draft september 2003webdocs.cs.ualberta.ca/~smillie/lgp30/lgp30rev.doc · web viewdepartment of...
TRANSCRIPT
January 2004
Revised.: July 2004, March 2005
The LGP-30 and its Effect on Computing at the University of Alberta
Keith SmillieDepartment of Computing Science
University of Alberta
AbstractSome computational methods at the University of Alberta before the introduction of electronic
computers are described briefly. The University’s first computer, the LGP-30 which was acquired in
1957, is described and examples are given of its use.
IntroductionFor almost forty-five years there was a small two-story brick building measuring some eighteen by
forty feet immediately to the north and west of the Arts Building on the University of Alberta campus. It
was built during the Second World War for testing aviation gasoline and cost $12,671.24 with an
additional $3987.17 for equipment. It was known, informally at least, as the “Gas Lab”. In 1954 the
testing work was moved to the newly constructed Alberta Research Council
laboratories on the campus. The Gas Lab was then occupied by the Faculty of
Arts and given the more formal name of Arts Building Annex. As one of its uses
was for housing mice used in experimental work by the Department of
Psychology, it was known to some as the “Mouse House”. It was demolished without fanfare during the
summer of 1986. No photographs appear to exist in the University Archives. All that remains are a few
bricks which were taken as souvenirs from the hole left by the demolition crew before it was filled. One
of these bricks is shown here.
Our present interest in this minor aspect of the University of Alberta’s history is because in the late
1950s the Arts Building Annex was home to the University’s first computer, the Royal McBee LGP-30.
At sometime during the 1960/61 academic year, quite possibly at the beginning, the computer and
supporting staff moved to the newly opened Physical Sciences Centre. The LGP-30 was replaced by an
IBM 1620 in 1961 although it continued to be used for teaching purposes for two more years. It was then
placed in storage, safely some of us thought, for about ten years when it disappeared without a trace.
The purpose of this paper is to attempt to reconstruct over forty years later the computing milieu at
the University of Alberta just before the arrival of the LGP-30 and to show some of the effects that it had
on computing practices. Unfortunately most of the documents relating to the use of the LGP-30 are lost
and all of the persons who were involved have retired and many have left the city. Fortunately a few are
still working and come to the campus regularly. Sadly though others have died or are in failing health.
However we shall do our best with help from those persons and records remaining to reconstruct a small
episode of local academic computing history.
As planning has already begun to demolish part of the Physical Sciences Centre, possibly a few notes
taken from a booklet1 prepared for its official opening on May 24, 1961 might help prevent it from
suffering a fate similar to that of the Arts Building Annex. The Centre was designed as a complex of a
Physics and Mathematics wing and a Chemistry wing at right angles to each other, and a shared lecture
wing and also an auditorium and library. The total floor area was 376,000 square feet. It was intended to
accommodate an expected total student population of 6,500 with some room for expansion. (The estimate
of the present total student population is 33,729.) The construction cost was $7,600,000 and an additional
$900,000 was allowed for equipment. Construction proceeded ahead of schedule and most of the space
was usable at the beginning of the 1960/61 academic year.
Before the LGP-30: MathematicsAn excellent account of mathematical and statistical computing in the Department of Mathematics at
the University of Alberta is given in a short departmental history2 written by E. S. Keeping. Professor
Keeping joined the Department of Mathematics in 1929 and retired in 1961, having been Head for the
previous seven years. He continued to lecture for the following nine years, and maintained his association
with the University for many more years. He died in 1984 in his eighty-ninth year.
An early set of mathematical tables used in the Department of Mathematics was Campbell's
Numerical Tables prepared by Professor J. W. Campbell who came to the University in 1920 and was
well-known for his work in astronomy and classical mechanics. The tables were divided into two parts,
the first giving tables for the common logarithm, square, cube and reciprocal, the circular functions, and a
short table of exponential and hyperbolic functions, and the second giving an extensive table of
hyperbolic functions which, according to Professor Keeping, Professor Campbell "himself calculated on a
hand machine". It was printed locally, first appearing in 1929, and was reprinted in 1946. At some
subsequent date Knott's Four-Figure Mathematical Tables, first published in 1900, was introduced into
the Department and was used until the mid 1960s. These tables were loaned to students during
examinations, and were clearly marked with bold red lines and a cautionary warning on the cover that
they were the property of the University of Alberta and were to be neither defaced or mutilated nor found
2
in "private possession". Eventually, these tables were remaindered in the University Bookstore for
twenty-five cents each.
The Department of Mathematics moved into the Physical Sciences Centre when it opened occupying
the fifth floor. There was office accommodation for 25 full-time teaching staff and about the same
number of part-time teaching assistants and graduates students. There was
also a computing room shown in the figure to the right which has been taken
from booklet already referred to1. (Professor Keeping is seated at the front.)
Professor Keeping was well aware of the advantages of the use of mechanical
desk calculators in statistical calculations, and has the following remarks
about their use in the Department:
Equipment is another item which has increased significantly in cost in recent years. For a
long time the laboratories in elementary statistics used small Monroe calculators [and also the
Multo calculator shown in Figure 2] which were cranked by hand, and there were only one or two
electric desk-calculators in the whole department, of rather old-fashioned type. By 1962 many of
the old calculators were almost worn out, and some newer types of hand calculators were
purchased. [These were probably the Swedish-made Odhner calculators which weighed thirteen
pounds and which accomplished multiplication and division by addition and subtraction with
repeated shifting.] A little later some improved electric calculators, such as the Frieden square-
root type, came along and then some desk electronic models were purchased.
The Multo Calculator shown here was also used for statistical calculations in the Department of
Mathematics until the 1960s. The only other reference that Professor Keeping makes to equipment prior
to the late 1950s is the following: “In 1955 the department purchased an electric kettle to be used in
preparing tea or coffee before the [Mathematics] colloquium, and the staff were assessed one dollar each
to pay for it.”
Professor Keeping was the co-author of a two-volume text on
mathematical statistics in which a large number of carefully
worked numerical examples were a prominent feature. In the
introductory chapter of the first volume3 there is a short section
entitled "Calculating Machines" in which we read the following:
"A calculating machine is constructed to add and subtract. By
means of continued addition or subtraction, operations involving multiplication, division, and square root
can also be performed with great speed." Later in the same section he gives the details of performing
repetitive calculations using as an example finding the values of "12 + 6x for x = 5, 7, 15, 12, etc."
Finally, it is interesting to note that Professor Keeping's lectures in the Faculty of Engineering had an
3
influence on the design of the pocket calculator. Two of his former students who were subsequently
employed by Hewlett-Packard remembered his lectures on reverse Polish notation and incorporated this
feature into the company's calculators4.
Before the LGP-30: AgricultureThe Faculty of Agriculture has always required calculating machines for the analysis of experimental
data. In Canada an excellent example of the importance of proper statistical designs and efficient
calculating procedures was given by Cyril H. Goulden who worked for the then Canada Department of
Agriculture, first at the Dominion Rust Research Laboratory in Winnipeg, Manitoba and later at the
Central Experimental Farm in Ottawa. In the 1952 edition of his Methods of Statistical Analysis, first
published in 1939, he makes the following remarks:
In the development of each procedure an attempt has been made to form a uniform method. After
general statements the algebraic development is given, and then is followed by a completely
worked-out example.
The book contains a very large number of numerical examples for everything from the simplest statistical
procedures to the very complex lattice designs used in agricultural work. From the late 1950s Goulden
took a keen interest in electronic computers, encouraging their use in the Department of Agriculture and
even speculating if the Department should build its own computer.
For many years the biometrics course in the Faculty of Agriculture at the University of Alberta used a
statistics manual prepared by Professor L. P. V. Johnson5. This carefully written document with its
numerous worked examples clearly follows the format adopted by Goulden. Indeed the first edition of
Goulden’s text is one of the 16 papers and books cited in the References.
The Department of Soil Science has carried out fertilizer trials since the early 1920s. However, it was
not until 1956 that statistical designs such as randomized blocks and balanced lattices were introduced.
For the first two years desk calculators were used for the calculations. One of the machines used was a
Monromatic which allowed the operator to accumulate at one time the sums of squares of two sets of
observations and their cross products. This was accomplished by entering pairs of observations, one pair
at a time with one datum on the left side of the keyboard and the other on the right. The cumulative sums
of squares would be displayed on the left and right sides of the accumulator dial with the sums of cross
products appearing between them. Also a counter mounted on one side of the calculator would show the
number of pairs of observations involved in these sums.
One final computational example from the Faculty of Agriculture is of interest because it forms a link
between the use of hand calculators and the LGP-30. It is a report entitled "On the meaning of marks"
written by Brian Hocking, Head of the Department of Entomology, and dated June 4, 1958. It gives the
4
results of an analysis of the variations in the distribution of the final marks in some selected courses at the
University of Alberta, and makes recommendations for removing inequities caused by these variations. In
the Procedures section of the report we read that "The first calculations were done on a desk calculator;
when a digital computer [the LGP-30] became available, this was used to complete the work." This report
is written in a style evocative of a gentler age at the University, and opens delightfully with the words
"For my personal guidance in trying to deal justly with my own classes, I set out to determine, during a
few of the less crowded hours last summer, ...".
Before the LGP-30: PhysicsThe first use of an electronic computer on the campus was in the Department of Physics which in
May 1957 established a link with the FERUT computer at the University of Toronto6. An account of this
work was reported in the Summer 1957 issue of The New Trail, the University of Alberta Alumni
Magazine, in a short article entitled "Electronic brain aids University research"7. It is of interest to quote
from this article as an example of how the computer was presented to the public:
Any problem that can be reduced to a numerical analysis and requires a numerical answer can
now be answered in minutes on the U. of A. campus through the medium of teletype and FERUT,
a high-speed digital electronic computer housed in the computation centre at the University of
Toronto.
A direct-line, teletype communication system with FERUT has been placed with the
University through the courtesy of the Canadian National Telegraphs, on an experimental basis.
Equipment is located in the Physics Department laboratory, basement floor, Arts Building.
This new type of "correspondence course" has students and physics professors at the U. of A.
preparing problems on teletype ticker tape every week. Then, on Thursdays at 5:00 p.m., a direct
line with the computation centre in Toronto is cleared for University use. Transmission time is
made available by the Canadian National Telegraphs and time on FERUT is paid for by the
National Research Council.
The mechanics of getting a solution to any mechanical problem are comparatively simple.
Previously prepared tapes of problems are fed through the teletype machine on the campus and
identical tapes are punched instantaneously in the computation centre in Toronto. FERUT is fed
the problem tapes which are then processed and FERUT feeds back an answer tape in typed
numbers in tabular form. The answer tape is fed through a Toronto teletype which activates the
keys of the machine on the campus.
There are many particular advantages of the new hook-up. Ambitious problems in the fields
of physics, mathematics, engineering, statistics, etc., can now be solved on the campus in a matter
5
of minutes, that otherwise would have taken months, perhaps years, of hard labour on a desk
calculator. Problems of a numerical nature can be solved much more accurately than previously,
and through the teletype medium, operators at both ends can converse.
It is interesting to compare the above account with some recollections made many years later by Don
Betts, a theoretical physicist at the University of Alberta who was very much involved in using FERUT
and later the LGP-30:
During the 1956/57 academic year [the theoretical physicists] decided to try to gain access to
FERUT. Don Scott, who was then Assistant Head of the Department, was enthusiastically
supportive of our goal and was instrumental in completing the necessary arrangements.
The director of FERUT, Dr. C. C. Gotlieb, agreed to let us have access to it, Canadian
National Telegraphs agreed to lend us a teletype machine and to provide free use of a telegraph
line from Edmonton to Toronto one evening a week, and the National Research Council provided
for the cost of time on FERUT. By April 1957 all arrangements were in place, and I was sent to
Toronto for two weeks to learn to write programs and prepare input tapes for the machine. I was
helped by Dr. B. H. "Trixie" Worsley and Miss Dorothy Goulding of the Computation Centre in
learning Transcode, a high-level language for FERUT which was written at the University of
Toronto. I did not learn the much more difficult machine-language programming. ...
The teletype machine was established in a glorified closet in the basement of the Arts
Building and the teletype link was first used on May 9, 1957. The official opening of the facility,
which included a ribbon cutting, was attended by the President of the National Research Council,
Dr. E. W. R. Stacie, the Director of the Western Region of Canadian National Railways, the
President of the University of Alberta, Dr. Andrew Stewart, representatives of the media, and by
all directly involved.
Throughout the summer we and our students would prepare programs and code teletype tapes
through the week for transmission on Tuesday evenings. [The article in the New Trail cited
previously gave the day as Thursday.] The two copies would then be run through a mechanical
comparator at the Toronto end. Any discrepancies were fixed up by the exchange of teletype
messages. The system worked tolerably well except when there was a thunderstorm anywhere
between Edmonton and Toronto!
Before the LGP-30: EngineeringThe experiences of other departments and faculties would appear to be similar to those of
Mathematics and Agriculture with calculations, many of them statistical, being performed on Marchant,
Munroe and Frieden calculators. It is unfortunate that few records appear to exist of this work. However,
6
a few remarks on engineering calculations are given in a history of the Faculty of Engineering4:
Their education was highly structured, free of options, very basic and without
computers. True, there were the standard six-place log tables and one volume of “Vega”
eight-place tables and ten-inch slide rules. The Faculty owned the computer of that age, a
30-foot spiral slide rule which gave five figures at the low end and four at the upper end
but it took a hefty pair of arms to muscle it about.
There was also a 42-inch wooden slide rule, yellow with black markings, which hung for many years in
the University Bookstore over the counter where slide rules were sold. During the early 1980s it was
borrowed occasionally to demonstrate the operation of a slide rule to students, many of whom had never
seen one. It is now part of a permanent exhibit of early calculating devices and machines in the offices of
the Faculty of Science8.
Acquisition of the LGP-30We have seen that in the mid 1950s professional and academic computing needs at the University of
Alberta were being met by mathematical tables, mechanical and electromechanical calculators, slide
rules, and then by a few physicists using the University of Toronto's electronic computer.
In May, 1957 the President, Dr. Andrew Stewart, appointed a "Committee on Electronic Equipment"
to make an assessment of computing needs at the University. The seven-member committee was chaired
by Don Scott, with Don Betts as Secretary and Professor Keeping as one of the other members. In July
the Committee recommended unanimously the purchase of an LGP-30 General Purpose Computer from
the Royal McBee Corporation of Port Chester, New York at a price of forty thousand dollars. At the end
of the same month two or three members of the Committee attended a three-day course on computers in
Calgary which was sponsored by IBM.
In September the Committee's recommendations were approved by the Board of Governors, the order
for the computer was placed, and the computer was installed the following month. The original
installation consisted of the computer with a Flexowriter console (a modified electric typewriter with a
mechanical paper tape reader and punch), a photoelectric paper tape reader and a mechanical paper tape
punch, and an additional Flexowriter for the preparation of program and data tapes.
The University of Saskatchewan acquired an LGP-30 the same month as did the University of
Alberta. The first Canadian university to have a computer was, of course, the University of Toronto which
in 1948 started to design and build its own computer which was operated briefly and then replaced in
1952 or 1953 by the FERUT computer. The second Canadian university was the University of British
Columbia which acquired its computer in March 1957.
The LGP-30 was installed first in the basement of the Arts Building where the Department of Physics
7
was then located. Harry Schiff, who was one of the physicists using the FERUT, can still remember going
with Don Betts to see the LGP-30 when, according to him, Don ran ahead of him down the hall to get the
first glimpse. The LGP-30 was soon moved to the Arts Building Annex. Initially it was operated on an
open-shop basis under the general supervision of the Committee on Electronic Computing with Don Betts
and another member of the Department of Physics taking a very active role. Users of the computer had to
do their own programming since only a few rudimentary programs were available.
The LGP-30 soon attracted the attention of many faculty at the University who recognized its
potential usefulness in their work. A brief article describing the computer and its use appeared in the Fall
1957 issue of The New Trail9. The last three paragraphs may give some indication of the enthusiasm that
greeted the installation of the computer:
University officials feel that the computer is sufficiently simple in operation to be used
generally by staff and students. It takes about two or three days to train an operator.
An indication of the scope of research opened up by this latest University acquisition may be
gained from the following illustration: An average type problem undertaken by a theoretical
physicist used to take about one month to solve through the use of a desk calculator. Now, the
same worker need spend about one day preparing his material to meet machine operation
requirements. Actual computations will take about one hour. Many problems which formerly
were considered too time consuming to attempt can now be tackled without hesitation.
The new digital computer will be of vital use to personnel in chemistry, physics, engineering,
plant science, social sciences, and other fields. Problems involving a number of variables can be
accurately solved in short order.
It was soon apparent that technical assistance would have to be provided to users of the computer, and
several students were hired to provide temporary support during the summer. Eventually the
administration of the computer and its use was made the responsiblility of a newly formed Computing
Centre with a mandate to offer computing services to the entire University. On November 1, 1960 Don
Scott was appointed Director of the Computing Centre.
Of the many students who provided programming assistance for the LGP-30 we shall mention only
two. One was Ursula Bielenstein (later Ursula Maydell) who worked part-time as a student and then in
1960 became the first full-time employee of the Computing Centre. She resigned a year later to begin
graduate work in statistics in the Department of Mathematics. The other was Bill Adams who came to the
University of Alberta in the summer of 1959 to complete an undergraduate degree begun at the University
of Edinburgh and who subsequently obtained one of the first M.Sc. degrees in computing at the
University. Upon completion of their graduate work both Ursula and Bill joined the academic staff of the
Department of Computing Science which was formed in 1964 and remained with the Department until
8
their respective retirements almost thiry years later.
During the 1960/61 academic year the Computing Centre moved to the fourth floor of the Physical
Sciences Centre when it opened. A corner of the Computing Centre with the LGP-30 is shown in the
figure to the left which has been taken from the booklet1 prepared for the
official opening which makes the following reamarks about the Computing
Centre:
The University Computing Centre is independent of the teaching
departments, and provides a service to the whole university by
undertaking many kinds of numerical problems too complicated to
be handled by old-fashioned desk methods. It employs a full-time
and part-time staff of 18, and is equipped at present with a Royal-McBee LGP-30 electronic
computer which runs nearly 24 hours a day. This will shortly be replaced by a faster machine to
increase the problem-solving capacity of the Centre.
The LGP-30 and its manualsWe shall give only a brief description of the LGP-30 as it has been well documented in contemporary
and later papers which will be mentioned at the end of this section. The LGP-30 was 26 inches deep, 33
inches high and 44 inches long, and because of its size was sometimes referred to as a "desk computer". It
weighed 800 pounds, and was mounted on castors for easy moving. It contained 113 vacuum tubes and
1350 diodes. The tubes were mounted on etched circuit pluggable cards which contained associated
components. There was a total of 34 of these cards which were of 12 different types. The computer
consumed 1500 watts, and could be plugged into any standard 115 volt 60 cycle outlet. Air-conditioning
was not required if the room temperature could be kept within a reasonable range.
The main memory consisted of a magnetic drum with a capacity of 4096 thirty-two-bit words
arranged in sixty-four tracks of sixty-four sectors each with one word in a sector. There were three
circulating registers: the Accumulator Register for arithmetic operations, the Instruction Register for
holding the current instruction, and the Control Register which held the address of the next instruction to
be executed. Input and output was by means of a Flexowriter, an electric typewriter with a ten-character-
a-second paper tape reader and punch. The contents of these three registers could be viewed as binary
numbers on a small CRT display. The clock speed was 120 kilocycles giving addition and multiplication
times, inclusive of storage access, of 8750 and 24 000 microseconds, respectively. Recommended staff
for an eight-hour shift was one maintainer-operator and a part-time programmer. Approximately 500
LGP-30s were manufactured and sold.
The internal operation was binary so that all data had to be converted from decimal to binary on input
9
and from binary to decimal on output. Numbers were assumed to be represented in fixed-point form and
be less than unity in absolute value. Negative numbers were represented as two's complements. The order
code consisted of sixteen one-address instructions. Programming was in machine language although a
compiler became available in about 1959.
There were very few reference manuals for the LGP-30. Indeed one user of the computer remarked
recently that he can't recall ever having seen a manual. There was a glossy illustrated printed Operations
Manual describing the LGP-30 logical structure, the console switches and auxiliary equipment. It had two
Appendices, the first a table entitled "Computer Codes on Typewriter Keyboard" and the second giving a
discussion of scaling and a carefully formatted table "Powers of Two" with the values of 2n and 2-n for
n = 1, 2, ..., 31. There were also a Programming Manual, Programming Class Notes, and a Subroutine
Manual Coding Sheets manual giving programs for input and other utility routines, and a set of floating-
point arithmetic routines. Published somewhat later than these manuals was a manual for an "algebraic
compiler and translator system" with the acronym ACT.
The Programming Manual consisted of fifty-six pages of typescript in a coil binding and was written
in understandable but somewhat unpolished prose. The introductory paragraph, entitled "What is
programming", contained the following interesting statements: "Programming the Royal Precision LGP-
30 is basically simple. Understanding certain problems requires certain knowledge, however
programming for the LGP-30 does not." (Many persons undoubtedly disputed these statements as they
read further.) A description of the structure and programming of the LGP-30 was preceded by a
discussion of organizing calculations on a hypothetical desk calculator. The remainder of the manual was
approximately evenly divided into sections on the structure of the computer, programming, number
systems including the scaling of fixed-point binary numbers, input-output procedures, and a summary of
the order code.
Three excellent papers on the LGP-30 published in 1956 and 1957 bear the name of Stanley Frankel.
(Frankel began his professional career at Los Alamos after completing a Ph.D. in physics at the
University of California in 1942, and was one of the persons overseeing the hand computing group and
later the IBM punched-card installation10,11. After the war he held a number of positions and from 1949 to
1954 was head of a digital computing group at the California Institute of Technology which was
responsible for the logical design of the MINAC computer which was subsequently licensed to
Librascope, Inc. and built and marketed as the LGP-30.) Frankel’s first paper gives a short introduction
to stored-program computers together with some discussion of the logical design of the LGP-3012. The
second gives a "substantially correct" and very thorough logical description of both the MINAC and the
LGP-30, and refers to the MINAC as a "breadboard model of a computer" of which the LGP-30 is the
production version13. The third paper, written by Frankel and James Cass, begins with an historical
10
introduction to electronic computers and then describes the structure and programming of the LGP-30 and
ends with an efficient machine-language program for finding the sine of a given angle14. The abstract for
this last paper nicely reflects the newness of the subject of electronic computers:
Here it is - a clear and thorough introduction to the theory, design, philosophy, programming
and use of general-purpose digital computers - based on the new LGP-30. This is a much-needed
contribution to the literature of the computer field.
There are several other references to the LGP-30 giving specifications and statistics of use 15 and also
the logical design16, 17. There is also an excellent Web site with many links to many other relevant sites 18.
Finally there is the author's history of the Department of Computing at the University of Alberta which
has been consulted often in the preparation of the present paper19.
A course in programming the LGP-30We might mention a course which Bill Adams gave for the Faculty of Extension in January and
February of 1960 as it was the first computing course of any type given at the University of Alberta20. The
course consisted of twelve two-hour lectures given on consecutive Tuesday and Friday evenings. A brief
discussion of these lectures may show both how an introduction to programming has changed in forty
years and also how some of the principles have remained the same.
The course began with a brief description of the structure of the LGP-30 and its components.
Machine-language programming was then introduced and illustrated with a few simple problems such as
the computation of the area of an annular ring, moving a contiguous block of data from one part of
storage to another, polynomial evaluation, and Newton's method for finding square roots. There was
considerable discussion of number representation, conversion and scaling. Finally subroutines and
subroutine linkages and input-output operations were discussed.
Although the craft of programming has evolved and many programming languages have come and
gone in the forty years since this course was given, the approach to programming and learning to program
have in many ways remain unchanged. For example, on the second page of the notes we find the
following comment: "This course is intended as a practical course in programming and it is of great
importance that assignments should be done; otherwise nothing will be learned." An enumeration of the
steps required for the solution of a problem on a computer begins with a carefully written statement of the
problem, and continues through to test calculations, debugging, and, finally, "specifications" which is
underlined three times. Much in Bill's notes when read today shows the importance, and the difficulty, of
obtaining the correct balance between the details of a rapidly changing technology and the principles
underlying its effective application.
11
Use of the LGP-30 at the University of AlbertaUniversity departments using the LGP-30 included Animal Science, Chemical Engineering,
Chemistry, Educational Psychology, Electrical Engineering, Entomology, Geology, Mathematics,
Physiology, Plant Science, Political Economy, Psychology, and Soil Science. It was also used by the
Alberta Research Council whose laboratories were on the campus. We shall limit our remarks to use by
the Department of Mathematics and the Computing Centre, and the Faculty of Agriculture except for one
final example from the Faculty of Engineering given at the end of the paper.
Until the Department of Computing Science was formed in 1964, M. Sc. degrees in computing were
awarded through the Department of Mathematics. The LGP-30 was used for the computations in some of
the theses which included the numerical solution of ordinary and partial differential equations, the
numerical solution of functional equations, and numerical integration. One person can remember that a
colleague in Calgary collected data on “miles of punched paper tape” which were transmitted to
Edmonton by teletype for processing.
We have already mentioned Brian Hocking’s study of the variations in marks and how the first part of
the calculations was done by hand and the final part on the LGP-30
when it became available. Beginning in 1958 analysis-of-variance
calculations in the Department of Soil Science were done on the
LGP-30 which replaced the manual methods used during the
previous two years as has been mentioned earlier. Listings of the
program and some of the results are still available, and a portion of
one is shown in the accompanying figure. Another study in the
Department of Soil Science on the effect of cropping systems and fertilizers on yields acknowledged the
usefulness of the LGP-30 for the analysis of the experimental results21.
Another example of the use of the LGP-30 is provided by a paper given by Fenton MacHardy of the
Department of Agricultural Engineering at Congrès International Technique du Machinisme Agricole
held in Paris in 196122. One of the examples given in the paper involved finding the optimal allocation of
tractor resources for preparing and planting a given area of land. Computationally the problem involved
the solution of a linear programming problem with ten variables including slack and surplus variables and
four constraints and required about two minutes on the LGP-30.
An LGP-30 simulator In order to run simple machine-language programs for the LGP-30 a simulator was written which
implemented the 16 machine-language instructions but had simplified input and output instructions
although the internal representation of instructions and data was still in binary. The simulator was written
12
in J, which has been described as a modern dialect of APL. Readers unfamiliar with J may wish to visit
the J Software Web site at
www.jsoftware.com ,
or the author's J Page at
www.cs.ualberta.ca/~smillie/Jpage/Jpage.html
which gives some introductory material on J as
well as the script file for the simulator. We shall
discuss here only two simple examples.
The first example is taken from Bill Adams'
lecture notes and involves the calculation of the
area of the annular ring between two concentric
circles of radii R and r. This area may be
calculated simply as (R2 - r2), or alternatively
(R + r)(R - r).
The annotated program is shown in the box. Its
execution on the simulator together with some
notes are as follows:
Tape1=: 0.12 0.1 0.314159
Memory 1
0 Load Ex1
0 Run Tape1
>FLEXOWRITER
0.0013823
The program is given as the 17-item list Ex1, the first few items of which are
+-----+-----+-----+-----+-----+-----+|i0000|h0017|i0000|h0018|i0000|h0019|...+-----+-----+-----+-----+-----+-----+ .
A data tape is prepared with suitably scaled values for R = 1.2, r = 1 and . Since the program is
very short, only one 64-word track of drum memory is defined with addresses 0000, 0001, …, 0063.
The program is loaded in memory beginning at location 0, and is then executed with the given data
starting in location 0.Finally the computed area of the annular ring which has an unscaled value of
1.3823 is displayed .
The second example is from Frankel14 and gives a subroutine for finding the sine of an angle
expressed in "quadrants", i.e., units of 90 degrees. The subroutine which is not shown here had been
optimally coded and gave the sine to an accuracy of about eight decimal places. The time given in the
13
0000 i0000 Read R0001 h0017 Hold R0002 i0000 Read r0003 h0018 Hold r0004 i0000 Read pi0005 h0019 Hold pi0006 b0017 Bring R0007 a0018 Add r0008 h0020 Hold R+r0009 b0017 Bring R0010 s0018 Subtract r0011 m0020 Multiply by R+r0012 h0021 Hold (R+r)(R-r)0013 m0019 Multiply by pi0014 h0022 Hold pi(R+r)(R-r)0015 p0000 Print0016 z0000 Stop
paper for the calculation was about 150 milliseconds. The subroutine was coded for the simulator together
with a small program for its use. The time taken for the simulated calculation using a 2.0 GHz PC was
about 6 milliseconds.
Phasing out the LGP-30
An article appearing in The Edmonton Journal on November 22, 1960 said that the University's
computer, which was referred to as a "30 computer", was being used twenty-four hours a day and seven
days a week. Furthermore, it stated that "the University expects shortly to launch an extensive computer
training program for students, and that arrangements are being made to obtain a faster unit for the newly
established centre". An additional LGP-30 was rented for a three-month period in August 1961 to meet
the demand from the Alberta Research Council. In May 1961 the University acquired an IBM 1620
which was intended initially primarily for research purposes. The LGP-30 was retained for teaching and
remained in use until the middle of 1963. Incidentally, the recommendation for the purchase of the IBM
1620 was the last official act of the Committee on Electronic Equipment which then ceased to function.
A Ph. D. thesis23 submitted in 1963 illustrates how the LGP-30 with its difficult machine-language
programming was being replaced by the IBM 1620 which could be programmed in Fortran. Both the
LGP-30 and the IBM 1620 were used in the thesis. The LGP-30 programs required a listing of 14 pages
and four pages of operating instructions and were used to process data collected on punched paper tape
which could be input directly into the LGP-30. On the other hand the listings of the two small Fortran
programs required only a page each.
Cited in the References to this thesis is McCracken’s A Guide to Fortran Programming published in
1961. This was the first of his many informative, well-written, attractively produced and relatively
inexpensive programming texts. What a contrast this book makes with many of the programming manuals
available at that time. Finally one cannot help but think how much better off would be today’s students if
present texts came up to the same high technical and literary standards set by Daniel McCracken.
AcknowledgementsI would like to thank the following persons for their help during the preparation of this paper: George
Ball, John Beamish, Don Bellow, Steve Hunka, Rob Lake, Fran Mather, John McDonald, June Panteluk,
Jim Robertson and Ken Walsh. I would also like to thank Peggy Kidwell, National Museum of Natural
History, Smithsonian Institution for kindly providing a number of LGP-30 manuals which were very
useful during the writing of the simulator.
14
References1. Anon., New Physical Sciences Centre (Chemistry, Mathematics, Physics). University of Alberta,
Edmonton, Alberta, 1961.
2. Keeping, E. S., A Short History of the Mathematics Department. Department of Mathematics,
University of Alberta, Edmonton, Alberta, 1971.
3. Kenney, J. F. and E. S. Keeping, Mathematics of Statistics, Part One. Third Edition. D. Van Nostrand
Company, Inc., Princeton, N. J., 1954.
4. Ford, George. Sons of Martha. University Faculty of Engineering 1913-1988. Faculty of Engineering,
University of Engineering, Edmonton, Alberta, 1988.
5. Johnson, L. P. V., An Introduction to Applied Biometrics. Burgess Publishing Co. Minneapolis,
Minn., 1950.
6. Williams, Michael R., "UTEC and Ferut: The University of Toronto's Computation Centre," IEEE
Annals of the History of Computing, vol. 16, no. 2, 1994, pp. 4 - 12.
7. Anon., "Electronic brain aids University research," The New Trail, vol. xv, no. 1, 1957, p. 3.
8. Smillie, Keith, “Computation: From Abacus to Silicon Chip”, IEEE Annals of the History of
Computing, 2202, vol. 23, no. 4, 2002, Web Supplement.
9. Anon., "Electronic computer aids research," The New Trail, vol. xv, no. 2, 1957, pp. 2 -3.
10. Gleick, James, Genius. The Life and Science of Richard Feyman. Vintage Books, Inc. New York,
1993.
11. Metropolis, Nicholas and Eldred C. Nelson, "Early computing at Los Alamos," IEEE Annals of the
History of Computing, vol. 4, no. 4, 1982, pp. 348 - 357.
12. Frankel, Stanley, "Useful applications of a magnetic-drum computer," Electrical Engineering, vol.
75, July, 1956, pp. 654 - 659.
13. Frankel, Stanley P., "The logical design of a simple general purpose computer," IRE Transactions on
Electronic Computers, March, 1957, pp. 5 - 14.
14. Frankel, Stanley and James Cass, "LGP-30 General-Purpose Digital Computer," Instruments &
Automation, vol. 29, Feb., 1956, pp. 264 - 270.
15. Weik, M., A Second Survey of Domestic Electronic Digital Computing Systems, Office of Technical
Services, United States Department of Commerce, Washington, D. C., 1957.
16. Andrews, Ernest G., William R. Arsenault and Harry D. Huskey, "General-Purpose Computers,"
Computer Handbook by Harry D. Huskey and Granino A. Korn, McGraw-Hill Book Company, Inc.
New York, 1962, pp. 20-1 to 20-34.
17. Bell, C. Gordon and Allen Newell, Computer Structures. Readings and Examples, McGraw-Hill
Book Company Inc., New York, 1971.
15
18. Thelen, Ed., LGP-30. http://ed-thelen.org/comp-hist/lgp-30.html, 2003.
19. Smillie, Keith, The Department of Computing Science: The first twenty-five years. Technical Report
TR 91-01, Department of Computing Science, University of Alberta, Edmonton, Alberta, 1991. Also
available at http://www.cs.ualberta.ca/about/history.php.
20. Adams, William S., LGP-30 Lectures. Unpublished lecture notes, 1960.
21. Toogood, J. A. and D. L. Lynch, "Effect of cropping systems and fertilizers on mean weight-
diameter of aggregates of Breton plot soils," Canadian Journal of Plant Science, vol. 39, 1959, pp. 151 -
156.
22. MacHardy, F. V., "Application of mathematical techniques to the problems of agricultural
mechanization of areas undergoing development," Congrès International Technique du Machinisme
Agricole, Paris, 1961, F1 pp. 3 - 24.
23. Bellow, Donald G., Anticlastic Behaviour of Flat Plates. Ph. D. Thesis, Department of Mechanical
Engineering, University of Alberta, Edmonton, Alberta, 1963.
────────────────────────────────────────────────────────────Keith Smillie is Professor Emeritus of Computing Science, University of Alberta, Edmonton, Alberta
T6G 2E8. His email address is [email protected].
16