College and Research Libraries


RALPH R. SHAW 

Machine Application at 

The University of Hawaii 

REGRETABLY IT Is NECESSARY to set the 
record straight about the experiment in 
computer charging at the University of 
Hawaii. The article purporting to deal 
with this topic in the May 1965 issue of 
CRL does not present an accurate ac-
count. It treats of many topics and of 
possible routines that have no relation-
ship with the syste~ actually installed 
and tested at the University of Hawaii, 
and if left uncorrected the article will 
seriously mislead the profession. 

This reply is limited to the minimum 
areas that must be discussed if the er-
roneous impression given by the article 
is to be corrected. These are: ( 1) accu-
racy of the statements in the article; ( 2) 
operational efficiency of the system ac-

. tually installed and tested; ( 3) program 
efficiency of the system actually installed 
and tested; and ( 4) outcome of the test. 

1. With regard to the accuracy of 
the factual statements in the article, I 
shall limit· ,my discussion to the state-
ments in just one column, even though 
many other statements are equally ques- . 
tionable. 

Taking the right hand column of page 
217, the first sentence states, «The auto-
mation of the order department routines 
is already flow-charted and approved." 
Fact: If the operations have actually 
been flow-charted the head of the Sin-
clair library, the head of the order de-
partment and the head of the cataloging 
department have no knowledge of the 
alleged flow-charts , and it can be stated 

Dr. Shaw is Dean of Library Activities 
in the University of Hawaii. 

with certainty that no system for this 
purpose has been submitted for con-
sideration, let alone being approved, 

_ and its implementation date is not being 
delayed by personnel vacancies. 

The second sentence in the next para-
graph says, «It was designed to operate 
at approximately the same cost as the 
unsatisfactory manual system which pre-
ceded it ... " Fact: The cost was more 
than twice as high as that for the system 
it succeeded. See part 2. 

The next sentence says, «A $25,000 
switchover budget was approved and at 
this writing it appears that a large por-
tion of this amount will be returned to 
the state at the end of the fiscal year." 
Fact: There is no basis for this statement 
as indicated by the facts in part 2, 
which were available early in April. 

The next sentence says, «Machine 
costs turned out to be considerably lower 
than expected as a result of unpredict-
able advantages gained through cooper-
ative use of certain items." Fact: There 
was no cooperative use - of any of the 
items rented by the library, and during 
the pe~iod of the test they were used for 
nothing other than circulation. Computer 
costs at the Computer Center were dou-
ble the preliminary estimate. 

There are many other statements and 
allegations in this article that are defi-
nitely open to question, but it does not 
seem necessary to pursue this further. 

2. Operational efficiency. Using the 
costs for- October through March, and 
projecting the costs for the year at the 
March rate when the costs were slightly 
lower because the debugging had pre-

/ 381 



382 I College & Research Libraries • September 1965 

sumably been completed, the annual . 
cost of computer charging for salaries, 
machine rental and supplies in the li-
brary was $86,498.52 plus cost of com-
puter time of $33,000.00, or a total of 
$119,498.52 per year. The cost of the 
system that the computer charging re-
placed was $53,027.73 per year. These 
figures are based on actual payroll rec-
ords for both periods, supplies and 
rentals shown on the university books, 
and time billed by the Computer Center. 
There were no activities added at the 
circulation desk during the period 1964/ 
65, so the net cost of computer charging 
was found to be substantially more than 
double that of the old manual charging 
system, which was no model of effi-
ciency. 

3. Program efficiency. The chief ob-
jection to the old charging system, as 
stated in the Black report, which the 
May article alleges to be the basis for 
the change, was, .. Indeed the present 
circulation system is under attack solely 
because of the difficulty of ascertaining 
where books may be if they are not on 
the shelves." This objection was based 
on the fact that it was, in the old system, 
necessary to consult multiple £les and 
that it sometimes took from fifteen to 
thirty minutes to locate a book. Under 
the computer charging system as in-
stalled at Hawaii, it was always a day 
before the charge got into the £les since 
the computer was run once a day only, 
and book charges from Friday afternoon 
at 4 PM through Monday at 4 PM did 
not become available until Tuesday 
morning. In fact the log kept of date of 
charge and date when the computer 
charge record became available to the 
circulation desk shows that the mean 
time for getting a charge into the com-
puter £le was 3.3 days, and that on sev-
eral occasions the delay ran two weeks 
or more before the location of a book 
could be determined. 

Overdues, which sound real tidy in 
the May 1965 article, were in fact re-

ported to the library only twice between 
the time computer charging actually 
started on October 8, 1964, and March 
31, which made a farce of the time con-
trol system. Overdues could not be sent 
even after the overdue printouts were 
received because they were so far out 
of date. 

While renewals, as stated in the ar-
ticle, can be handled by running the 
book card back through the machine, 
the article fails to point out that this 
requires that the book be brought back 
to the library to make the machine-read-
able book card available. An alternative 
would be to repunch the book card. 
This would be difficult with information 
obtained over the telephone and would 
require the punching of three cards 
which would then be sent to the Com-
puter Center and would be converted 
into one book card and returned a week 
later after which the renewal could be 
run into the system. The net result is , 
that renewals could only be handled if 
the borrowers brought the books back 
instead of their being handled by tele-
phone. 

Another major factor in terms of pro-
gram efficiency is the fact that one-day 
loans or reserve-book loans could not 
be handled in the computer charging 
system as designed and operated, which 
meant that more than half the charging 
continued to be done manually as be-
fore, resulting in multiple charging sys-
tems in use at the same time. 

4. Outcome of the test. After the data 
noted above and other results of the test 
were studied by everyone concerned, the 
rental of the equipment was cancelled 
as of May 6, 1965, and a simple charging 
system using an IBM 026 punch for 
about four hours per week and a tabu-
lator at the Computer Center for about 
two hours per week was put into opera-
tion on June 1, 1965. This system uses 
a uniform method for all materials 
charged and provides time control and 

( Continued on page 398) 



398 I College & Research Libraries • September 1965 

a key to the library on assumption of 
responsibility for the collection during 
that period of time. Cooperation be-
tween the individual scientists, the de-
partments, and the library in collecting 
material is essential. As individual col-
lections increase to an unmanageable 
size, they should cumulate into depart-
mental reading rooms, and when these 
reach a critical size the collections should 
fade into control and supervision by the 
library. The location of science collec-
tions, departmental or library owned, 
should be as central as possible to all 
departmental laboratories and connected 
by passageways if necessary. An ideal 
solution to this problem would seem to 
be the following description of a sciences 
building complex: 

. . . a great wheel, or pentagon, in which 
the science library would be the hub, the 
radiating spokes the major disciplines-
physics, mathematics, chemistry, the biolog-
ical sciences-and the rims the laborato-
ries."8 

Finally, if the library's users include 
students, a strong argument can be made 
for centralizing science collections for 
their convenience. They will be taking 
courses in a variety of subjects, and in an 
era when ever-increasing quantities of 
materials must be covered in each sub-
ject any time saved for students would 
be an educational advantage. 

In conclusion it should be stressed 
that a university or college library's main 
function is to provide the user with those 
books that have a frequency of use that 
is so low, or a cost that is so high, as 
to prohibit individual purchase or own-
ership and to increase the educational 
value of each item by adding to the com-
pleteness of the collection. Within those 
sciences that need to use books and jour-
nals in a laboratory situation it is neces-
sary that individual scientists and de-
partments assume ownership and respon-
sibility for some types of library mate-
rials. • • 

s Shera, op. cit., p. 42. 

University of Hawaii 
(Continued from page 382) 

location of materials charged within five 
minutes at most. Three professional posi-
tions have been transferred from the 
circulation department to other places 
where they are needed. Two of these 
positions were filled and the personnel 
are now working in other departments, 
the third is in process of being filled in 
another department. 

Caveat: This is not a generalization 
about computer charging. It is a factual 
report on the system as it was actually 
applied at the University of Hawaii. It 
may very well be that some system still 
to come and some machine configuration 
still to come may give us better control 
of loans for less money, but that remains 
for the future. 

This is not a condemnation of the 
use of computers in libraries when suit-
able. We have, in fact, just completed 
the systems work and programing, in-
cluding two dry runs, and have convert-
ed to computer handling of our account-
ing records for book purchases, which 
have to be distributed against some fifty 
accounts. The time studies on this appli-
cation indicate that the use of a key 
punch for two hours per day plus the 
use of a 1401 computer for not more 
than ten minutes per month will save 
us a net of somewhat more than £our-
man-years while giving us better and 
more prompt control of our book ex-
penditures. As fast as we can develop, 
or anyone can show us, other applica-
tions that will in fact release time for 
library services or money for books we 
will put them into operation. This, how-
ever, requires rigorous systems work 
and careful costing so that we can be 
reasonably certain that the change will 
result in an improvement. The kind of 
reliable and helpful information that is 
needed is not provided by articles such 
as the one under discussion. • •