WORLD CONFERENCE ON RECORDS AND GENEALOGICAL SEMINAR COMPUTER INPUT AND OUTPUT TOOLS AND TECHNIQUES

• Document Type: CREST
• Collection: General CIA Records
• Document Number (FOIA) /ESDN (CREST): CIA-RDP73-00402R000100140011-7
• Release Decision: RIFPUB
• Original Classification: K
• Document Page Count: 21
• Document Creation Date: December 19, 2016
• Document Release Date: July 19, 2006
• Sequence Number: 11
• Publication Date: August 5, 1969
• Content Type: REPORT

Approved For Release 2006109/25 : CIA-RDP73-00402R000100140011-7?
. WORLD CONFERENCE
ON RECORDS
AND GENEALOGICAL SEMINAR
"Record Protection in
an Uncertain World"
Salt Lake City, Utah, U.S.A.
5-8 August 1969
Part. I I
Automatic Document Reading; Optical Scanning, an answer
to the Computer Input Problem?
by Melvin M. Farr, B.A., J.D., LL.M.
Approved For Release 2006/09/25 : CIA-R 7 40011-7,q - 6 & 7 b
COPYRIGHT? 196-9 THE GENEALOGICAL SOCIETY OF THE CHURCH OF JESUS CHRIST OF LATTER-IIAYS}C.INTS, TNC,:,
 Approe8$V 1rTEVe1NPT?AND905TPUT TBBLJ A?NB ?GNIQ4 gg011-7
Part I I
Automatic Document Reading; Optical Scanning, an answer
to the Computer Input Problem?
by Melvin M. Farr, B.A., J.D., LL.M.
INTRODUCTION
Much has been said and written about the population
explosion, but an even larger explosion is the "Information
Explosion". It is reported that every five years the amount
of printed information doubles. More and more of this
information is finding its way into the computer, along with
myriad of other information; such as information about bank
accounts, financial records, traffic violations, etc.
Keeping pace with this information explosion, computer
manufacturers continue to announce faster and faster
computers. They have developed disc drives that read in one
hour, information that would take one keypunch operator 85
years to prepare. This being the case it is easy to see why
"Input" is considered to be one of the key bottlenecks in
computer systems.
It was the recognition of this bottleneck that caused
the development of various optical scanners. Each striving
for the ideal in systems design, "the capturing of data at
its source in a machine readable form.". Some have been
more successful than others in this attempt and it is this
fact that makes it difficult for potential users to know
Approved For Release 2006/09/25 : CIA-RDP73-00402R000100140011-7
 Approved For Release 2006/09/25 : CIA-RDP73-00402R000100140011-7
what to do in regards to optical scanning. If there were
but one scanning company, with only one machine, reading
only one type of recorded media it would be relatively easy
for one to determine whether or not he had a scanning
application; and once an application was recognized there
would be no company selection problem. This, of course, is
not the case. There are several companies, some having more
than one machine, having different "capabilities", and
reading different types of recorded data, making it
difficult to know what to do in regards to optical scanning.
The purpose of this paper is to present general
information on scanning with the hope that it might aid
those who have potential scanner applications to understand
some of the factors they should consider in determining if
optical scanning can aid them in reducing the "Input"
bottleneck.
I. FINDING A SCANNER APPLICATION
The potential for scanner use, obviously, exists anytime
there is a system that does tie, or should tie into a
computer. The ideal as mentioned before, is where the data
can be captured at its source in machine readable form.
e.g., the first time the information is recorded on paper
have it in a form, mark sense, handprinting, typewritten,
etc; so that it can be read by machine and put into computer
-2-
Approved For Release 2006/09/25 : CIA-RDP73-00402R000100140011-7
 Approved For Release 2006/09/25 : CIA-RDP73-00402R000100140011-7
readable form. In determining whether there is a scanning
application several factors must be considered. Some of
them are: The type of scanning equipment, the system
itself, the capability of the people working the system, the
cost of accomplishing this by scanning vs. other methods.
A. TYPES OF SCANNING EQUIPMENT
1. Bar Code Readers. This type of equipment, reads
"Bar Codes" printed mechanically. This is usually
accomplished through manually operated imprinters or "Data
Recorders."
A well known example of this method, is some oil company
credit cards. Account identification information is
embossed on a card in bar code form (a series of vertical
long and short dashes,) that represent an account number.
This card is put in an imprinter at the time a sale is made
and variable information, i.e., the amount being charged by
the customer is entered into the imprinter key board by the
station attendant. The charge slip is put in the imprinter
and when operated the imprinter transfers the account
number, amount of purchase, plus other information to the
charge slips. The record of purchase is put through an
optical scanner and the information is converted to cards,
paper tape, or magnetic tape and from these entered into the
computer for processing.
Approved For Release 2006/09/25 : CIA-RDP73-00402R000100140011-7
 Approved For Release 2006/09/25 : CIA-RDP73-00402R000100140011-7
This is a good use of Bar Code Readers, in that it
permits entry of the data into machine readable form at the
source where the transaction takes place and the first time
it is recorded. In addition, the cost of the imprinter is
not great (less that $100.00) and it is very simple and
fast, yet reliable.
2. Mark Readers This equipment reads pencil marks
e.g./- ,
entered by a person on a card or paper. These marks are
entered on the forms in predetermined fields (areas on the
forms). The presence of a mark in a given area serves to
represent a given quantity or meaning, dependent upon the
way the system is designed. A very common use for mark
readers is in connection with meter reading of gas or power
company meters. The "Meter Reader" marks the meter value on
a customer account card, e.g. The customer card would
probably have four verticle columns of numbers, each column
having the digits 0 to 9 down the card. To represent the
numbers 2468 the meter reader would put a mark through the 2
in column one (the column farthest to the left). Through
the 4 in the second column, through the 6 in column three,
and through the 8 in the fourth column (farthest to the
right). The card is then passed through an optical mark
reader and the information is punched into the card,
converted to paper tape, or magnetic tape, so that it can
Approved For Release 2006/09/25 : CIA-RDP73-00402R000100140011-7
 Approved For Release 2006/09/25 : CIA-RDP73-00402R000100140011-7
then be processed by computer. (Some mark readers read
optically, while others read it by "sensing" the marks on
the paper or card. For the latter, special electrographic
pencils must be used.)
A big advantage of this approach is that the "Imprinter"
(pencil) is very inexpensive and mobile, permitting the
person, responsible for making the entry, to go to many
places to gather information, and there make an entry that,
at its source, is machine readable. A problem with this
method is that in dealing with marks one is dealing with
something with which people are not accustomed and it is
therefore, more susceptible to people errors.
3. Character Recognition Equipment These machines read hand
printed or machine printed characters. The object of this
equipment is to permit a person to record data in the form
in which humans are accumstomed to seeing and using it, and
then have the scanning equipment "translate" it to computer
language for processing.
HAND PRINTING
Hand printing technology is in an infant stage, and
requires that the person printing use great care in the
preparation of the data. Reading of handprinted alpha
characters is not yet very good. Since this is the case,
applications using hand printing are rather limited; though
Approved For Release 2006/09/25 : CIA-RDP73-00402R000100140011-7
 Approved For Release 2006/09/25 : CIA-RDP73-00402R000100140011-7
there are some where department store clerks hand print
customer account numbers and purchase amounts on a sales
check, which is then optically scanned and converted to
computer language for processing. Hand printed character
recognition holds out promise since there are so many
businesses where the source of entry is made by a person
using a pen or pencil. As with mark readers, the input
device (pencil) is very inexpensive and mobile. Unlike the
mark readers, however, by reading hand printed characters
information is read in the form in which people are
accustomed to seeing and using it. For this reason it is
easier for people to work with, and since it is in "English"
a copy of the hand printed entry form can be given to a
customer as a receipt.
MACHINE PRINTING
The reading of machine printing is quite advanced and
several companies have announced equipment that is capable
of reading machine printed characters. Most of the machines
are single font machines. These are machines that are
capable of reading only one font. A font is a set of
characters of the same type, size, and style, such as
Underwood Distinctive Elite, IBM Prestige Elite, etc. There
are, however, some "Multifont" scanners that are capable of
reading two or more fonts; an example of one of these is
Approved For Release 2006/09/25 : CIA-RDP73-00402R000100140011-7
 Approved For Release 2006/09/25 : CIA-RDP73-00402R000100140011-7
the machine being installed at Management Systems in Salt
Lake City, Utah. It is capable of reading 19 different
typewriters (Those typewriters that have basic pica.or elite
fonts), 1403 type style, and USASI-1 type style. It reads
pica and elite upper and lower case characters along with
certain special characters.
Some of the applications using machines to read machine
printed information are:
(1.) To read text material prepared by typewriters for
newspapers.
(2.) To read customer bills being returned with a
payment, the bill having been printed on a 1403,
mailed to the customer, returned, and the amount of
payment encoded on the bill with a bank proof
machine.
(3.) To read airline passenger and audit tickets.
(4.) For the conversion of large amounts of information
to computer readable form. This is accomplished by
the use of typewriter and scanners rather than
keypunching.
Advantages that exist for the use of scanners for the
conversion of large amounts of information over keypunching
are as follows:
Approved For Release 2006/09125r: CIA-RDP73-00402R000100140011-7
 Approved For Release 2006/09/25 : CIA-RDP73-00402R000100140011-7
(1.) The input device (typewriter) is much cheaper than
a keypunch machine, and more mobile.
(2.) The amount one has to pay for a good typist is less
than that paid a good keypunch operator, due to the
fact that typing is easier, and a typewriter takes
less skill to operate efficiently than a keypunch.
(3.) It is easier to correct typing errors than
keypunching errors.
(4.) Typing is faster than keypunching.
(5.) Record length is practically unlimited (not limited
to 80 characters as in keypunch).
4. MICR READERS (Magnetic Ink Character Recognition) These
are not "Optical Scanning" machines, since they do not read
optically; but through the recognition of "magnetic"
patterns as the paper is passed under the magnetic head with
the head touching the paper. These are used mostly in the
Banking Industry and since they are not optical scanners
will not be discussed further.
B. The System Itself In considering whether there is a
scanner application, the first question, that must be
answered is "should the information being considered be
processed by computer?" If it should be, then the next step
is to "determine where the information orginates" Once the
origin is found, the question is, "can the system be
Approved For Release 2006/09/25 : CIA-RDP73-00402R000100140011-7
 Approved For Release 2006/09/25 : CIA-RDP73-00402R000100140011-7
designed so that the original entry is in a scanable form?"
The-greatest possible benefit is derived from scanning if
the source record, that has to be prepared as a basic
element of the system, can be prepared in scanable form. An
excellent example of this is the above mentioned gasoline
credit card application. As a part of doing business the
gasoline attendant must:
1. Record the purchase as to
a. The amount of purchase
b. The customer making the purchase
2. Give the customer a receipt of the purchase for his
record.
All of these are accomplished, at the source, through
the use of an embossed card and an imprinter. In addition,
the method used is about as simple as could exist,
considering what has to be done.
If the source data cannot be entered in machine readable
form, but must be "copied" (keypunch, etc.) for computer
use, scanning could still be the best method because of the
above mentioned advantages of typing over keypunch.
C. CAPABILITY OF THE PEOPLE WORKING THE SYSTEM
A system should not be built around personalities or
individuals, but one must consider the capabilities of the
type of person that is to perform the job, and the
Approved For Release 2006/09/25 : CIA-RDP73-00402R000100140011-7
 Approved For Release 2006/09/25 : CIA-RDP73-00402R000100140011-7
environment in which the job will be performed. For
example, in the gasoline credit card application, factors to
consider and test are:
1. Can the station attendant, who probably has only a
high school education or less, accurately operate
the imprinter.
2. Will his hands be clean enough so that he will not
get grease on the charge slip, that might interfere
with the scanner's reading of the slip?
If this application required that the attendant type the
information, it would probably not work, since relatively
few attendants know how to type accurately and it would also
be too slow.
Tests should be conducted to determine whether the
typical employee working the system has the capability to
prepare the entry in machine readable form.
D. COST OF SCANNING VS. OTHER METHODS
Once it has been determined that a potential scanning
application exists, the question of cost comes up. Although
it has been said that typing has advantages over keypunch;
if.the volume of the input data is sufficient, say, for only
one keypunch machine, it would be difficult, if not
impossible to cost justify a scanner. Therefore, one should
conduct a study to determine the number of transactions
-10-
Approved For Release 2006/09/25 : CIA-RDP73-00402R000100140011-7
 Approved For Release 2006/09/25 : CIA-RDP73-00402R000100140011-7
needing entry to the computer, along with the average number
of characters per transaction. Then using this information,
determine the comparable cost for various methods of input.
Some factors to be considered are:
1. Can the source entry be prepared in a scanable
form, as a by product of the normal system, thereby
eliminating the need to "copy" it (keypunch, etc.)
2. What will the error and reject rate be on the
application if a scanner is used? In scanning an
error (substitution) is where the scanner mis-reads
a character, e.g.. a B as an 8 an E as an F, etc.
A reject is where the scanner cannot read a
character and therefore outputs a reject symbol,
such as a question mark, in place of the character.
3. Under the system as planned, what is the method
(and its cost) of finding and correcting errors?
Some possible methods are as follows:
a. As the material is read on the scanner a
print-out of what was read by the scanner
is prepared on a high speed printer.
This print-out is visually compared
against the transactions and errors and
rejects are noted. Corrections are then
typed and re-entered into the system.
-11-
Approved For Release 2006/09/25 : CIA-RDP73-00402R000100140011-7
 Approved For Release 2006/09/25 : CIA-RDP73-00402R000100140011-7
b. The scanner output tape is edited on the
computer for obvious errors.
(1.) Incorrect account numbers (found through
use of check digits.)
(2.) Incorrect amounts. (found through batch
balancing procedures)
(3.) Numeric information in an alpha field,
etc.
c. A combination of the above two methods
4. What method will be used for entering reject
characters? Some possible methods are:
a. Entry in connection with error correction as
discussed above.
Through the use of a CRT terminal, the
transactions are displayed and the reject
characters are entered through the terminal.
NOTE: Consideration should be given to the
effect of an error or reject on the system if
it is not corrected. It might be found that
errors and rejects in some fields will not be
a problem. If this is the case those fields
would not be checked for errors or rejects.
-12-
Approved For Release 2006/09/25 : CIA-RDP73-00402R000100140011-7
 Approved For Release 2006/09/25 : CIA-RDP73-00402R000100140011-7
5. The number of keypunches and verifiers- (operators
6.
and equipment) that will be displaced if scanning
is used.
The cost of the scanner rental. In figuring this,
recognize that in an eight-hour day the scanner
will not be running eight hours due to forms
handling, paper jams, etc.
7. Scanner operator costs, including overhead; such as
sick leave, vacations, holiday, and supervision.
8. Space cost for the equipment, keypunch, scanner,
etc., along with facility costs, such as air
conditioning, power, etc.
II SELECTING THE OPTICAL SCANNER
Once a scanning application has been found, and it is
known what type of scanner is needed, e.g. bar code reader,
mark reader, etc. the next step to be taken is to select
the specific scanner to be used.
Some of the steps to be taken in this phase are the fol-
lowing:
A. Select from the many companies, three or four that have
a scanner to fit the system. There are many scanning
companies manufacturing many different types of scanners, at
different prices. Some of the companies are:
Approved For Release 2006/09/25 : CIA-RDP73-00402R000100140011-7
 Approved For Release 2006/09/25 : CIA-RDP73-00402R000100140011-7
ADDRESSOGRAPH MULTIGRAPH CORP.
1200 Babbitt Road
Cleveland, Ohio 44117
COGNETRONICS CORP.
333 North Bedford Road
Mount Kisco, New York 10549
CONTROL DATA CORPORATION
8100 34th Ave. 5
Minneapolis, Minn. 55440
CUMMING-CHICAGO CORP.
4740 N. Ravenswood Ave.
Chicago, Illinois 60640
FARRINGTON MANUFACTURING CO.
5881 Leesburg Pike
Falls Church, Va. 22041
GENERAL ELECTRIC CO.
3500 North Central Ave.
Phoenix, Ariz. 85012
HONEYWELL EDP
60 Walnut Street
Wellesley Hills, Mass. 02181
IBM CORP.
112 East Post Road
White Plains, N. Y. 10601
Approved For Release 2006/09/253 A-RDP73-00402R000100140011-7
 Approved For Release 2006/09/25 : CIA-RDP73-00402R000100140011-7
INTERNATIONAL COMPUTERS LTD
839 Stewart Ave.
Garden City, Long Island, N. Y. 11533
MEASUREMENT RESEARCH CENTER
P.O.B. 30
Iowa City, Iowa 52240
THE NATIONAL CASH REGISTER CO.
Dayton, Ohio 45409
NATIONAL COMPUTER SYSTEMS
1015 South 6 Street
Minneopolis, Minn. 55415
OPTICAL SCANNING CORP.
Newton, Pa. 18940
PHILCO-FORD CORPORATION
3900 Welsh Road
Willow Grove, Pa. 19090
R.C.A.
Cherry Hill, New Jersey
RECOGNITION EQUIPMENT INC.
1500 West Mockingbird Lane
Dallas, Texas
REPUBLIC ADVANCED TECH. SYSTEM GROUP
9754 Deering Ave.
Chatsworth, Cal. 91311
Approved For Release 2006/09/25 : CIA-RDP73-00402R000100140011-7
 Approved For Release 2006/09/25 : CIA-RDP73-00402R000100140011-7
SCAN-DATA CORP.
800 East Main Street
Norristown, Pa. 19401
UNIVAC DIVISION OF SPERRY RAND
Roseville, Pennsylvania
A letter could be written to the above companies, asking
for information about the particular scanner, bar code
reader, mark reader, etc. needed for the application.
Information about the application should be passed on to the
companies with this letter. In response to this letter it
should be possible to narrow down the number of companies
that have scanners that interest you. Factors to consider
are:
1. If your application is a bar code application, does
the company have a bar code reader?
2. What is the speed of the scanner? It must be fast
enough to handle your daily transactions but not so
fast that there is too much excess capacity.
3. What is the cost of the scanner? Does it fit in
with the cost estimates you have made.
B. Run Tests With the Scanning Companies.
Once the selection has been narrowed down to a few
companies, tests should be run with, these companies to
satisfy yourself that their "promises" in fact, can be met.
-16-
Approved For Release 2006/09/25 : CIA-RDP73-00402R000100140011-7
 Approved For Release 2006/09/25 : CIA-RDP73-00402R000100140011-7
Before running a test it is important to find out from the
companies:
1. The Brand of paper to be used (should it be bond or
sulphite?). Most paper companies put out a
scanning paper and the scanning manufacturer should
know which is the best for his particular scanner.
2. The dolor of inks that are "blind" to the scanner.
The form should be printed in this color.
3. The minimum and maximum form size.
4. The Location of the "read" areas on the forms.
(Margin requirements, etc.)
5. If a machine print reader, what type of ribbon
should be used (carbon or fabric ribbon?) What is
the best brand?
C. Visit users of the particular scanner you are interested
in. Upon narrowing your choice down to a particular
scanner, to help you finalize your selection; it is well to
make arrangements to talk with people who are using the
scanner you have selected. This can be done through
correspondence, but is better accomplished via telephone,
and at least one visit to a user location is recommended.
In this manner much valuable information can be gained from
those who have "been down the road" and can point out
pitfalls to watch out for.
-17-
Approved For Release 2006/09/25 : CIA-RDP73-00402R000100140011-7
 Approved For Release 2006/09/25 : CIA-RDP73-00402R000100140011-7
III SIGNING A CONTRACT WITH THE SCANNING COMPANY.
When the particular scanner has been selected a contract
should be signed with the scanning company. Elements that
should be present in the contract are:
1. Whether it is for a lease or purchase (if for a
lease, a purchase option should be provided for).
2. Maintenance support to be provided. What and at
whose expense?
3. What customer training and support is to be
provided by the scanning company.
4. Delivery date.
5. What minimum facilities must be provided for the
scanner, and at whose expense.
6. Cost of the scanner.
7. Who is responsible for delivery costs (the
purchaser usually bears the expense).
8. Who is responsible for providing insurance on the
equipment?
9. An understanding as to cost for running second and
third shifts including maintenance costs.
10. Guaranteed performance level for your particular
application.
a. Minimum numbers of documents to be processed
in a given time.
Approved For Release 2006/09/25 : CIA-RDP73-00402R000100140011-7
 Approved For Release 2006/09/25 : CIA-RDP73-00402R000100140011-7
b. Maximum number of errors and rejects per given
number of characters (or maximum number of
documents with rejects out of a given number
of documents).
c. Mean time between failure
d. Mean time to repair the equipment.
e. Size of documents handled.
f. Paper weight, etc. handled.
g. Jam rate of the paper handler.
h. Multiple feed rate of the paper handler.
i. Other criteria relied upon by you in selecting
the particular scanner.
11. An understanding as to an acceptance test, that
will measure the scanner's performance as agreed
upon.
12. Other items as agreed upon.
IV. ACCEPTANCE TEST
When a contract has been signed for the purchase or
lease of a scanner, arrangements should be made for an
acceptance test that thoroughly tests the performance of the
machine, to make certain that it meets all the agreed upon
performance levels. If at all possible the test should be
-19-
Approved For Release 2006/09/25 : CIA-RDP73-00402R000100140011-7
 Approved For Release 2006/09/25 : CIA-RDP73-00402R000100140011-7
conducted, using "live" data, of a large enough sample to
cover all conditions that might arise. If live data cannot
be used, then data should be used that imitates as nearly as
possible "live" conditions. In planning an acceptance test
it would be well to formalize it by putting the plan in
writing. The plan should clearly indicate how the
acceptance test will be conducted including a clear
definition of each one's responsibility (the scanner vendor
and the customer).
CONCLUSION
Optical scanning is a very good solution to the input
bottleneck, when it is properly applied. As with any other
good business tool it has its limitations and it is
therefore important that careful study and thought be given
to any potential scanning application, to make certain that
the right equipment for the right job are brought together.
-20-
Approved For Release 2006/09/25 : CIA-RDP73-00402R000100140011-7