College and Research Libraries


By FREMONT RIDER 

Microcards and Punched Card Filing 

Mr. Rider ts librarian, Wesleyan Uni-
versity. 

A BOUT ONCE a month for the last three 
.£"1_ years the writer has received a letter 
from someone suggesting what a splendid 
idea it would be to file microcards by one 
or another of the punched card sorting 
techniques. I have had to reply that the 
filing of microcards, or for that matter the 
filing of any sort of catalog cards, was not 
practicable by any of the punched card 
systems with which I was familiar, and in 
the case of most of these letters I tried to 
explain why. Since the volume of this 
punched-card-fi1ing correspondence is now 
rather bulky, and since the subject is one 
that, for obvious reasons, is of considerable 
practical interest to librarians, it has seemed 
worth while to put in the record this brief 
discussion of it. 

It must be repeated at the start that this 
discussion reflects only such mechanical · 
sorting methodologies as are now in use, 
for around the corner there may be some-
thing so new and so different as to make 
mechanical catalog card filing practicable. 
By practicable we, of course, do not mean 
possible, for mechanical card filing is al-
ready possible enough. By practicable we 
mean more efficient. 

Proponents of the mechanical filing of 
catalog cards have failed to realize at its 
full value the fact that all punched card 
systems were developed to be, and still are, 
sorting systems, not filing systems. There 
is a very real difference between these two 
words, and in thi difference lies the im-
practicality of punched card techniques for 

JULY~ 1949~ PART I 

filing. If we are faced with a file of ten 
million cards, divided up into a thousand 
or more categories, or combinations of cate-
gories, some of them consisting of a few 
hundred cards only, some of them contain-
ing perhaps hundreds of thousands of cards, 
and we want to segregate from that file 
any one of these categories it is obvious 
that we have a sorting job to do, not a 
filing job. By punching many holes on 
these ten million cards, or by intricately 
notching one or more of the edges, in either 
case according to a predetermined code, 
we can sort the cards mechanically. 

The filing of catalog cards presents an ~ 
entirely different problem. From a library 
catalog we never have occasion to sort out, 
in answer to any patron's request, hundreds 
of thousands of cards. For one to want 
even a hundred cards at any one time under 
any one category (i.e., under any one au-
thor or subject heading) is an event of 
great rarity. Most of the time what one 
seeks in the catalog is one card--or two or 
three cards. By using present catalog meth-
ods, we may, · on the average, take a minute. 
to find that one card that we seek-granted, 
of course, both a well-made catalog and an 
intelligent searcher! On the other hand 
to search that one card out of the catalog 
mechanically, although it is something that 
is technically possible~ is possible only at 
a prodigious waste of time and money. 
Furthermore this waste is one that mounts 
up in every phase of the mechanical sorting 
process. Let us analyze it. 

First, merely as a preliminary to mechani-
cal searching thereafter, every card that we 
file in our catalog must be individually 

233 



punched or notched to meet the sorting re-
quirements of our predetermined code. I 
have no cost figures on this particular phase 
of punched card technique, but surely this 
item is in no sense an inconsiderable one. 
It is strongly to be suspected that it would 
be considerably more costly than the typing 
of the catalog card was. 

Second, this punching or notching takes 
up an appreciable proportion of the total 
·space on the card Notching systems de-
mand for themselves the exclusive use of 
almost one-half of the total over-all area of 
the card. Hole-punched systems, can in 
some cases, be sprinkled over the face of 
the card without interfering too much with 
its normal use: but with ordinary catalog 
cards such an over-all sprinkling would 
probably be impracticable, and with micro-
cards it would certainly be. In other 
words, with every catalog card which is 
intended to be mechanically filed, a definite 
-and a very substantial-proportion of the 
total area of the card has to be allocated 
to notches or punch holes. If this propor-
tion be half of the area then it means that, 
in the case of microcards, we would have 
to double the number of cards in our file 
(which means in turn doubling not only 
the cost of the microcards themselves but 
also the cost of the filing equipment and 
file space rental used by them) simply to 
effect-we hope-some saving in an im-
portant, but still only subsidiary, cataloging 
cost, i.e. card filing and searching. 

We now come to a third question. Is 
it possible for us to cod.ify, according to 
any practicable notching or punched hole 
system, the interminable ramifications and 
variations of heading which are to be found 
in any large dictionary catalog? Person- · 
ally I doubt it. 

It is one thing to have a punched hole 
pattern that will fish out mechanically, from 
a file of personal name cards, all blue eyed 
males 6 5 years old using false teeth-or any 

other similar combinations of sortings that 
one may desire to make. But it is quite 
another thing to have a punched hole sort-
ing pattern that will mechanically fish out 
for us, out ef ten million cards, two en-
tries under the authorship of "Frederick 
Alexander Stapleton, 1822-83." 

It is true that permutations do quickly 
run into ·astronomical totals. It may be 
true that a ten million card catalog does 
not have - more than a million different 
headings. Unfortunately for punched card 
coding systems, those million headings have 
a complexity of diversity that stolidly defies 
any sort of purely numerical transliteration. 
Most punched card systems deal primarily 
with surnames. But a full-fledged dic-
tionary catalog requires the mechanical 
separation not only of surnames, but of 
full name entries (equipped with birth and 
death date differentia) ; and the separation 
of these from subject headings, some of the 
latter as many as I 2 words long and them-
selves coupled with numerical and symbol-
istic differentia, from title entries each sev-
eral words long, from series entries and 
other entries. Remember that all of these 
different forms of entry are now mixed in 
a single concatenated but minutely differ-
entiated sequence. I would hate to be given 
the task of plotting a pattern of holes to 
unravel it. 

We have not yet arrived at the fourth 
phase of our problem, the heart of it, the 
real nub of punched-card catalog card im-
practicability. All mechanical sorting sys-
tems demand that all the cards involved 
in a given sorting process be removed from 
thei~ drawers and run through some kind 
of a sorting machine. This machine may 
be the simplest sort of gadget imaginable 
(in one system a long steel needle suffices 
to do the trick). Or it may be a very 
compli<.:ated electrical contraption, equipped 
with all sorts of counting, tabulating, throw 
out, and cross-checking devices. All these 

234 COLLEGE AND RESEARCH LIBRARIES 



devices have, however, as their common 
denominator the above stated demand for 
removal. And it is this demand for re-
moval that finally throws out, as imprac-
ticable, all mechanical methods .of catalog 
card searching. 

Let's take a specific example. . You as 
a reference librarian are asked to ascertain 
if your library possesses a copy of Penn-
sylvania Genera of Permian Brachiopoda 

by one William Selzer. Which process for 
the answering of this question is the easier 
one? To do what you do now, i.e. to go 
to your ten million card catalog and to look 
in it, under the author entry, "Selzer, Wil-
liam," for a title beginning "Pennsylvania 
Genera?" Or to resort to a mechanical 
searching process, which means to take all 
of your catalog's ten million cards out of 
their ten thousand catalog drawers and to 
run them through a mechanical sorting ma-
chine (then putting them back into their 
drawers again) until finally the machine 
come& to that one "William Selzer" card 
that you are looking for, and is automati-
cally thrown out for your inspection? The 
answer is obvious. 

Not so obvious at first glance, however, 
are a host of additional collateral difficulties. 

• 

We must remember for one thing that a 
large library catalog is usually servicing 
not one card searcher but a dozen--or a 
hundred-searchers simultaneously. If 
each one of these hundred searchers is to 
be mechanically served, it is clear that not 
one but a hundred sorting machines are 
going to be required, working simultane-
ously to find for each patron the single 
card--or the two or three cards--that each 
one wants. But for each patron (with ten 
million cards to be gone through) the 
automatic searching process is clearly going 
to be a matter of hours, taking the handling 
time of dozens of library attendants, and 
creating an enormous confusion of cards 
and files simultaneously going through the 
machine searching routine. 

Finally, fifth, what of the physical wear 
and tear on your cards if they are going to 
be run constantly through these sorting 
machines? Obviously this runmng 
through is going to involve a certain amount 
of card erosion. 

All · of which would seem to sum up to 
this: that mechanical card filing applied to 
library card catalogs sounds appealing, but 
as yet cannot be deemed a practicable 
proposition . 

National Program in Cataloging 
(Continued from page 232) 

for each type of librarianship-school, pub-
lic and college or university. 10 Insofar 
as cataloging is concerned, are there not 
lessons to be learned from the past? Since 
Jewett's time have we not got rather far 
away from the institutional point of view? 
In the brief account here presented of cer-
tain phases of cataloging history, we have 
learned that the institutional point of view 

10 "Southeastern Conference on Library Education, 
February 29-March 6, 1948." Atlanta, Ga., South-
eastern Library Association, 1948, p. 4. 

JULY~ 1949~ PART I 

prevailed in the pioneer period of catalog-
ing history, that it steadily lost ground 
after 1876, and that since 1908 it has been 
overwhelmed by a national pattern in card 
production--a pattern which has been 
actually detrimental to the welfare of some 
institutions. In the present renaissance in 
cataloging history, we hope that the cata-
loging interests of all types of libraries will 
find an adequate niche in a sound national 
program. 

235