|
|
 |
Abstracts
|
|
The
Justice Web Collaboratory, Ronald W. Staudt
Legal
Aid and Unrepresented Litigants - Building Legal Decision Support
Systems for Victoria, John Zeleznikow
Public
Access to Justice, Ronald W. Staudt
Judicial Document Drafting Assistants:
14 Years After JEDA, Karl Branting
An Argumentation Shell for Supporting
the Development and Drafting of Legal Arguments, John Yearwood
and Andrew Stranieri
Data
Mining in Law with Association Rules, Author Andrew Stranieri,
Donald Berman
A Tale of Two Systems, David E. Woodin, Esq.
|
To
Be Taken At Face Value?, Michael Bromby
Sentencing
Support Technology In Pursuit of Public Safety, Michael Marcus
The
Paradoxes of JDSS: Ask Not What IT Does to Judges, but What
Do Judges Do to IT?, Cyrus Tata
JDSS’s for Criminal Sentencing: Sentencing
is Not Only a Judicial domain, Marius J.A. Duker
"Cyberjuries: The Next New Thing?", Nancy Marder
Towards a taxonomy of judicial thinking
FOO Check Teck
Discretionary Theory, Tunde Meikle,
Nanyang Technological University,
Evaluating Legal Decision Support Systems
Maria Jean J. Hall and John Zeleznikow
|
The Justice
Web Collaboratory
Ronald W. Staudt
Associate Vice President,
Law, Business, & Technology
and Professor of Law
Chicago-Kent College of Law
rstaudt@kentlaw.edu
The Justice Web Collaboratory1
(JWC) is a partnership between Chicago-Kent College of
Law and the National Center for State Courts (NCSC)2.
The JWC was formed to use the tools of the World Wide
Web to create a laboratory for collaboration aimed exclusively
at American judges. The JWC is organized as a communications,
publications and education web site with the primary goals
of:
1. Creation of an online community for
judges that encourages them to learn about the Internet
and to use its capabilities in appropriate ways to enhance
the performance of judges on the bench and in chambers;
2. Delivery of educational content and
research material to judges with special initial focus
on topics relating to technology, science and computing,
including case management, electronic filing, data integration
and knowledge management;
3. Establishing a living and vibrant laboratory
for improving the justice system in the United States
with special focus on public access to the courts.
The centerpiece of the Collaboratory is JudgeLink.org, a
web site designed exclusively for American judges. The site
is built to include a judges' information clearinghouse, continuing
judicial and legal education programs and an online think
tank. JudgeLink is designed to provide a secure forum for
the sharing of knowledge and exchanging of ideas between judges
of varying backgrounds and geography. This forum can serve
as a "just in time" learning network for judges who are unable
to consult with knowledgeable colleagues on issues of judicial
administration, technical information and other matters. This
password protected, members-only forum supports discussions
of developments across different jurisdictions as they occur,
without concern that conversations will be taken out of context
or publicized prematurely. Secure discussions of hot topics
are facilitated in a chat room, listserve and threaded discussion
forums. Since this debut, the Judicial Conference Center has
experimented with a variety of topics such as: courts in the
age of the internet, evidentiary issues related to technology,
politicizing the judiciary and pro se representation.
1. See
http://a2j.kentlaw.edu/, last visited May15,2001.
2. The Dixon Conference on Using the Internet to Improve the
System of Justice held on April 23-24, 1998 triggered the
development of the Justice Web Collaboratory and provided
initial funding through the National Center for Automated
Information Research. The responsibility for the management
of the Justice Web Collaboratory is coordinated by Ronald
Staudt, Associate Vice President for Law, Technology and Business
and Professor of Law, Illinois Institute of Technology, Chicago-Kent
College of Law and Roger Warren, the President of the National
Center for State Courts. Operational leadership is handled
by Chicago-Kent College of Law's Manager of the Justice Web
Collaboratory, Todd Pedwell and Marcia Koslov, Director, Knowledge
Services, National Center for State Courts.
Legal Aid
and Unrepresented Litigants - Building Legal Decision Support
Systems for Victoria
John Zeleznikow
Donald Berman Laboratory for Information Technology and Law,
Applied Computing Research Institute, La Trobe University,
Bundoora,
Victoria, Australia, 3083.
johnz@.cs.latrobe.edu.au;
The Donald Berman Laboratory for Information
Technology and Law has a long history of building Legal Decision
Support Systems in conjunction with Victoria Legal Aid. In
building such systems we have developed two significant tools
for placing legal decision support systems on the World Wide
Web:
a) A Knowledge Management Shell (based on Toulmin’s
Theory of Argumentation) to maintain and update legal knowledge;
b) A knowledge based system shell, WebShell
that enables KBS to be developed and executed on the web.
WebShell models knowledge using two modelling techniques;
decision trees for procedural type tasks and argument trees
for tasks that are more complex, ambiguous or uncertain.
With these tools we have built systems that
advise upon:
1) The distribution of property following divorce;
3) Eligibility for refugee status in Australia;
4) Eligibility for legal aid in Victoria;
5) Sentencing of criminals.
One of the major goals of Victoria Legal Aid
(VLA) is to improve efficiency. Placing an Eligibility for
legal aid system on the World Wide Web means VLA lawyers can
focus upon offering advice and litigation support, rather
than upon administering eligibility tests.
With state and federal governments decreasing
their financial support for Legal Aid, many indigent litigants
are being forced to represent themselves. In conjunction with
VLA, we are developing Legal Decision Support Systems on the
World Wide Web to help litigants represent themselves. In
particular, we have noticed that as in the United States,
in the mid-1990s at least one party was self-represented in
more than two-thirds of domestic relations.
Public Access
to Justice Project
Ronald W. Staudt
Associate Vice President,
Law, Business, & Technology
and Professor of Law
Chicago-Kent College of Law
rstaudt@kentlaw.edu
Startling statistics expose a huge shift by the customers
of courts away from the traditional methods of seeking resolution
of their disputes. More and more customers of the courts
have abandoned lawyers and are coming to court on their
own, pro se or pro per. The numbers are staggering. In some
jurisdictions nearly 80% of all divorces are prosecuted
with one or both of the parties unrepresented.
At the beginning of the new millennium, three venerable
institutions supported by three highly respected foundations
launched a project that will attempt to apply the lessons
of the new economy to the problems raised by these new court
customers. The new project called, Meeting the Needs of
Self-Represented Litigants: A Consumer-Based Approach, is
self consciously interprofessional. The world's leading
PhD granting School of Design, the IIT Institute of Design,
is providing the core methodology and professional design
staffing for the project. ID is joined by the country's
most technologically advanced law school, Chicago-Kent College
of Law, and the country's premier court research institution,
National Center for State Courts. The project seeks no less
than the redesign of courts for unrepresented litigants.
Perhaps it is too constraining to state the objective as
a redesigned court. The project is self consciously aimed
at getting close to the customer, understanding through
ethnographic observation, the needs and desires of the customer.
The project is unconstrained in the selection of solutions,
tools and new prototypes to meet the needs of this customer
group.
Months of customer observation by Design students and Law
students have poured boxes of data into a systematic design
process, a process developed and refined by Professor Charles
L. Owen. Through May 2001, the team of law students, design
students and National Center for State Court researchers
led by Paula Hannaford worked to refine the observations,
tune the proposed solutions and shape a new system. The
design process promises to produce a new approach that is
"different- freshly imagined to match the best of new technology
to emerging needs and interests, better- thoroughly and
systematically thought through for all users, and right
- sensitively positioned to meet environmental, personal,
social, and cultural needs"1
1. Charles L.
Owen, "Design, Advanced Planning and Product Development",
Chicago: Institute of Design, Illinois Institute of Technology,
p. 13.
Judicial Document
Drafting Assistants: 14 Years After JEDA
Karl Branting
Fellow of US Supreme Court
A judicial document-drafting assistant (JDDA) is a computer
system that helps judges draft orders or other routine judicial
documents by (1) insuring that the judge rules on each issue
necessary for resolution of the case before the judge and
(2) for each resolution of each issue, presenting the judge
with "boilerplate" text corresponding to the issue. Despite
the development 14 years ago of a JDDA with a demonstrated
ability to speed decision-drafting, this technology is essentially
unknown to judges today. This presentation describes the
central design elements of JDDAs and discusses the factors
affecting their acceptance.
An Argumentation
Shell for Supporting the Development
and Drafting of Legal Arguments
John Yearwood and Andrew Stranieri
This paper describes an argumentation shell to support
the formulation, representation and drafting of legal arguments.
The shell can be used to capture generic arguments in many
legal domains as well as to assist decision makers in constructing
their own actual arguments. The primary objective of this
research is to demonstrate that knowledge represented using
the generic/actual argument model (GAAM), a variant of Toulmin's
argument structure can be used to: support the development
of complex arguments, add context and increase specificity
for the retrieval of relevant documents, incorporate background
knowledge, assist in the drafting of documents that represent
arguments made and provide a structure for complex inferences
requiring a range of mechanisms. The shell can be used to
support decision making in a range of legal domains including
discretionary domains.
Data Mining in
Law with Association Rules
Author Andrew Stranieri, Donald Berman
Laboratory for Information Technology and Law,
La Trobe University, Australia
Data mining algorithms for the automated
discovery of association rules can be applied to data from
legal databases. The rules, that identify associations between
all possible combinations of variables are discovered automatically.
We demonstrate how interesting rules can prompt analysts
to formulate hypothesis for further investigation in Australian
family law. However, the techniques can also be applied
to directly support judicial reasoning. In that application,
association rules are invoked within a knowledge based system
shell to facilitate consistency and to suggest ways in which
arguments for desired outcomes can best be made. This is
demonstrated with hypothetical examples from sentencing.
A Tale of Two
Systems
David E. Woodin, Esq.
Due Process Software
P.O. Box 433
Catskill, NY 12414
dewoodin@gungaweb.com
Classical legal expert systems crafted in
the late 1980's from declarative languages such as Prolog
exhibited elegant approaches to problems within a limited
domain, however they were rather homely in appearance,
fragile in operation, required a willingness on the part
of the user to engage in some level of participation with
the machine, and consequently failed to achieve much market
penetration among lawyers.
With the advent of the web, a proliferation
of free and commercial on-line legal sources raises the
legitimate question of whether there is any place for
automated advisory systems in an online world seemingly
saturated with legal information
The thesis to be argued is that there is
indeed more need now than ever for such systems.
In a brief, but hopefully compelling presentation,
it will be demonstrated how the benefits of a present
web-based commercial application in New York State criminal
law spring directly from its roots in the prior rule-based
system, and its actual use in producing portions of the
web-based system.
Speculated is the conclusion that a practical
promise of classical rule-based systems may lie in their
employment as "intelligent authors" of conventional applications.
To Be Taken
At Face Value?
Michael Bromby
Glasgow Caledonian University
Biometric identifiers such as fingerprints, blood and DNA
samples, traces of hair and fibre can help in deciding whether
a suspect is linked to a crime scene. Before such scientific
procedures were identified and made commonly available for
legal proceedings, descriptions by eyewitnesses and the
victim were often the only means of suspect identification.
With the advent of closed circuit television (CCTV), many
crimes are now recorded by cameras in the public or private
domain, leading to a new form of forensic identification
– facial biometrics. Decisions on how to view and interpret
forensic evidence are important for both prosecution and
defence, not least for the judge and jury who must decide
the entire case.
Forensic evidence is often quantitative, with biometric
identifiers being measurable to some extent, and then compared
to establish similarity. In many cases, similarity is expressed
in statistical terms or as the probability of finding such
evidence given certain presumptions; the face value of these
figures may not mean much to a layperson such as a juror
who may be open to influence by either side of the adversarial
judicial process. The use of ‘hard science’ in court can
often be the point at which the members of the jury are
persuaded beyond all reasonable doubt. The common perception
of scientific testimony being unquestioningly true often
raises the issue of ‘mystic infallibility’ of such scientific
evidence whereby the jury may accept evidence too readily.
Whilst the merits of scientific technique can aid the secure
conviction of criminals, problems relating to the accuracy
and reliability of identification software may lead to miscarriages
of justice through convicting the innocent.
This paper looks at the use of CCTV to identify suspects
in the UK. By comparing eyewitness identification to computerised
facial recognition systems, it will be shown that the latter
may be more reliable for suspect matching. However both
matching reliability and admissibility may be questioned
in court, with or without the aid of expert witnesses. It
is currently possible to engage automated tracking of suspects
and questions arise as to the effects on society of such
invasions of personal privacy or ubiquitous mass profiling.
With future technological developments, it may be possible
to automatically identify patterns of behaviour through
CCTV cameras, aiming to predict whether an incident will
arise, thus allowing resources to be managed more effectively.
With such a wide-spread CCTV presence in the UK, the potential
of facial matching software could lead to the establishment
of a national face database that, unlike a DNA database,
could extend beyond forensic ‘trail’ evidence, leading to
identification of fraud, the uncovering aliases, or as a
means to validate identity.
Sentencing Support
Technology
In Pursuit of Public Safety
Michael Marcus
Circuit Court
Multnomah County (Portland)
Oregon, USA
Sentencing support tools represent the highest
calling and hold the greatest potential for decision support
technology in the courts. Recidivism statistics dramatically
reveal the dysfunction of sentencing discretion; legislative
and popular ballot initiatives have responded by encroaching
on that discretion. Other information technology offers
to supply sentencing judges with greater access to the
same information which has so far failed to reverse this
dysfunction: offender histories and, more recently, program
descriptions. Decision support technology has rarely been
aimed at sentencing decisions, and when applied to sentencing
generally compounds the error by presupposing that judges
(or sentencing models) should emulate past judicial performance.
Sentencing support tools properly constructed
and applied, however, allow judges and others involved
in sentencing to aim specific correctional responses at
the offenders most likely to be diverted from criminal
behavior by those responses. These tools therefore promise
to increase the public safety performance of sentencing
judges and correctional providers (public and private);
to justify the restoration of judicial sentencing discretion;
to improve public policy decisions deploying correctional
resources; and to alter the culture of criminal justice
by allowing rationality a full role in the process.
The Paradoxes
of JDSS:
Ask Not What IT Does to Judges,
but What Do Judges Do to IT ?
Cyrus Tata
Co-Director Centre for
Sentencing Research Law School
Strathclyde University
Glasgow, Scotland, UK
E-mail: Cyrus.Tata@strath.ac.uk
www.law.strath.ac.uk/CSR
At first blush JDSS’s imply the determination
of judicial decision-making. JDSS’s will, we have assumed,
affect judicial decision-making. Information Technology
might change the nature of decision-making, but an overlooked
important question is how does the decision-making process
affect the interpretation of information technology?
Rather than focusing attention on to what
extent and in what ways DSS ought to affect judicial decision-making,
this paper advocates the need to examine more seriously
how judges affect apparently immutable information and
knowledge. Far from the dream (or nightmare) of the computer
affecting and determining decision-making, it is possible
to detect that JDSS’s may be far more benign in practice
than either their designers or users had hoped or feared.
Informed by a critique of AI approaches,
this paper advocates a less behaviorist/ ‘black-box’ approach
in favour of using empirical research into the interpretive
way that legal decision-makers use discretion. The paper
aims to reveal paradoxes of JDSS’s including: judicial
resistance and data ‘neutrality’; the nature of legal
‘knowledge’ in IT; and performative functions of JDSS’s.
The paper is illustrated by observations
from ten years’ research and development work in designing
and implementing JDSS in sentencing.
JDSS’s for Criminal
Sentencing: Sentencing
is Not Only a Judicial Domain.
Marius J.A. Duker
(PhD, Department of Criminal Law, Free University Amsterdam)
Abstract for a presentation on JDSS’s and sentencing in
the Netherlands
The Dutch Public Prosecution develops prosecutorial
guidelines for sentences to be demanded at court. A rule
based JDSS supports the application of these prosecutorial
guidelines. The judiciary itself develops sentencing guidelines
and it develops a database in which elaborated judgements
(casuistry) are stored. Gaining consistency is seen as
mainly a task of judges themselves. Hence one speaks of
Judicial Decision Support Systems. In other words the
focus is on supporting judicial reasoning. Keeping attention
to the particularities of the case at hand, judicial reasoning
in sentencing could imply explaining the sentence by referring
to an applicable guideline and sentences as imposed in
like cases on the basis of that guideline. Nevertheless
in practice judicial reasoning actually implies the valuation
of relevant facts of the case at hand while from such
reasons alone any specific sentence can be derived. As
a consequence of this practice the said database primarily
aims at consistent judicial reasoning from the facts.
That approach necessitated cumbrous indexing of cases
by relevant facts in order to give insight in judicial
reasoning about those facts in like cases. Following that
approach the factual complex of like cases as such doesn’t
constitute the relevant information, but the judicial
reasoning about separate facts in those cases does. This
promotes scientific analysis of judicial reasoning in
like cases but it doesn’t promote an explanation why the
judge imposes for instance six years and not eight years
of imprisonment. This analytic approach together with
considering sentencing as a judicial domain hinders an
effective control towards the judge from parties in a
criminal procedure. One should therefore not aim at analysing
judicial thinking but at a practice in which parties in
individual cases derive from casuistry the margins of
appreciation (sentence bandwiths) for the judge to impose
sentences. One should not focus on supporting judges but
on the legislator’s responsibilities and on the possible
influence of prosecutors and lawyers towards the judge’s
decision-making. By elaborating the democracy principle
I propose that judges give reasons if they depart from
sentence-bandwiths in prosecutorial guidelines as approved
by parliament. And by elaborating the fair trial principle
I propose that in order to concretize, in addition to
the said bandwiths, the sentence to be imposed, the defence
and the prosecutor can derive arguments from relevant
casuistry as stored in a generally accessible database.
Prosecutorial guidelines and generally accessible casuistry
then are complementary. This approach simplifies the development
of prosecutorial guidelines and JDSS’s. The database is
not to be seen as a Judicial Decision Support System but
as a generally accessible collection of authoritative
cases which are relatively so few that indexing can be
very restricted. Furthermore, the proposed complementarity
of guidelines and casuistry only requires prosecutorial
guidelines to be formulated in general terms and doesn’t
require prescriptions for valuing relevant factors (i.e.
standardization). In that case a JDSS for applying these
guidelines is of no need.
"Cyberjuries:
The Next New Thing?"
Nancy Marder
Chicago-Kent College of Law
West Adams Street
Chicago, IL 60661-3691
tel.: (312) 906-5136
fax: (312) 867-2878
nmarder@kentlaw.edu
"Cyberjuries" lie at the nexus of a hallowed, legal
tradition and a new, revolutionary technology. The traditional
jury has long occupied a central role in the American
justice sytem: in criminal cases, it decides whether
the State will take a defendant's life or liberty; in
civil cases, it decides whether the defendant will be
held liable.
The traditional jury has been assigned this critical
role because of its unique institutional features. The
jury is a democratic institution in that it draws its
members from the citizenry rather than from officialdom;
its members come from the entire community rather than
from select groups; and jurors engage in group deliberation
and reach consensus through discourse rather than individual
votes. These features, which I have explored in other
papers, make the traditional jury uniquely situated
so that it is more likely than other institutions to
arrive at the right answer, to satisfy the parties,
and to render a verdict that will win community acceptance.
Although the traditional jury in the United States
has remained largely unchanged since our nation's founding
(with the exception of who can serve), we are now witnessing
the explosive growth of the internet, which is likely
to produce profound changes. The internet provides a
new means of access, of interaction, and of creating
and sustaining communities; in doing so, it unleashes
forces that challenge traditional institutions, such
as the jury.
This paper explores the challenges that the internet
poses to the institution of the traditional jury and
examines two related questions: Is there a significant
role for the "cyberjury"? And are there modifications
to traditional juries that can be made with the explosive
growth of the internet?
The paper begins with a brief description of the nascent
forms of cyberjuries that have recently appeared on
the Web, which allow parties to present civil disputes
to anyone who wants to serve as a "cyberjuror." The
paper then analyzes how and when cyberjuries might play
a positive role in resolving disputes, which institutional
safeguards they could borrow from traditional juries
to improve their decisionmaking, and which institutional
features traditional juries could borrow from cyberjuries
to streamline their processes.
One goal of this paper is to show that a cyberjury,
particularly if enhanced with traditional jury safeguards,
can serve as a dispute resolution mechanism for certain
kinds of disputes. Eventually, cyberjuries will be able
to contribute to society's judicial processes.
A second goal is to argue that traditional juries should
make use of the internet to remain a strong institution
in the 21st century. For example, one threat to the
jury system is the substantial sacrifice that traditional
jury service requires of jurors. To the extent that
summonses, questionnaires, and voir dire can be done
via the internet, traditional jury service will become
less onerous.
Finally, this paper explores how cyberjuries can function
in parallel with traditional juries, in ways that are
potentially harmless and harmful. This paper is the
first to explore how the internet might revolutionize
the traditional jury.
Towards a Taxonomy of Judicial
Thinking
FOO Check Teck
Nanyang Technological University,
School of Mechanical and Production Engineering
50 Nanyang Avenue
Singapore 639798
Email: mctfoo@ntu.edu.sg
Despite the availability of software
(eg. case base reasoning, CBR methodology as in CLIPS
or Java-based, object-oriented methodology) that may
be applied to support judges in formulating their
decisions, the very process of judicial thinking remains
little explored. Computer modeling work in this area
seems ipso facto to be driven more by the availability
of given technology (eg. computational linguistics;
informational retrieval of legal knowledge but based
on computer model). The technology adapted for legal
purposes appear able to support judges in making certain
classes of decisions (eg. in areas of sentencing,
personal damages). Yet it may be equally fruitful
o approach the problem of modeling from perspective
of judicial thinking: how judges think? This raises
the interesting question: do judges think differently?
If so how differently. Interestingly, for example,
recent experimental research that involved the mapping
of brain functions (using fMRI) captured how different
parts of the brain are “fired” depending on what had
been running through the brain. In this paper the
author argues for a more balanced approach: moving
research from the perspective of thinking then to
the modeling of judicial thinking. One remote possibility
that may be explored is to involve a statistically
significant number of judges in empirical, interview
studies of judicial thinking process. Another novel
approach may be to map the thinking processes of an
eminent judge such as Lord Denning from United Kingdom.
Such a study may yield insights of whether there is
a discernible pattern of judicial thinking across
cases in different areas of the law - contracts, equity,
administrative law and equity.
Discretionary
Theory
Tunde Meikle
University of Ballarat/Australian Catholic University
t.meikle@aquinas.acu.edu.au
In our work with the Refugee Review Tribunal
(RRT) of Victoria domain experts have revealed important
aspects of their decision making. Some of these aspects
related directly to the content of the legal domain and
the structure of the knowledge used in coming to decisions.
Using a modified Toulmin argumentation structure, this
knowledge is being implemented in a version of software
called EMBRACE. But as with all implementation, user experience
and ultimate acceptance depends on more than this knowledge
in the machine. It is also important that other aspects
of decision making are captured and supported by the software.
From our work we have found these aspects to be discretion,
consistency and resolution. Briefly, discretion indicates
the extent to which a decision maker is directed towards
a choice, consistency refers to the similarity of process
that the decision maker applies to each case, while resolution
is the granularity of focus on the case problem. It is
the first of these aspects, discretion, that is the subject
of this paper.
Our decision makers expect their exercise
of discretion to be unimpeded, so an understanding of
discretion as a theoretical construct is necessary in
order to effectively incorporate support for it into a
computerised system. A concise meaning of how discretion
is experienced by our experts has been explained by them
as the condition in which they are not directed (by their
legislation) towards a particular outcome. By outcome
the expert is including all choices along the way, not
just a final ‘solution’. In the course of processing a
single case, a decision maker (who works individually
at the RRT), will have to make many choices that lead
them to gather information that will develop into a case
for or against granting refugee status for an applicant.
There is no rule book that guides this process, and the
legislation that applies is general at best. This generality
is also referred to as ‘open texture’, and it is this
aspect that is often used to characterise a domain that
is called discretionary. This is because where there are
terms that can be defined only by use, then it meaning
is not fixed arbitrarily but is interpreted in context.
The choices about how to achieve this is the responsibility
of the RRT decision maker for example, and as autonomous
agents they require decision support to preserve this
capacity.
But beyond asking practitioners, there is
relatively little critical writing upon the construct
of discretion. It is clear that the higher the court,
the more the participants have it, and depending on the
legal system, one is more or less bound by the previous
choices of higher courts, this is the principle of binding
precedent. So at the core it appears discretion is a matter
of the capacity to exercise choice, although it may be
an informed choice. What is intriguing is the relationship
between information and choice, so that given more information
that must betaken into account, choice may be so constrained
as to be effectively directed towards an outcome. The
very condition our domain experts would not want. This
is one complexity. This paper further examines this and
other concepts related to discretion, decision making
in law and the implications these have for the design
of decision support systems.
Evaluating Legal
Decision Support Systems
Maria Jean J. Hall
Donald Berman Laboratory for Information Technology and
Law
Applied Computing Research Institute,
La Trobe University Bundoora,Vic., Australia 3083
Phone: 61.3.98168194
Fax: 61.3.94793060
jean_hall@one.net.au
and
John Zeleznikow
Donald Berman Laboratory for Information Technology and
Law
Applied Computing Research Institute,
La Trobe University Bundoora, Vic., Australia 3083
Phone: 61.3.94791003
Fax: 61.3.94793060
johnz@cs.latrobe.edu.au
Whilst much research has been devoted to
theories of knowledge representation and argumentation
for building legal decision support systems, very little
time has been spent on the essential issue of evaluating
such systems. In this paper we investigate the state of
the art of evaluation of legal decision support systems.
Using some principles from software engineering,
we introduce techniques for evaluating legal decision
support systems. We discuss how these principles can be
applied in evaluating two systems or laboratory has built:
Split_Up and GetAid.
|
|
