harvard.edu
In April, 1994, an electronic e-print archive and server for papers on
the topics of computational linguistics, natural-language processing,
speech processing, and related fields was put into operation. The
CMP-LG server is organized by Stuart Shieber of Harvard University in
collaboration with Paul Ginsparg of Los Alamos National Laboratory
(LANL), with hardware and software provided by LANL. Although the
server has no affiliation with the ACL at this time, we provide a
summary of its structure and operation here because of the natural
interest that the membership has, and in consideration of potential
participation of the ACL in the future.
The server accepts e-mail submission of papers, typically in LaTeX,
and makes them available along with Postscript versions by e-mail,
ftp, gopher, and WWW. Subscribers receive summary listings each day
of papers added to the server. Essentially all operations, including
subscription and processing of submissions and mailings, are handled
automatically by software running at Los Alamos.
After two months of operation, the server has some 660 subscribers
from 36 countries receiving the daily listings of contributed papers.
Summary statistics on the subscribership are given below.
Coverage of the research in the area is difficult to gauge, but as of
June 9, well over half of the 50 papers in the regular and student
sessions of the ACL-94 meeting were available on the server. A total
of 64 papers were archived as of June 9, 1994, for an average of about
one paper per day.
We are pleased to mention the cooperation of the Consortium for
Lexical Research in New Mexico State University, which has agreed to
serve as a site for archiving large data sets and other ancillary
materials that are beyond the scope appropriate for storage directly
on the server. In addition, Eric Iverson at NMSU has integrated the
cmp-lg server with the ACL-94 on-line information, to provide a
prototype of an on-line proceedings for ACL-94.
Some worries have been expressed as to copyright issues for papers
submitted to the server. In general, the submitter is responsible for
legitimacy of submission. In the particular case of ACL proceedings
papers, the copyright agreement with ACL specifically preserves the
authors right to distribute electronic preprints freely. Discussions
with the executive committee and MIT Press on arrangements for
electronic distribution of papers published in Computational
Linguistics are in process.
A second set of worries has to do with the effect of the server on the
income of the ACL, by virtue of potential reductions in the purchase
of proceedings and memberships. A good indicator is the much longer
experience that the high-energy physics community has accumulated with
the hep-th and related servers. Anecdotal evidence indicates that
there has been no substantial effect of the server on subscriptions to
physics journals or membership in scholarly societies from the
operation of the hep servers. There is little reason to expect a
dramatic effect on ACL membership either, especially where ACL is seen
to be supportive, rather than obstructive, of such scholarly pursuits.
As the server is well underway, no action is required of the ACL at
this time to foster the continuation of this service. However, there
are several ways in which the ACL could aid in its promotion,
including, but not limited to: cooperation on copyright issues, and
advertisement and promotional aid. The executive committee may want
to consider such aid at the annual meeting.
Summary of cmp-lg Subscribership by Internet Domain
as of June 8, 1994
Domain Name Count
------------- --------------------------------- -----
1. at Austria 3
2. au Australia 11
3. be Belgium 6
4. bitnet 2
5. br Brazil 2
6. ca Canada - Ontario 20
7. ch Switzerland 10
8. cl Chile 1
9. com Commerical Internet clients 104
10. aol.com America On Line Clients 3
11. cz Czech Republic 1
12. de Germany 58
13. dk Denmark 2
14. edu Academic Internet clients 191
15. es Spain 25
16. fi Finland 7
17. fr France 30
18. gov Government Internet clients 5
19. gr Greece 1
20. hk Hong Kong 4
21. hr 1
22. ie Ireland, Rep. of (Eire) 5
23. il Israel 8
24. is Iceland 1
25. it Italy 5
26. jp Japan 22
27. kr Korea, Rep. of (South) 4
28. mil Internet Military clients 2
29. mx Mexico 1
30. net Internet Network providers 4
31. nl Holland (Netherlands) 22
32. no Norway 2
33. nz New Zealand 4
34. org Internet Organizational clients 6
35. pl Poland 3
36. pt Portugal, Madeira & Azores 4
37. ro Romania 4
38. se Sweden 11
39. su C.I.S. (Former U.S.S.R.) 1
40. th Thailand 1
41. tr Turkey 4
42. tw China, Rep of (Taiwan) 4
43. uk England (U.K.) 61
44. za South Africa, Rep. of 1
Summary of cmp-lg Subscribership by Host
top 20 hosts only
as of June 8, 1994
Host Count
--------------------------------------- -----
1. cogsci.edinburgh.ac.uk 12
2. research.att.com 10
3. dfki.uni-sb.de 8
4. das.harvard.edu 8
5. cs.columbia.edu 7
6. watson.ibm.com 7
7. divsun.unige.ch 6
8. coli.uni-sb.de 6
9. linc.cis.upenn.edu 6
10. ai.sri.com 6
11. let.ruu.nl 6
12. sfs.nphil.uni-tuebingen.de 6
13. parc.xerox.com 5
14. cl.cam.ac.uk 5
15. prl.philips.nl 5
16. stl.research.panasonic.com 5
17. cs.toronto.edu 5
18. ims.uni-stuttgart.de 5
19. ccit.arizona.edu 4
20. aisb.ed.ac.uk 4
=============================================================================
SURVEY OF COMPUTATIONAL LINGUISTICS COURSES
Bonnie Dorr
bonnie@umiacs.umd.edu
The Survey of Computational Linguistics courses was published as a
supplement and distributed with the 20(1) volume of Computational
Linguistics Journal this spring. The survey serves as a follow-on to
the Directory of Graduate Programs written by Martha Evens two years
ago.
In addition to actual responses from individuals, statistics on each
question have been compiled to provide the reader with a general
impression of how courses are being taught. A set of indices at the
end of the supplement provide: (1) a list of faculty members and the
universities where they work; (2) two lists of universities, one
alphabetized by university name and the other by
country/state/province; (3) a list of software available for
distribution, along with the e-mail address of the contact person.
Two important questions have been raised: does ACL want to continue
doing this, and if so, how often? In addition, Martha Evens has
offered some suggestions concerning future instantiations of the CL
Survey. One is that we might be able to satisfy this need with a file
available for anonymous ftp --- which we advertise in the Finite
String --- plus printouts that could be mailed to people who did not
access to net mail. If the survey were made electronically available
we would need to have a plan for executing updates and we would need
to decide who would maintain the compiled listing.
=============================================================================
GRADUATE DIRECTORY
Martha Evens
mwe@math.nwu.edu
The most recent Directory of Graduate Programs in Computational
Linguistics was published in paper form at the end of 1992.
I have a number of questions about what the next step should be,
if the Executive Board decides to produce another version.
1. Do you want to store this information in a file available for
ftp or in some other form accessible by email? If this file
were stored in machine-readable form, the information could be
provided much more rapidly, it could be refreshed more often,
it could be accessed and searched in a number of different
ways.
A. Would you want it to be stored on the ACL server? and/or
on a Sun Workstation at IIT?
B. Do you have preferences for formats and index types?
It is now indexed by country and state or province
and by faculty name. It could also be indexed by the
language of instruction or the languages processed or
or the degrees granted or other factors.
C. It would be possible to include more information in
machine-readable form. Do you have a policy? Should we
post whatever comes from each university within certain size
limits?
D. Do you have a schedule preference? For example, we could
request information formally once a year and accept it and
post it whenever it arrived.
2. Do you want to print a paper version? This makes it
accessible to libraries and to people who do not have email.
I suspect that this is a small group and that ad hoc paper
subsets could be printed on request quite cheaply.
3. Do you want to authorize me to keep on doing whatever you want
done? I have Tex and ASCII versions, form letters, etc.,
which I would be most willing to turn over to anyone.
=============================================================================
NLP SOFTWARE REGISTRY
Elizabeth Hinkelman
hinkelman@dfki.uni-sb.de
The Registry continues to serve as an informational facility
for natural language processing software, for an increasing range
and number of international queries. It has recently added a
World-Wide Web page to its distribution mechanisms.
Since the beginning of the year, we have also been involved in
the Relator project, an EC-sponsored pilot study on mechanisms for
sharing natural language resources. The emphasis in this phase
is on the design of a suitable institution that would ensure
quality and availability of a range of resources for the twelve
European languages. We are primarily involved in the software
evaluation and distribution area of this project.
The Registry itself remains fully international in character,
and expects to be more able to concentrate on its original target,
information for the academic community, as the more industrially-
oriented organizations mature.
=============================================================================
SIG GENERAL MATTERS: ACL SPECIAL INTEREST GROUP SPONSORSHIP
Doug Appelt, Vice President
appelt@ai.sri.com
The ACL provides support for a number of special-interest groups
(SIGs) that are organized and managed by members of the association who
have specialized research interests in various subfields of computational
linguistics. The ACL has encouraged members to form these special interest
groups since 1990. Because of the increasing number of such groups that
are seeking official recognition we have decided to codify the guidelines
this year by providing a model constitution for newly-forming SIGs to adopt.
During the past year, we have received two requests by members for
the formation of new SIGS. These requests include
SIG-NLL (Special Interest Group in Natural-Language Learning, organized
by David Powers)
SIGPHON (Special Interest Group in Phonology, organized by Steven Bird)
Both groups have sumitted draft copies of constitutions for approval, both
are in the stage of creating final drafts.
The following guidelines have been adopted for all ACL-sponsored SIGS:
1. The function of a SIG is to encourage interest and activity
in specific areas within the ACL's field, including linking this
with related areas, by such means as workshops, newsletters etc.
2. Proposals for a SIG shall be approved by the ACL Executive.
(A note on the procedure for proposing a SIG is attached.)
3. A SIG shall have a statement of purpose and constitution
approved by the Executive, and be designated as an ACL SIG.
4. A SIG shall have at least two elected offices, Chair and
Secretary, and hold elections for these positions every two
years. These two officers must be members of the ACL. The
election procedure must be specified in the constitution. A SIG
shall have a Liaison Representative (who may be an officer)
responsible for communication with members and with the
Executive.
5. A SIG shall provide a written report to the Executive each
year before the ACL Annual Meeting.
6. A SIG must have at least twenty-five members. It shall
maintain a list of its members and supply this to the Executive
along with its annual report.
7. The Executive may close a SIG on such grounds as failure to
meet required procedures, inactivity over a period of years, or
divergence from the interests of the ACL.
8. A SIG's statement of purpose, current officer information,
and annual report shall be published in the Finite String.
9. A SIG may hold workshops in conjunction with ACL meetings,
under the normal arrangements for workshops associated with such
meetings (Independent of other workshops it may hold).
10. A SIG may have access to the ACL memberhip list, and SIG
announcements may be included in normal ACL publicity.
11. A SIG must request Executive approval for any activities
jointly sponsored with any other non-ACL organisations.
12. The ACL will not provide any automatic financial support for
a SIG beyond that implicitly covered by the means already
mentioned, such as publicity. A SIG must cover its own running
costs, and any significant proposed expenditures must be approved
by the Executive. A SIG may request support for a Workshop
according to the normal ACL workshop guidelines, but funding for
such a workshop is not guaranteed. A SIG is not required to have
subscriptions from its members. Any proposal to support a SIG via
member subscriptions must be explicitly approved by the
Executive.
The following constitution has been proposed by the executive committee for
newly forming SIGs. It covers the main points in the guidelines,
while allowing flexibility for SIGs to adapt to their individual
circumstances:
CONSTITUTION OF THE ACL SPECIAL INTEREST GROUP ON
I. STATEMENT OF PURPOSE
The purpose of the Association for Computational Linguistics
(ACL) Special Interest Group on (the SIG)
shall be to promote interest in ; to provide members of the ACL
with a special interest in with a means of exchanging news of
recent research developments and other matters of interest in ;
and to sponsor meetings and workshops in that appear to be
timely and worthwile, operating within the framework of the ACL's
general guidelines for SIGs.
II. ELECTED OFFICERS
The elected officers of the SIG shall consist of a President, a
Secretary and . The President and Secretary shall be
members in good standing of the ACL.
The term of all elected officers of the SIG shall be
.
The duties of the President shall be
(1) To have primary executive authority over actions and
activities of the SIG.
(2) To prepare a written report on the activities of the SIG
for the Executive Committee of the ACL, for presentation to the
the ACL at its Annual Business Meeting.
(3) To designate a Liaison Representative for the SIG,
who shall be primarily responsible for communication
with members of the SIG, answering inquiries about the
SIG, and communication with the Executive Committee
of the ACL.
The duties of the Secretary shall be
(1) To maintain a membership roster of the SIG.
(2) To be responsible for any moneys awarded to the SIG by the
ACL; to collect and manage any dues that may be required by the
organization; and to present a written annual report on the SIG
finances to the Executive Committee of the ACL.
(3) To act as a Liaison Representative, who shall be
be primarily responsible for communication with members of
the SIG, answering inquiries about the SIG, and
communication with the Executive Committee of the ACL.
III. ELECTION OF OFFICERS
All officers of the SIG shall be elected by a vote of the
membership. The vote shall take place at least