To:       Serial Colleagues
From:     Jean Hirons, Acting CONSER Coordinator, Library of
          Congress
Subject:  Digital reproductions and multiple file formats
Date:     October 17, 1995

This message is being posted to CONSERLST, SERIALST and INTERCAT;
please excuse the duplication.

The CONSER (Cooperative Online Serials) Program's Task Force on
Electronic Resources is investigating a number of issues related
to the cataloging of electronic serials in order to make
recommendations to the Policy Committee, which is meeting in
early November.  One subgroup has been charged with recommending
policies for the description in the CONSER database of multiple
file formats and digital reproductions.  We realize that firm
policies cannot be agreed to at this point; however, we would
like to provide recommendations based on current experience.
Please share your thoughts with us on the topics and questions
below.

1.  Multiple file formats.  CONSER is currently using one record
to represent all file formats for serials (e.g., ASCII,
Postscript, hypertext), including those with html encoding.
This is based on a belief that the file formats in which a serial
is issued are likely to change (at least in the near future as
technology develops) and that the information given in one record
is adequate to inform patrons of what is available.  This policy,
which was adopted last spring, has met with wide-spread approval
and a few strong voices of dissent.  Are there further thoughts
on this issue?

2.  Digital reproductions.  Some libraries and companies are
beginning to digitally reproduce serials in whole or in part and
we expect this to be a growing trend.  What constitutes a
"digital reproduction" is the first issue to be addressed.  If
only partial contents are reproduced (e.g., text of a news
broadcast), is this a reproduction or a new item?  For purposes
of this discussion, assume that the entire contents have been
digitized (with the possible exception of ads) for either the
entire serial or a significant portion of the issues.

We are fully aware of the displeasure with the lack of a multiple
versions solution for microforms and know that many libraries are
now using local holdings records to cover microform reproduction
holdings.  While this works well for local needs it does not
serve the national catalog and cannot be adopted as CONSER
policy.  What are the options for digital reproductions and are
they different enough from microforms to enable a different
approach for CONSER?  Basically, the options are:

     1) Treat like microforms--separate record based on original
          with 533
     2) Catalog separately but base the description on the form
          of item--use 534 for details of original (this would be
          the pure AACR2 approach)
     3) One record approach--add a 530 and 856 to the record for
          the original (the 530 would indicate something like:
          "Also available online ..." )
     4) Indicate locally in holdings records (cannot be a
          "CONSER" approach but may be the most widely used)

I've listed options 1-3 below with my thoughts on the pros and
cons (these are not complete) and some questions.

1) Treat like microforms

     Pros:
          similar approach to all reproductions
          can use a GMD of [computer file]
          can give all details of the computer file (538, 856
               etc.)
          can indicate who did the reproduction (533)
          clear what your library has access to

     Cons:
          separate record needed

     Questions:
          records for the original could be cloned but what about
          the fixed fields--should they reflect the original (as
          is current policy for microforms) which would mean
          serial 008 and CF 006 or should they reflect the
          overall policy for format integration--CF 008 and
          serial 006?

          are libraries willing to follow this approach?

          how much of the reproduction information is important
          to libraries and patrons?

2) Catalog separately, base record on reproduction

     Pros:
          conforms with AACR2
          computer file can be fully described
          clear who did the reproduction

     Cons:
          doesn't conform with our treatment of other types of
               reproductions
          requires a separate record

3) One record approach--add 530 and 856 to record for original

     Pros:
          doesn't require a separate record or extensive
          cataloging
          fast and easy to do
          may suit needs of patrons better than two records

     Cons:
          doesn't provide for full description of the computer
               file (e.g., no GMD, 007, 538, etc.)
          relies on the original record with the potential of
               "mucking it up"!
          may be unclear to a patron what the library has or has
               access to, particularly if the original is not
               held
          "holdings"(i.e., access) are unclear in the national
               database
          may be more difficult to later change, if a real
               multiple versions solution is fully implemented

     Questions:
          would a 530 and 856 be sufficient to meet the needs of
               your patrons?
          would it be sufficient for library access, control,
               etc.

          how important is the information that could not be
          included (538, 007, issuing body, etc.)?

I realize that this is a lot to think about and many libraries
have yet to deal with these issues, but we appreciate your
thoughts.  Please reply to the list, if possible by Oct. 25th.

Jean Hirons
Acting CONSER Coordinator
Serial Record Division
Library of Congress
(202) 707-5947
jhir@loc.gov