iassist quarterly

Introduction

An Archivist's

Challenges:
Adapting to Changing

Technology and

Management Techniques

Over twenty years ago, the National Archives of

the United States embraced the concept that

automated records were actually records which

could be considered permanent within the

meaning of the Federal Records Act and set

about collecting them. Since then it has

confronted problems incident to finding these

automated records, acquiring them, preserving

them and making them available to the public.

Previous papers have discussed access to public

automated records in the normal sense; that is,

the ability of the researcher to get at them. In

this paper I wish, however, to discuss the

National Archives' acquisition process as a form

of access.

by Donald Fisher Harrison'

National Archives and Records Administration

Washington, D.C., United States of America

This paper addresses three threats to the

acquisition of machine-readable records: the

threat of an onslaught of hardware and software

incompatibility, the threat of discontinuity within

textual records series brought about by

end-users with microcomputers and the threat

brought about by new management techniques

from the Paperwork Reduction Act of 1980.

Archivists ought to view these threats as

challenges. When overcome, the challenges will

have presented the Archives with the

opportunit>' to create a better collection of

automated records.

'Presented at lASSIST/IFDO Internationa!

Conference May 1985, Amsterdam.
** The views expressed in this paper do not
necessarily correspond with those of the
National Archives and Records Administration.
I wish to acknowledge that some of the material
has been developed out of long standing
collaboration with fellow archivist Dr. W. Jon
Heddesheimer, but any mistakes in concept or
fact are mine.

Software and hardware dependency

The first challenge to the National Archives is

well publicized and needs no significant

introduction in this treatise. The Archivist of

the United States, confronted with the research

community's complaint that valuable data were

being created by Federal agencies without any

consideration for their preservation or

dissemination to the public, established in the

Spring 1986



(assist quarterly - 9

1960's the forerunner of today's

Machine-readable Branch. This branch was

given the task of inventorying Federal data

bases and deciding how best they should be

preserved for posterity. We accessioned a

number of machine-readable data files created

in the 1960's. Some of these files were

dependent on other outside factors and could

not be read on their own. Three examples of

software and hardware dependency illustrate our

initial problems.

The first example came early in our

organizational being. We received over

thirty-five machine-readable data systems from

the Office of the Secretary of Defense and the

Office of the Joint Chiefs of Staff. These

systems were encoded in a data base

management system called the National

Information Processing System (NIPS). They
caused serious problems in access and handling

and a considerable backlog in the accessioning

workload.

NIPS was devised for generalized file handling

using languages designed to support user

requirements in six components. It afforded any

data center the capability of reporting long and

involved statistical manipulations on extremely

short notice to a variety of users. However, the

software was compatible only with IBM
computers.

The presence of NIPS files suggested serious

difficulties in providing a uniform reference

service to researchers and brought up the whole

question of software dependent files. To retain

the files in NIPS would constitute a precedent

Since researchers by and large preferred to use

their own utility software, transportable files

would afford a range of options that encoded

files would not Last but not least maintaining

large inventories of software would add to the

preservation costs and require more shelf space.

For all these reasons we decided to decode the

files. It appears now, with hindsight that

despite the fact that these files were unique and

very valuable, we should have insisted that the

material be transportable before being accepted

by the National Archives.

The second example was the National Archives'

accessioning of a microfilm series of records

containing pictures of captured North

Vietnamese documents. These were filmed in

Saigon during the war by the Combined
Document Exploitation Center on 94 oversize

(13-inch) rolls of 35mm microfilm, each roll

1(XX) feet long. The documents were on one

side of each frame, with digital bar codes on
the other side to provide indexing and control

information.

Soon after we received the microfilm we
discovered to our chagrin that the material was

hardware dependent in a system known as "File

Search." Four configurations of this machine

had been manufactured and sold to Federal

agencies in the 1960's. The last model

(generation four) had a small computer in it It

could therefore provide a printout by reading

the bar code on the film strip, transferring it to

magnetic tape, which in turn could be

manipulated and dimiped on to paper. The
machines cost $250,000 new and were used only

by military and intelligence agencies, as far as is

known.

It was only after this information was made
available to us that we discovered that other file

systems were known to exist in this environment

and were equally unreadable without any

machines in existence to retrieve data. These

included some important files in the Navy Sea

Systems Command (in Arlington, Virginia) and

the Navy Oceanographic Command (in Bay St

Louis, Mississippi), including the Defense

Intelligence Agency. Recently we have

discovered the existence of an intact File Search

model in salvage channels. We have requested

that it be turned over to the National Archives,

and we think we have the technical expertise to

restore the model to operating condition.

Spring 1986



10 - iassist quarterly

The third example entailed the 1960 Decennial

Census, offered to the National Archives by the

Census Bureau in the mid 1970's. These

records were created by a UNIVAC II-A

computer, of a generation that had been

effectively phased out of use in the Federal

government after the tapes had been created. It

has been reported that once the tapes became

available for transfer to the National Archives,

only two such machines existed to read them,

one in Japan and one in the Smithsonian

Institution.' Eventually a reasonable approach

was agreed to by the National Archives and the

Census Bureau, to convert the data into a

compatible format, making them available for

preservation in our vaults.

These three examples are illustrative of the long

term problems aeated by hardware and

software dependency of records created in the

1960's, when computers were maintained in

relative isolation from each other. It was a

period in which data managers were concerned

with the CTcation and the use of computer

products and were by and large ignorant of the

long term value of these products as Federal

records. It can be said that the letter of the

law — the fact that the tapes were handed over

to the National Archives — was carried out
The fact that the tapes were unreadable because

of software and hardware dependency was a

new problem that had never been faced with

paper acquisitions. For their part, agencies were

understandably reluctant to dispense funds solely

for the benefit of depositing these records in

the National Archives. Thus reason has had to

prevail in our dealings on transfer of the tapes,

and no one solution can be applied in all cases.

Small computers and office automation

The second challenge to the smooth flow of

records into the archives stems ironically from

the very machine meant to facilitate

administrative operations in the modem office.

For several years, most federal agencies have

been extending the advantages of their word

processing pools by placing terminals at the

hands of management officials, giving fingertip

control to their own records creation. Office

automation (AO), more aptly termed "the

paperless office", is based on a series of

compatible, menu-driven programs utilizing a

centralized data base for common shared-use

data and unique smaller databases for individual

users. These systems have the ability to transfer

data and information between data bases

through a network or a distributed system.

The advantages of such a system are obvious.

Federal managers frequently need information

suddenly and immediately, and often the

demands for this information come after the

staff has left for the day or the weekend.

Managers would like the ability to search for

the data or reports they need through an

indexed automated bibliographic/numeric data

base, access and use the appropriate software to

perform simple to moderately complex analyses

of this data (e.g. forecasts, conelations, etc.), use

graphics to illustrate their results, access word

processing/office automation tools to produce a

memo in the appropriate format, and finally

send this report/memo electronically to the

recipient's office, all without the necessity of

using the phone, typewriter, or staff that are not

avjulable.

'Commiaee on the Records of Government,

Report Washington, DC, March 1985, pp.
86-87.

Keeping all this in the system can cause an

archival "log jam". The designers and the users

of paperless office systems are frequently

ignorant of the paper systems they are replacing

and the archival need for intellectual continuit>'.

Outside contractors compound the problem. In

Spring 1986



iassist quarterly - 11

the absence of any other information the

hardware and software dependency problem has

reemerged in the small computer world, and

agencies are finding that transportability cannot

cross the boundaries between offices. Software

now provides end users with ultimate fingertip

access. This allows handcrafted programming

and instant manipulative gratification. The same

person who creates data on the system can now
dispose of it with equal ease. By closing the

gap between the user and the machine, the

system eliminates the apparent need for the

data middleman, lo say nothing of the records

manager who, under other circumstances, looked

after standardized formats, ensured traditional

records disposition practices and provided for a

continuity of records series in the agency.

Thus the danger inherent in OA is that the

practice concentrates on the information as it is

used immediately after creation without making

a record of actions taken. It is said to parallel

the dangers of telephone use when first

introduced. With that instrument, mjinagers

needed go through no intermediate device for

communication. Telephones assured privacy of

communication and were sheltered from the

public record. The comparison with OA is

evident Just as managers could converse at the

push of a telephone button, so they do now
with electronic mail. Further, if one of the

parties is absent, there need be no callback,

because the mail has already been delivered

electronically.

Like the telephone, the OA challenge is to find

a way to record the communication. With the

small computer, software must be devised to ask

the user for a determination of the ultimate

value of the information before it is ever keyed

into tlie system. This software has been

integrated into the planning for OA systems in

most Federal agencies. Whether or not it solves

the problem in practice remains to be seen.

Information resources management

The third challenge to a smooth transfer of

records to the archives now comes in the form
of an application of new management techniques

to the CTeation and the use of information

within the Federal establishment. This new
methodology typically accommodates the reality

that government must function with less

personnel and with individuals of lesser skill

and training by altering the way agency missions

are carried out The Paperwork Reduction Act

of 1980 was rightly concerned with a problem

that had existed for some time in that the

Federal government was preoccupied with the

physical problems associated with the large

volume of paper records created. The authors

of the bill reasoned that managers should have

been concerned with how the information was

being controlled and how it could be shared

with the maximum number of sources. Thus
the new law espoused intellectual control

vis-a-vis physical control, regardless of the

medium on which the information had been

stored. In order to do this a number of

organizational changes have taken place in

Federal agencies, each a bit different from the

next, in which an "information Czar" has been

placed at the highest levels to control access

and dissemination of all information, regardless

of the medium. This new arrangement has now
been entrenched for four years.

A typical arrangement has been established to

combine the former functions of "automation,

communications, office automation, records

management, publications, audiovisual activities

and other information activities, services and

facilities." An information management plan is

usually mandated beginning with a problem
analysis, designing a model information system,

constructing the "architecture" which produces a

program and provides guidance for a budget

request Under this concept, every information

system will have centralized management The

Spring 1986



12 - iassisl quarterly

"single manager" concept has been extended to

encompass all information, defined as "... all

processes by which the user may receive,

display or project desired information...

(including) voice, text, graphics, audiovisual,

video teleconferences, micrographics, files,

records management, optical discs and other

forms of published information."'

In many ways, the single manager system makes

a lot of sense. The information manager is in a

unique position to disseminate information

within an agency to avoid duplication of efTort

— or better, to avoid disparate and conflicting

data creation. By being organizationally placed

at the highest level, the IRM provides

information for important decisions and controls

a sizeable portion of the agency's budget

Furthermore, the concept will ease the path of

liaison between the agencies and the National

Archives. As we began to accession records in

machine-readable form in the 1960's, we
became increasingly aware of the presence of

the data manager as a viable records aeator

and manager. Between 1961 and 1980, the

Machine-readable Branch frequently

communicated with the data manager directly

when it was not able to get required

information any other way. Furthermore, in the

first half of this decade, we became more and

more . concerned with deahng directly with

government managers since they were creating

(and destroying)' information without

acknowledging either the Federal Records Act

or their agency records administrator. With the

advent of the IRM principle, however, the

Archives need only deal again with one official,

who, if properly briefed on the urgency of the

problem, will coordinate the actions of the

records manager, the data manager and the end

user.

'Draft Army Regulations 25-1 and 25-5.

1984.

Conclusions

Technology has created new solutions to old

problems, but in the process has itself aeated

new problems. The archival community is thus

confronted with unique challenges to its

traditional role as keeper of the records, which

requires our attention. Some measures come to

mind as actions to stem the tide.

First, the archivist must keep professional pace

with the proliferation of computing technology,

not only as it is practiced in Federal agencies in

this decade, but also as many writers envision

that it might be practiced 25 years from now.

Reading the literature is not enough. It

requires a shrewd selection of educational

services and an on-going dialogue with other

archivists. This must include at a minimum the

study of software, hardware and storage media

as trends develop. An archives must be capable

not only of receiving machine-readable records

in various modes and written on various media,

but also of serving its users with a multiplicity

of anangements.

This leads to the second measure. The archivist

must determine far enough ahead in time in

what mode and on which medium these new

records will appear as candidates for acquisition.

To do this, archivists must assert their

professional needs to the creators of records

throughout the life-cycle of the records.

Furthermore, the requirement to deposit tapes

and other media in the National Archives

should be anticipated and budgeted by Federal

agencies.

Third, the archivist must reach end users by

some means, to ensiu-e standardization of

practices and procedures. It is vitally important

to overlay records management practices on the

uses and outputs of small computers and of

office automation systems. This might include

commimicating with procurement ofTicers and

Spring 1986



iassist quarterly - 13

IRM officials to standardize hardware and

software packages which would be

interchangeable within and between Federal

agencies.

The fourth, and by no means the least

important, point is that the IRM developments

in Federal agencies, formed as a result of the

Paperwork Reduction Act of 1980, must be

influenced by direct communication with

archival officials.

IRM managers have been imbued with the

immediate needs of the agency information

program in mind — the here and now concept

There is always the danger that not enough
planning will be conducted for the ultimate fate

of records. By the way they maintain certain

modes of information, IRM officials can

influence the disposition, and in turn, the

configuration of future holdings of the National

Archives."

RECORDS ACQUISTION LIASON FOR FLOW OF RECORDS
BETWEEN

THE NATIONAL ARCHIVES AND FEDERAL AGENCIES

/'

ADMINlSTftATOn

BEFORE I960

X

RECORDS \
'' WiCHIVIbT

' ftOMINIETRATOfl^

mDP

1

I

CtIO /

\_ MAIIACER USER /
\^ y

——

_

1

-"''^

's IHFORTUTION
J

\ RE'SDURCES /
\ rW-i^GER /

1981 - 1985

Spring 1986