Experience of using GATE for NLP R&D.
Hamish Cunningham, Diana Maynard,
Kalina Bontcheva, Valentin Tablan, Yorick Wilks
Department of Computer Science and
Institute for LAnguage, Speech and Hearing,
University of Sheffield, UK
GATE, a General Architecture for Text Engineering, aims to
provide a software infrastructure for researchers and developers
working in NLP.
GATE has now been widely available for four years. In this paper, we
review the objectives which motivated the creation of GATE and the
functionality and design of the current system. We describe some of
the ways in which GATE has been used during this time, and examine the
strengths and weaknesses of the current system, identifying areas for
This paper relates experiences in projects that have used GATE (General
Architecture for Text Engineering) over the four years since its initial
release in 1996.
We begin in section 2
with some of the motivation behind this type of system, and then give
a definition of architecture in this context (section
3). Section 4 briefly describes GATE;
section 5 covers a range of projects that have used
the system; and section 6 examines some additional ways
in which GATE has been used. These experiences form the input to section
7, which discusses the system's strengths and weaknesses.
If you're researching human language processing you should
probably not be writing code to:
- store data on disk;
- display data;
- load processor modules and data stores into processes;
- initiate and administer processes;
- divide computation between client and server;
- pass data between processes and machines.
A Software Architecture for language processing
should do all this for you. You will have to parameterise it, and sometimes
deployment of your work into applications software will require some
low-level fiddling for optimisation purposes, but in the main these
activities should be carried out by infrastructure for the language
sciences, not by each researcher in the field.
We can go further and say that you shouldn't have to reinvent components and
resources outside of your specialism if there is already something that
could do the job. A statistician doesn't need to know the details of the
IEEE Floating Point computation standard; a discourse processing specialist
doesn't need to understand all the ins and outs of part-of-speech tagging
(or worse still how to install a particular POS tagger on a particular
If you're a professional mathematician, you probably regard a tool like SPSS
or Mathematica as necessary infrastructure for your work. If you're a
computational linguist or a language engineer, the chances are that large
parts of your work have no such infrastructural support. Where there is
infrastructure, it tends to be specific to restricted areas. GATE, a General
Architecture for Text Engineering [ Cunningham et al. 1997], represents an attempt to fill
this gap, and is a software architecture for language processing R&D.
We now have four years of experience with GATE, work on which began in
1995, with a first widespread release late in 1996. The system is currently
at a pivotal point in its development, with a new version due for release.
The experiences reported in this paper contribute to the requirements
analysis and design of the new system.
3 Infrastructure for Language Processing R&D
What does infrastructure mean for Natural Language Processing (NLP)? What
sorts of tasks should be delegated to a general tool, and which should be
left to individual projects? The position we took in designing GATE is to
focus on the common elements of NLP systems.
There are many useful tools around for performing specific tasks such as
developing feature structure grammars for evaluation under unification, or
collecting statistical measures across corpora. To varying extents, they
entail the adoption of particular theories. The only common factor of NLP
systems, alas, seems to be that they very often create information about
text. Developers of such systems create modules and data resources that
handle text, and they store this data, exchange it between various modules,
compare results of test runs, and generally spend inordinate amounts of time
pouring over samples of it when they really should be enjoying a slurp of
something relaxing instead.
The types of data structure typically involved are large and complex, and
without good tools to manage and allow succinct viewing of the data we work
below our potential. At this stage in the progress of our field, no one
should really have to write a tree viewing program for the output of a
syntax analyser, for example, or even have to do significant work to get an
existing viewing tool to process their data.
In addition, many common language processing tasks have been solved to an
acceptable degree by previous work and should be reused. Instead of writing
a new part of speech tagger, or sentence splitter, or list of common nominal
compounds, we should have available a store of reusable tools and data that
can be plugged into our new systems with minimal effort. Such reuse is much
less common than it should be, often because of installation and integration
problems that have to be solved afresh in each case [ Cunningham et al. 1994].
In sum, we define our infrastructure as an architecture, framework
and development environment, where an architecture is a macro-level
organisational pattern for the components and data resources
that make up a language processing
system; a framework is a class library implementing the architecture;
a development environment adds graphical tools to access the
services provided by the architecture.
GATE version 1.n does three things:
- manages textual data storage and exchange;
- supports visual assembly and execution of modular NLP systems plus
visualisation of data structures associated with text;
- provides plug-in modularity of text processing components.
The architecture does this using three subsystems:
- GDM, the GATE Document Manager;
- GGI, the GATE Graphical Interface;
- CREOLE, a Collection of REusable Objects for Language Engineering.
Figure 1: Gate Architecture
GDM manages the information about texts produced and consumed by NLP
processes; GGI provides visual access to this data and manages control flow;
CREOLE is the set of resources so far integrated. Developers working with
GATE begin with a subset of CREOLE that does some basic tasks, perhaps
tokenisation, sentence and paragraph identification and part-of-speech
tagging. They then add or modify modules for their specific tasks. They use
a single API for accessing the data and for storing their data back into the
central database. With a few lines of configuration information they allow
the system to display their data in friendly graphical form, including tree
diagrams where appropriate. The system takes care of data storage and module
loading, and can be used to deliver embeddable subsystems by stripping the
graphical interface. It supports modules in any language including Prolog,
Lisp, Perl, Java, C++ and Tcl.
5 Projects involving GATE
Over the past 4 years, GATE has been used in a number of projects, not
only within the University of Sheffield, but also by a variety of
external institutions. In this section, we outline some
of the main projects that have used GATE, and examine its
performance in each case.
Goal: ECRAN (Extraction of Content: Research at Near-market)
[ Basili et al. 1997] was a 3-year EU funded research project with the main
aim of carrying out Information Extraction (IE) using adapted lexicons.
Participants: Thomson-CSF (Paris) (project co-ordinators),
SIS (Smart Information Systems, Germany), University of Sheffield,
University of Rome La Sapienza, University of Geneva,
NCSR ``Demokritos'' (Athens)
Description: GATE was mainly used in this project to
implement a general word sense disambiguation engine based on a
combination of classifiers.
Benefits: The modular architecture of GATE allowed this
to be carried out very rapidly.
Drawbacks: Two main disadvantages were found with
GATE. (1) The architecture was under development at the same time as
the word sense disambiguation engine. (2) The speed of database access
for the Tipster database was found to be slow for large amounts of
lexical data. The solution used was to store large amounts of lexical
data separately from GATE as gdbm hash tables.
Goal: The aim of the Cass-SWE project (A Cascaded
Finite-State Parser for Syntactic Analysis of Swedish) [ Kokkinakis and Johansson-Kokkinakis1999] was to create a
parsing system for fast and accurate analysis of large volumes of
Participants: Språkdata/Göteborg University,
Description: Cass-SWE implements a grammar as a modular set
of 6 small grammars. GATE is used to integrate all the required
software components into one system prior to parsing, and to enable
the results to be visualised in a user-friendly environment.
Benefits: GATE allows the tagging process to be carried out
sequentially, and enables modification of individual elements without
disruption to others. Using GATE as a visualisation environment also
enables the results of Cass-SWE to be further used in applications
such as IE tasks and additional semantic
Drawbacks: There were a few initial difficulties
understanding the workings of the GATE system, but problems originally
thought to be caused by GATE were later traced to the CASS parser.
Goal: The aim of the GIE (Greek Information Extraction)
project [ Petasis et al. 1999] was to develop a prototype named entity
recognition model for Greek.
Participants: NCSR ``Demokritos'' (Athens), University of
Description: The GIE system is based on the VIE system
provided with GATE, but requires different language-specific resources
such as gazetteers and grammars. Using GATE enables non-language
specific resources to be reused from the English version, thereby
saving time and effort.
Benefits: GATE facilitated significantly the
integration of existing and new modules in GIE, as well as the
validation of the final demonstrator. It was generally found to be
fast, easy to use and powerful.
GATE's demand for system resources as document size increases can
become a serious limitation. Complex compilation processes made the
embedding of static modules difficult. GATE also has some difficulties
supporting non-Latin languages. mostly relating to the GUI. Many
minor possible improvements to the GUI and to GATE in general (such as
the addition of new features) were identified during this project.
Goal: AVENTINUS (Advanced Information System for
Multinational Drug Enforcement) is an EU funded research and
development programme set up to build an information system for
multinational drug enforcement.
Participants: SIETEC (Germany), ADB (France),
Amt für Auslandsfragen (Germany), Bundeskriminalamt (Germany),
Sprakdata Gothenburg (Sweden), Institute for Language and Speech
Processing (Greece), INCYTA (Spain), University of Sheffield.
Description: AVENTINUS aims to collect information from
distributed international sources, using advanced linguistic
techniques to improve IE, involving multimedia resources and
Goal: SVENSK [ Olsson1997, Olsson et al. 1998, Gambäck and Olsson2000] was a 4-year project aimed at
developing an integrated toolbox of language processing components and
resources for Swedish.
Participants: SICS (Swedish Institute of Computer
Science), NUTEK, Uppsala University, Göteborg University,
PipeBeach AB., Telia Research AB, IBM Svenska AB
Description: The toolbox is based on the GATE language
engineering platform and incorporates language processing tools
developed at SICS or contributed by external sources.
Benefits: Each component has a standardised interface,
so users have the choice of working within GATE or selecting and
combining supplied components for integration into a user
application. GATE is useful in that it is not committed
to any particular type of data or task. The emphasis on modularity was
also found to be particularly appealing.
Drawbacks: GATE was at the time still in its early
phases and had some problems with very large-scale resources.
Specification of byte offset and I/O requirements for different
modules was also difficult.
Goal: LOTTIE (Low Overhead Triage from Text using Information
Extraction) was a demonstrator project for the GATE infrastructure. It
aimed to provide proof-of-concept by implementing demonstration
software dealing with the major technological problems involved in
Participants: University of Sheffield
Description: LOTTIE did not itself use GATE, but formed
a basis on which to prototype initial versions of release 2.
Parts of it were real, based on a
project in a different domain, and parts of it served as a test case
for GATE development and as a demonstration of future possibilities.
Goal: The aim of Eudico was a distributed multimedia
infrastructure supporting annotation of speech and video corpora
[ Brughman et al. 1998].
Participants: Max Planck Institute for Psycholinguistics
(Nijmegen, Netherlands), University of Sheffield
Description: Eudico enables transcriptions of utterances to
be time-aligned with speech and video data, so that dynamic and
simultaneous viewing and editing is possible. Integration with GATE
was carried out in order to benefit from GATE's ability to represent,
store and visualise linguistic data.
Benefits: The flexibility of GATE's data
model enabled the seamless integration between EUDICO's time-based
data and GATE's offset-based annotations. This enabled the
representation, manipulation and display of time-aligned
transcriptions into GATE's viewers, allowing the user to manipulate
the different types of data simultaneously in a uniform environment.
Drawbacks: There is a certain lack of support for
distributed/remote access to the document manager. Therefore in a
client-server environment, the entire data has to be sent over the
network instead of just the parts that are needed.
Goal: LaSIE [ Wilks and Gaizauskas1999]is an advanced large-scale IE
system, performing named entity recognition, coreference resolution,
template element filling and scenario template filling.
Participants: University of Sheffield
Description: LaSIE was designed specifically to work
within the GATE architecture, and led to the free distribution of
its counterpart, VIE, a base-line IE system. LaSIE modules within GATE
have also formed part of other customised projects within the EC
Fourth Framework (AVENTINUS and ECRAN).
Goal: EMPathIE (Enzyme and Metabolic Path Information
Extraction) was an 18-month research project aimed at applying
Information Extraction technology to bioinformatics tasks.
Participants: Dept. of Information Studies & Dept. of
Computer Science (University of Sheffield), Glaxo Wellcome plc.,
Description: EMPathIE aims to extract details of enzyme
reactions from articles in biomedical journals. The IE system is
derived from LaSIE and was developed within the GATE
Benefits: The embedding of EMPathIE within the GATE
environment means that many modules can be reused. EMPathIE thus makes
use of many of the LaSIE modules, and itself produces modules which
have been used for other related projects. Using GATE therefore
enables much of the low-level work in moving IE systems to new domains
to be carried out effortlessly.
Goal: TRESTLE (Text Reuse, Extraction and Summarisation
for Large Enterprises) [ TRESTLE2000] is a 2-year project involving
IE from electronic alerting bulletins distributed daily throughout the
Participants: Glaxo-Wellcome plc, University of
Sheffield Dept. of Computer Science and Dept. of Information Studies.
Description: TRESTLE is based on the
LaSIE IE system, but requires different
domain-specific resources, such as gazetteers and ontology, and
substantial modification of the discourse interpreter and template
Benefits: GATE provides domain independent linguistic
components for TRESTLE, the most important of which is the semantic
parser. Named Entity recognition requires only the installation of
domain specific gazetteers.
Strong computing skills are
necessary in order to make the most of the system.
Goal: PASTA (Protein Active Site Template Acquisition)
[ K. Humphreys and Gaizauskas2000] extracts information about protein structures
directly from scientific journal papers, and stores them in a template.
Participants: Depts. of Computer Science, Molecular
Biology & Biotechnology, and Information Studies (University of Sheffield).
Description:The system has been adapted to the molecular
biology domain from pre-existing IE technology such as LaSIE. The
progress so far demonstrates the feasibility of developing intelligent
systems for IE from text-based sources in the pursuit of knowledge in
the biological domain.
Benefits: The use of a common database for storing
intermediate results offers several advantages. GATE allows simple
integration of heterogeneous system components and algorithms. The
user interface is also attractive.
Drawbacks: GATE version 1 can be slow and memory hungry,
and is not very robust in face of changes in
versions of the C compiler used (gcc).
5.12 German Named Entity Recognition
Goal: German Named Entity Recognition [ Mitchell1997] was
an MSc project to adapt part of the LaSIE system to deal with German,
and to test whether the architecture was suitable for processing a
language other than English.
Participants: Dept. of Computer Science, Sheffield University
Description: The system followed the same general
architecture as LaSIE, but with modifications to various modules such
as the grammar and tokeniser.
Benefits: Using the GATE architecture meant that only fairly
minor modifications to individual modules were necessary, and rule
adaptation was easy. The evaluation of LaSIE as a tool for processing
other languages was very positive (as borne out by the later
development of M-LaSIE, a multilingual IE system).
The ability to group modules into blocks for
processing would be a useful addition, as would an easier method of
inserting new modules in the correct place.
6 Usage of GATE
In addition to the projects described above, GATE has been used as a
research tool, for applications development and for teaching. Some
examples of these uses are given below.
Our colleagues in the Universities
of Edinburgh, UMIST in Manchester, and Sussex (amongst others)
have reported using the system for
teaching, and the University of Stuttgart produced a tutorial in German
for the same purposes (see
Numerous postgraduates in locations as diverse as Israel, Copenhagen and
Surrey are using the system in order to avoid having to write simple things like
sentence splitters from scratch, and to enable visualisation and management
Turning to applications development,
ESTEAM Inc., of Gothenburg and Athens are using the system for
adding name recognition to their MT systems (for 26 language pairs) to
improve performance on unknown words.
Syntalex Ltd., of London, are developing a product that automatically applies
amendments to legal documents within the GATE framework.
Both British Gas Research
and Master Foods NV (owner of the Mars confectionery brand)
used the LaSIE system for competitor intelligence
systems. LaSIE was configured as an embedded library using GATE's
[ Gotoh et al. 1998] report experiments using
Named Entity tagged language models for large vocabulary connected speech
recognition. The modelling data was created by running LaSIE
as a batch process using GATE's command line interface.
Competitor intelligence researchers in Finland are using GATE in the
BRIEFS project [ Keijola1999] (Brief driven Information
retrieval and Extraction for Strategy), which concerns business
intelligence for companies. Although they had some initial
difficulties getting the system to run on their platform (Linux Red
Hat 6.0), they were able to get good results in a short period of
time, which was one of their main reasons for choosing GATE in the
The Polytechnic of Catalunya in Barcelona have used GATE as part of
the ITEM project for multilingual IE and retrieval
for Spanish, Catalan and Basque. They used GATE as a framework for
integrating their tools for different languages, and as a
visualisation tool for the results. They found it particularly useful
for providing a user-friendly environment for non-experts, although
its slow speed in processing large corpora was a problem.2
GATE has also been used for
Information Extraction in English, German, French, Spanish, Greek and
In these and other projects, researchers have contributed a diverse
and growing set of components to CREOLE, the collection of language
CREOLE now includes:
- the VIE3
English IE components
(tokeniser and text structure analysers; sentence splitter; several POS taggers;
morphological analyser; chart parser; name matcher; discourse interpreter);
the Alvey Natural Language Tools morphological analyser and
the Plink parser;
the Parseval tree comparison software;
the MUC scoring tools;
French parsing and morphological analysis tools from Fribourg and INRIA;
Italian corpus analysis tools from Rome Tor Vergata;
a wide range of Swedish language processing tools.
The main deficiency in this set is a bias towards language analysis, and towards
Processing Resources above Language Resources.
A partial list of GATE licensees (of which there were over 300 at the end of 1999)
is available at http://www.dcs.shef.ac.uk/nlp/gate/users.html.
Along with the experiences cited above, this list indicates that take-up of
the system is healthy and that LE R&D workers have found the system useful
in many contexts.
7 Strengths and Weaknesses
The strengths of the final version 1 release of GATE are that it:
- facilitates reuse of NLP components by reducing the overheads of integration,
documentation and data visualisation;
facilitates multi-site collaboration on IE research by providing a modular
base-line system (VIE) with which others can experiment;
facilitates comparative evaluation of different methods by making it easy to
facilitates task-based evaluation, both of `internal' components such as
taggers and parsers, and of whole systems, e.g. by using materials from the
ARPA MUC programme [ Grishman and Sundheim1996]
(whose scoring software is available in GATE, as is
the Parseval tree scoring tool [ Harrison1991],
and a generic annotation scoring tool [ Rodgers et al. 1997]);
provides a reasonably easy to use, rich graphical interface;
contributes to increased software-level robustness, quality and efficiency in
NLP systems, and to their cross-platform portability (all UNIX systems, Windows
NT and Windows 95; native support for Java, C++ and Tcl);
contributes to the portability of NLP systems across problem domains by
providing a markup tool for generating training data for example-based
learning (it can also take input from the Alembic tool [ Day et al. 1997] for
unifies the two best approaches to managing information about text by combining
a TIPSTER-style database with SGML input/output filters (developed using tools
from Edinburgh's Language Technology Group [ McKelvie et al. 1997]).
The principal problems with version 1 are that:
- It is biased towards algorithmic components for language
processing, and neglects data resource components (PRs vs. LRs).
It is biased towards text analysis components, and
neglects text generation components.
The database implementation is space and time inefficient.
The visual interface is complex and somewhat non-standard.
The task graph generation and management process does not scale beyond small
``GGI suffers from the scaling problem [ Burnett et al. 1987], as the size of the
custom graph quickly becomes unmanageable'' [ Rodgers et al. 1997].
Only the annotator component model is extensible; adding new viewers or tools is
Installing and supporting the system is a skilled job, and it runs better on
some platforms than on others (UNIX vs. Windows).
Sharing of components depends on sharing of annotation definitions (but
isomorphic transformations are relatively easy to implement).
It only caters for textual documents, not for multi-media documents.
It only supports 8-bit character sets.
Module reset cascades through all previously run PRs that made non-monotonic
Work is currently in progress on Version 2 of GATE, which aims to
combat some of these problems and extend the range of the system [ Cunningham2000].
Further details of requirements for this type of system, and how to
evaluate them, can be found in [ Cunningham et al. 2000].
Based on the collective experiences of a sizeable user base across the EU
and elsewhere, the system can claim to be a viable infrastructure for certain
sections of the field. Given further development, we hope that it can take
on this role for a wider variety of tasks.
This work was supported by EPSRC grants GR/K25267 and GR/M31699.
- [ Basili et al. 1997]
R. Basili, M. Pazienza, P. Velardi, R. Xatizone, R. COllier, M. Stevenson,
Y. Wilks, O. Amsaldi, A. Luk, B. Vauthey, and J. Grandchamp.
Extracting case relations from corpora.
ECRAN Deliverable 2.4 version 1.
- [ Brughman et al. 1998]
H. Brughman, A. Russel, P. Wittenburg, and R. Piepenbrock.
Corpus-based research using the Internet.
In First International Conference on Language Resources and
Evaluation (LREC) Workshop on Distributing and Accessing Linguistic
Reseources, Granada, Spain.
- [ Burnett et al. 1987]
M. Burnett, M.J. Baker, C. Bohus, P. Carlson, S. Yang, and van Zee P.
Scaling Up Visual Languages.
IEEE Computer, 28(3):45-54.
- [ Cunningham et al. 1994]
H. Cunningham, M. Freeman, and W.J. Black.
Software Reuse, Object-Oriented Frameworks and Natural Language
In New Methods in Language Processing (NeMLaP-1), September
1994, Manchester. (Re-published in book form 1997 by UCL Press).
- [ Cunningham et al. 1997]
H. Cunningham, K. Humphreys, R. Gaizauskas, and Y. Wilks.
Software Infrastructure for Natural Language Processing.
In Proceedings of the Fifth Conference on Applied Natural
Language Processing (ANLP-97), March.
- [ Cunningham et al. 2000]
H. Cunningham, K. Bontcheva, V. Tablan, and Y. Wilks.
Software Infrastructure for Language Resources: a Taxonomy of
Previous Work and a Requirements Analysis.
In Proceedings of the 2nd International Conference on Language
Resources and Evaluation (LREC-2), Athens.
- [ Cunningham2000]
Software Architecture for Language Engineering.
Ph.D. thesis, University of Sheffield.
- [ Day et al. 1997]
D. Day, J. Aberdeen, L. Hirschman, R. Kozierok, P. Robinson, and M. Vilain.
Mixed-Initiative Development of Language Processing Systems.
In Proceedings of the 5th Conference on Applied NLP Systems
- [ Gambäck and Olsson2000]
B. Gambäck and F. Olsson.
Experiences of Language Engineering Algorithm Reuse.
In Second International Conference on Language Resources and
Evaluation (LREC), pages 155-160, Athens, Greece.
- [ Gotoh et al. 1998]
Y. Gotoh, S. Renals, R. Gaizauskas, G. Williams, and H. Cunningham.
Named Entity Tagged Language Models for LVCSR.
Technical Report CS-98-05, Department of Computer Science, University
- [ Grishman and Sundheim1996]
R. Grishman and B. Sundheim.
Message understanding conference - 6: A brief history.
In Proceedings of the 16th International Conference on
Computational Linguistics, Copenhagen, June.
- [ Harrison1991]
Evaluating Syntax Performance of Parsers/Grammars of English.
In Proceedings of the Workshop on Evaluating Natural Language
Processing Systems, ACL.
- [ K. Humphreys and Gaizauskas2000]
G. Demetriou K. Humphreys and R. Gaizauskas.
Two applications of information extraction to biological science
journal articles: Enzyme interactions and protein structures.
In Proc. of Pacific Symposium on Biocomputing (PSB-2000),
- [ Keijola1999]
BRIEFS - Gaining Information of Value in Dynamical Business
- [ Kokkinakis and Johansson-Kokkinakis1999]
D. Kokkinakis and S. Johansson-Kokkinakis.
Cascaded finite-state parser for syntactic analysis of swedish.
Technical Report GU-ISS-99-2, Dept. of Swedish, Göteborg
- [ McKelvie et al. 1997]
D. McKelvie, C. Brew, and H. Thompson.
Using SGML as a Basis for Data-Intensive NLP.
In Proceedings of the fifth Conference on Applied Natural
Language Processing (ANLP-97), Washington, DC.
- [ Mitchell1997]
Named Entity Recognition in German: the identification and
classification of certain proper names.
Master's thesis, Dept. of Computer Science, University of Sheffield.
- [ Olsson et al. 1998]
F. Olsson, B. Gambäck, and M. Eriksson.
Reusing Swedish Language Processing Resources in SVENSK.
In Workshop on Minimising the Efforts for LR Acquisition,
- [ Olsson1997]
Tagging and morphological processing in the svensk system.
Master's thesis, University of Uppsala.
- [ Petasis et al. 1999]
G. Petasis, G. Paliouras, V. Karkaletsis, C.D. Spyropoulos, and
Resolving part-of-speech ambiguity in the greek language using
In Proc. of the ECCAI Advanced Course on Artificial Intelligence
(ACAI), Chania, Greece.
- [ Rodgers et al. 1997]
P.J. Rodgers, R.J. Gaizauskas, K.W. Humphreys, and H. Cunningham.
Visual Execution and Data Visualisation in Natural Language
In IEEE Visual Language, Capri, Italy.
- [ TRESTLE2000]
The TRESTLE project.
- [ Wilks and Gaizauskas1999]
Y. Wilks and R. Gaizauskas.
Report on epsrc research grant on the large scale information
extraction research project.
Technical Report GR/K25267, University of Sheffield.
1 (Matti Keijola, personal communication, October 1999)
2 (Jordi Batalla, personal communication, March 1999.)
3 A `Vanilla IE' system related to LaSIE
File translated from TEX by TTHgold, version 2.34.
On 9 Aug 2000, 13:14.