OCLC's recent copyrighting of its data base focuses attention on an issue of significance for all libraries engaged in creating machine—readable records--both in current cataloging and in retrospective conversion. The issue is that of ownership, or a library's right to use and re-use the records it has created.
Little attention has been focused on this issue until recently. This is understandable. Most libraries joined a bibliographic utility to improve cataloging productivity rather than for the express purpose of creating machine-readable records. A library interested in automating and embarking on a retrospective conversion project confronts an almost overwhelming mountain of decisions. These encompass the nature of the data base: full-length or brief records, MARC or non-MARC; the cataloging standards to be followed: AACR2 or a mix of AACR1 and 2; the extent of recataloging; and the approach to authority control.
Once a library's requirements in these areas have been determined, an appropriate method of conversion must be chosen, an equally daunting task. How will the conversion be undertaken? Will all records be created de novo by keying, or will a matching and extraction option be used? Should such matching and extraction be done off line or online? If off-line, how will the search keys be entered: on typewriters using optical character recognition (OCR), or on personal micros? If an online approach is sought, what system should be used: a local stand-alone system, a commercial bibliographic service, or one of the bibliographic utilities? Will the work be performed by the library's existing staff, by staff hired especially for the conversion, or by staff supplied by a contractor? Where will the personnel be accommodated: in the library or offsite?
Unfortunately, these questions cannot be resolved academically. The answers depend on the appropriateness of available resource files of machine-readable records, and the cost of the various alternatives.
Among the plethora of immediate issues and concerns, questions of the library's right to future use and reuse of the finished file may easily be accorded a lower priority. This tendency is reinforced by the fact that a library usually undertakes conversion with a very specific purpose in mind. Most often, this is to use the file to load into an automated system to support current operations, or to provide input to a state or regional project for the development of a COM or online union catalog. Again, the immediacy of, and commitment to, that purpose discourages consideration of the other possible uses that the library might wish to make of the file in the future. Such future uses might include withdrawal of the file from a shared local, commercial or utility system for mounting on a competing system; duplication of a subset of the file on a different support system for a special purposes union list; or contributing the file to a shared automated system to support cataloging and/or conversion by other libraries. The spectrum will vary depending on the individual library.
The only factors which are constant are the likelihood—a likelihood that increases as the range of automated system operations and approaches increases—that a library will wish to use its file of machine—readable records for purposes other than those envisaged when the file was created, and the fact that a machine-readable file of library holdings represents an expensive and valuable resource. These constants dictate that a library pay careful attention to reuse conditions when assessing its options for retrospective conversion. A resource file which offers a 95 percent hit rate when matched against the library's collection and supports conversion for $1.00 per record may not be the attractive answer that it first appears if it carries significant reuse restrictions.
Different producers and suppliers of machine-readable bibliographic records have different reuse conditions and policies. The MARC tapes produced and distributed by the Library of Congress are in the public domain and thus are freely available for reuse. However, if obtained indirectly from a service organization which undertakes additional editing or input to the records, the situation may change.
A different situation obtains in relation to the “availability” of machine—readable records which present Library of Congress cataloging that was not included in its machine-readable record service. These manual records have been converted into machine-readable form by Carrollton Press and, while the cataloging on which the resulting REMARC file is based is in the public domain, the machine-readable records resulting from the conversion are not. Carrollton seeks to recoup the expenses of the project by selling the records to libraries at $.50 per record. To ensure a continuing market, the company has developed a contract which clearly defines how a library which purchases REMARC records may use them:
5.1 The Library may, without restriction or limitation reproduce or transfer in hard copy or microform the information contained in the REMARC . . . records delivered to the Library. The Library may also transfer in machine—readable form an abbreviated portion of each REMARC . . record, provided that the total number of characters transferred shall not exceed cumulatively one-half of the total number of characters of the . . . record for each title. All reproductions or transfers of the REMARC . . . records, no matter what their form or method, shall contain the REMARC identifier code.The REMARC agreement imposes usage restrictions which appear to accommodate many of the potential reuse requirements of libraries, allowing the records to be output in COM and other non-machine-readable listings, and permitting applications such as machine-readable reporting to union catalogs. However, the conditions which would apply to a library mounting its REMARC records on a local automated system shared by other libraries which may wish to access the records for cataloging purposes have to be negotiated. Many libraries have successfully negotiated such arrangements, and Carrollton Press has, on occasions, agreed to supply a small number of REMARC records in very specialized subject or language groupings for unlimited reuse by libraries. In at least one case of a specialized language grouping, the editors have verified that this reuse waiver included the right for the libraries to mount the REMARC records on a publicly accessible bibliographic utility where they could be used to support cataloging by other libraries.5.2 Except as provided in Section 5.1, the Library shall not transfer in machine—readable form the REMARC . . . records into the possession of any individual, corporation, firm, partnership, association, library, network, utility or any other person unless such person has entered into an agreement with Carrollton concerning the use, reproduction and transfer of its PEMARC data base.
The Library may allow online access to its REMARC data base for reference or interlibrary loan purposes, but shall not allow such access for the purpose of transfer to the possession of another person of REMARC data in machine language of more than 50 percent of the characters in any record. The Library shall use its best efforts to prevent unauthorized transfer of REMARC records to parties with no agreement with Carrollton, . . ."
Until recently, the major source of restrictions on usage were those imposed by record producers such as. Carrollton Press. The situation appears to have changed with OCLC's registration of ,its data base. While the utility has been silent on the practical implications of this action, approved by the Copyright Office in mid-March 1984, it serves as notice to libraries that the reuse of records obtained from the utility might be more stringently policed in the future.
OCLC recently introduced new contractual language into its retrospective conversion agreement for libraries performing such activities on the OCLC system. The contract states, in essence, that a library undertaking such conversion:
- assigns to OCLC a perpetual royalty-free worldwide right to use the records created in the conversion in the OCLC data base, and
- has the right to use the records in any manner it may elect . except that it agrees that it will not transfer the record to a third party other than:
- other members of OCLC
- a vendor that performs services for the library that requires the use of the records; in which case, the vendor is required to sign an agreement to the effect that it will not retain a copy of the record in its files
- a non—OCLC member subject to conditions to be negotiated with OCLC.
Because of the relatively recent introduction of the new contract, the editors were not able to identify any examples of record transfers made under the conditions of the third exemption. However, enquiries made of OCLC suggest that it might be possible to negotiate reuse conditions appropriate to a library's future needs. Spokespersons for OCLC stress that the utility's interest is in protecting its commercial viability rather than in limiting the activities of libraries.
Although many sectors of the library community strongly opposed OCLC's registration of its data base, a number of libraries have apparently followed suit and applied for copyright registration of their own data bases. Again, the implications of these actions are as yet unclear.
In reaction to the protests that greeted OCLC's notification of application for copyright, RLIN—-the other major utility--announced that it would not seek to copyright its data base.
The overall implications appear clear: a library should ensure that the value of its data base is not undermined by usage restrictions not thoroughly understood at the time the file was created. For many libraries the cost of developing and maintaining a data base of machine-readable bibliographic records is at least equal to the cost of hardware and software. The data base has a potential life expectancy much greater than the five to ten years of the hardware and software in a local system.
