In the late 1960s, the Library of Congress started work on two major achievements. The massive National Union Catalog Pre-1956 Imprints eventually reached 754 volumes and took 130 feet of shelf space. Computerizing library cataloging soon rendered it obsolete. Henriette Avram, who never went to library school, led the team that devised MARC (MAchine Readable Cataloging).

The primary purpose in developing MARC was to automate printing library catalog cards, but Avram and her team also intended it for libraries to be able to print their lists of  new acquisitions.

The Library of Congress supplied card sets for its entire holdings to any American library that ordered them. That saved libraries the expense of cataloging those materials themselves. Avram did not envision putting the entire library catalog online and making cards obsolete, but her work made it possible.

In fact, MARC enabled the online public access catalog, which saves so much time and space. Patrons had to visit the library at times when it was open in order to use it. Now, anyone with internet service can use the catalog and interact with it without leaving home. Or use it anywhere in the library without leaving where they’re working.

The MARC format drove innovation both in library technology and  computer technology. Magazine mailing labels show the computer’s ability to handle textual data in the late 1960s. They use capital letters only, minimal punctuation, and limited field sizes.

A library catalog needs the full title of an item and names associated with it (including authors and publishers). And so it needs not only upper and lower case letters and a full range of punctuation, but also diacritical marks. Even though the MARC record was at first limited to English-language materials, some authors’ names used them.

Who was Henriette Avram?

Avram was born Henriette Regina Davidson in Manhattan in 1919. Nothing in her background could have predicted the course of her career or her influence. Although she gained in the fields of computer programming and library cataloging, she had no formal education in either. She aimed at first for a medical career.

Davidson married Herbert Avram in 1941. In the early 1950s, the couple moved to Washington, DC, where he took a job with the National Security Agency. Instead of continuing her medical studies, she studied mathematics at George Washington University. It didn’t take long before she, too, took a job at the NSA. There, she learned computer programing. She later observed,

Learning programming in those days was not like it is today (1989). It was a bootstrap operation. You were on your own with far less than perfect tool to learn from. The men—women—were quickly separated from the boys, and the numbers of people who made it through were few indeed.

Avram joined the Library of Congress in 1965 and took charge of the library’s exploration of automating its systems. To do that, she had to learn about cataloging, a discipline with very intricate rules. The Library of Congress collection had materials in hundreds of languages, some that don’t use the Roman alphabet. It also had a wide variety of formats—not only books and serials but also maps, sound recordings, films, and more.

Avram took charge. She was energetic, well organized, and persuasive. At the same time, she was diplomatic and able to inspire her team and all the organizations she worked with to cooperate on a shared vision.

Before describing her work on MARC, we will look in more detail about what library cataloging means.

An overview of library cataloging

Cataloging is divided into descriptive cataloging and subject analysis. Descriptive cataloging transcribes such information as the format, author, title, publisher, place of publication, and size. Subject analysis includes subject headings and classification numbers.

Descriptive cataloging operates according to a set of rules. But library-land had completely reconceived and rewritten its rules more than once by the time Henriette Avram started working at the Library of Congress.

Subject analysis does not follow formal rules. The Library of Congress, the National Library of Medicine, public libraries, and others all had their own classification systems and headings.

All this information fit on 3×5-inch catalog cards. But each card conformed to whatever rules were in force at the time it was made. Some cards were hand-written in a formal “library hand,” some typed. and some typeset.

Human eyes have no trouble dealing with all the variation represented in a standard card catalog. Computers have to be programmed to understand every minute detail.

Controlled vocabulary in library cataloging

Cataloging rules require that every person or organization must have one and only one authorized form of their names. John Adams, for example was the second President of the United States. But John Adams is also a 20th- and 21st-century composer. And those are only two of myriad men named John Adams represented in library collections.

Names change over time. Women represented in the catalog get married. Men might also change names. Prolific German theologian Joseph Ratzinger became Pope Benedict XVI. Corporate bodies change their names, too. The authorized form of the name must be the most current form. Some fiction authors adopt one or more pen names.

Behind the scenes, therefore, the Library of Congress maintains a name authority file. Every name change requires changing the corresponding name headings in every cataloging record. It also maintains a subject authority file for Library of Congress Subject Headings.

In the days of card catalogs, changes to names or subject headings required someone to pull entire card sets, change the headings, and refile the cards. But now, authority records  are also MARC records. Name fields in bibliographic records link to authority records. Changing an authority record automatically changes all the bibliographic records linked to it.

Shared cataloging

The Library of Congress is the largest library collection in the US, but plenty of libraries own materials that the Library of Congress does not. Any library could obtain Library of Congress card sets. The National Union Catalog collected the holdings of numerous other important libraries. Local cataloging departments could consult it to find and copy cataloging from these other libraries.

The National Union Catalog Pre-1956 Imprints represented only one portion of the National Union Catalog. It also comprised holdings compiled in several other smaller chronological units. All these volumes contained page after page of pictures of catalog cards.

The Ohio College Library Center pioneered the idea of shared MARC records among multiple libraries in the 1970s. It has since grown to a mammoth worldwide organization now known simply as OCLC.

OCLC member libraries need only find a master record on OCLC and attach their holdings. Any changes made to the master record instantly appear in every local catalog that uses it.  MARC and OCLC have revolutionized interlibrary loan as much as cataloging.

Henriette Avram and the development of MARC

The Library of Congress began to investigate automating its internal operations in the late 1950s. Its Council on Library Resources (CLR) published a study in 1963 that recommended formation of a group to “design and implement the procedures required to  automate the cataloging, searching, indexing, and document retrieval functions.”

Planning for MARC began under Henriette Avram’s direction in 1966. Her team also solicited participation of other libraries and selected 16 partners. These included government, academic, public, school, and special libraries from all over the country (and one in Canada).

The plan was to start distributing cataloging in September 1966, which gave Avram’s team just eight months to invent a computer program for the purpose and begin to use it. It amounts to learning to design an airplane and fly it at the same time. The pilot program lasted until June 1968, a full year after the projected end date.

Library collections contain much more than books, but designing the project required haste. It meant that it would begin with books and start on other formats only after problems with books had been discovered and solved.

By the end of the project, the library had distributed about 50,000 machine-readable records for books in the English language. Implementation of a national distribution system soon followed. The Library of Congress began  specifying MARC descriptions for formats other than books when it issued formats for serials and maps in 1970. It also started to create and issue MARC records in non-English languages, and eventually, for languages that don’t use the Roman alphabet.

What is the MARC format?

This is what a MARC record looks like to a cataloger:

Each three-digit number down the left hand side of the record designates a particular database field. Some ranges of numbers represent names of people, government entities, or corporations. Others represent the call number, the title and its variants, publisher information, the size of the item, the series if any, and various kinds of  notes. The MARC format also has a number of control fields that mean something only to the computer.

The two digits to the left are called indicators. Some fields have no indicators defined. Others use only the first or second indicator. Some use both. How does the catalog know to alphabetize by the first significant word of the title, not the or a or an? An indicator tells it how many characters to skip.

Then the actual content of the field begins. All those letters and numbers with a dollar sign before them are subfields. Each little bit of information must go into the  proper subfield.

All this granularity enables searching and sorting in as many ways as anyone needs. Imagine what it took for Avram’s team to put it together in just a few months!

Conversion of older records to MARC

Once large-scale distribution of MARC records began, it became necessary to convert all earlier cataloging records to the new format. And that not only in the Library of Congress collection, but every other library.

So as the Library of Congress had long taken the lead in shared cataloging and distribution of card sets, Henriette Avram also took the lead in “retrospective conversion.” Any library could get MARC records for items in the Library of Congress collection. Therefore, other libraries only needed to perform their own conversion on materials the Library of Congress did not yet have MARC records for.

MARC specifies various levels of cataloging. Not all libraries need the level of detail of Library of Congress cataloging. Using MARC, they can choose a suitable level and contribute their records to the central database.

International adoption of MARC and its implications

Avram recognized that if the MARC system would be used outside the Library of Congress, then it ought to be a standard. Therefore, she worked closely with both the American Library Association and the American National Standards Institute.

The British National Bibliography showed interest in devising its own MARC format early on. So it obtained recognition of its version from the British Standards Institute. The two institutions submitted their standards to the International Organization for Standardization. MARC’s acceptance there made it possible for other nations to develop their own versions of MARC.

Planning for international cooperation in cataloging predates MARC. The International Conference on Cataloguing Principles met in Paris in 1961 to agree on what have become known as the Paris Principles.

As part of implementing them, the International Federation of Library Associations began work on International Standard Bibliographic Descriptions in 1969. It then published the first draft in 1971. It is difficult to understand how such work would have been possible without development of the MARC standard.

ISBD made it possible for anyone familiar with it to identify the title, author, publisher, and other information without knowing the language.

The growth of online catalogs in general and OCLC particular made much of this work obsolete. OCLC has gone far beyond the concept of a national union catalog. Major national libraries from all over the world cooperate with it. And so its mammoth database is rightly called WorldCat. It includes at least some holdings from libraries in more than 170 countries.

Henriette Avram and her team knew that their project could revolutionize librarianship. They just had no idea how far it would reach.

Sources:

The evolving catalog: cataloging tech from scrolls to computers / Karen Coyle, American Libraries. January 4, 2016
Henriette D.  Avram, modernizer of libraries, dies at 86 / Margalit Fox, New York Times. May 3, 2006
MARC; its history and implications / Henriette D. Avram, Library of Congress MARC Development Office, 1975
“Mother Avram’s Remarkable Contribution”: Henriette D. Avram / Lucia J. Rather and Beacher Wiggins, American Libraries. October 1989