A Feasibility Study on the Computer Retrieval of
the Law of Quebec
or
The Care and Feeding of your Friendly
Neighbourhood Legal Computer
Robert Vineberg *1
Have you ever wished that there was a truly well-developed
index of Quebec law that could lead you to all the jurisprudence
relevant to a particular matter? Or have you ever spent hours
poring through cases, searching for that one judgment directly on
point? When one realizes that the jurisprudence of Quebec currently
amounts to approximately 120,000,000 printed words, and is growing
at the rate of 1,700,000 words per year, the enormous difficulty of
“thorough” legal research immediately becomes apparent. Luckily,
the ultimate legal scholar is waiting on the steps of our profession.
Though his parentage is obscure, his mental capabilities are aston-
ishing. He can read more in a minute than you can in a day, and
he can print about fifteen times as quickly as the fastest legal stenog-
rapher. He’s a thoughtful fellow too, one who would .almost in-
variably be able to answer your most complex legal question in less
time than it took to ask it.
Why is such a legal “wunderkind” not working for you right
to
now? The problem is, of course, his education. His eagerness
absorb great quantities of information requires several talented
instructors and considerable sums of money. While the cost of such
instruction and upkeep is not small, the rewards are commensurately
large. With the Quebec legal system today undergoing rapid evolu-
tion, there could be no more fitting symbol of progressive thought
and approach to our law than to be the first area of the world to
achieve the complete computerization of its law and jurisprudence,
so that the most current answer to every legal inquiry would be
but computer-seconds away.
How does one go about educating a computer? The first stage
is filling its memory with facts. Of the various systems developed,
Of the Senior Board of Editors.
‘With
thanks to Mr. Robert Shapiro of I.B.M. Canada who provided much
of the data used herein.
McGILL LAW JOURNAL
[Vol. 13
certainly the most satisfactory is the storing of the full text of
every Quebec case and statute in the computer. This is superior to
the headnote or digest-stor.ing methods in that it eliminates the
inadequacies of condensation at the input stage, and allows for the
great convenience of a full-text print-out at the output. How can
this massive task be done? While optical scanners have been devel-
oped which can “read” a printed page into the computer, their capa-
bility is as yet limited to only certain kinds of type of too narrow
a range for our purposes. Thus, until the development of a more
flexible scanner, this task must be performed entirely by key-punch
operators typing the full text of the material into machine-readable
form. It has been estimated that such a prodigous operation would
require approximately 15,000 man-days, at a projected cost of
$250,000.
Once the computer’s memory has been filled, it must be taught
to think, to sort the information and produce the desired responses.
Such programs, as developed in the United States, often take the
following form. The computer is first instructed to read the material
and compile a (bilingual) dictionary of every word used therein,
eliminating non-significant words such as connections, proper names
and numerals, and assign an index number to each word, indicating
its sources. Thence a Thesaurus, or Synonym Dictionary is fed into
the computer, instructing it that when ordered to search for a word,
this should include a list of its synonyms and their French-English
equivalents. Thus, the computer knows the locations of every word,
as well as the locations of each of their synonyms.
The final step is the search and print-out by the computer. The
most efficient and oft-employed search procedure is the Key-Words-
in-Combination approach. By this method the computer is asked to
find the cases in which a certain combination or sequence of words
appears, and other words do not appear. Thus for the simple example
of a man slipping on an icy sidewalk, one might ask the computer
for all cases which “icy” or “slippery”, “sidewalk” or “street” and
“fall” and their synonyms appear, but none of those wherein both
“automobile” and “collision” (or their synonyms) appear. The ma-
chine may then print out the citations or the full texts of the
relevant cases, whichever is desired.
Hence with great accuracy and speed, the computer can be
trained to perform the most laborious and time-consuming tasks of
legal research. What of its cost ? The price of such a system is
only partially determinable. While one can estimate the cost of
manually supplying the computer with the necessary basic data at
$250,000, the cost of programming is at present virtually impossible
No. 3]
STUDY ON THE COMPUTER RETRIEVAL
515
to evaluate. A major determining factor would be the availability
of programs developed by others, which could result in large savings.
These would include, for example, Irving Kayton’s SYNDIG, or
Synonym Dictionary Generator.2 The cost of computer time and
equipment, using the IBM 360/40 and associated hardware, would
(176 hours). However, this cost
approximate $14,000 per month
drops by almost 90% if an “overtime shift” is used on a computer.
That is, if a computer is used 16 hours a day rather than 8, the
rental charge is roughly 110% of the 8 hour cost. Finally, the costs
of operating persomel, once the system is functioning, may be
estimated at $4,500 per month.
Who could undertake such a task? It would seem that the most
appropriate institution would be either a university or the provin-
cial government. Both are already major users of computers and
have the skilled staff required to establish such a system. Nor would
such a program lack remunerative appeal, for if the capital require-
ment were $500,000 and 50 “requests” were made per day, each
being charged $25, the system would pay for its maintenance ex-
penses plus a return on investment of approximately 25%.
Thus, with a legal data centre only a telephone call away, and
cases being fed directly from the “redaction” offices of the court-
houses to the computer, legal research, of primary importance not
only to the practitioner but to the jurist and academic as well, would
become greatly simplified and streamlined. The road to computeriza-
tion could bring little but benefits to Quebec law, and its attainment
may signify a breathtaking achievement in the search for a modern
and truly contemporary legal system.
2 1. Kayton, “Retrieving Case Law by Computer”, 35 George Washington Law
Review 1.
