To: CONSRLST, EMEDIA, VPIEJ-L, SERIALST
          From: Bill Anderson
          Re: Cataloguing Records for Remote-Access Electronic Serials
          Date: February 14th

          Many thanks to Wayne Jones for furthering the discussion on
          multiple versions of e-serials, and to the others who provided
          fruitful comments on what I find to be a complex situation.

          I think Wayne makes an excellent point with regards to the
          preferred arrangement of *sub-records* for various versions
          linked to a single bibliographic record.  Also, the two criteria
          suggested--substantial difference in content and significant
          technological/access differences--is interesting.  Mr. Lamp
          offered a different criteria:  a change analogous to edition
          change, or significant change in the **nature** of the material.
          The hypertext version is significantly different in nature, in
          his view.  And Priscilla Caplan has noted that USMARC does allow
          for small variations in format to be included on a single record.

          In the six weeks since the original posting first appeared I have
          investigated a number of e-serials to categorize the various
          types as to their access methods and the number of versions in
          which they are available.  I considered the version question
          separate from the access issue since single versions can have
          multiple access methods and multiple versions may be found at a
          single location.  What I discovered from the batch I investigated
          is that about half are available in more than one version, and
          about a third offer multiple access methods--email, FTP, remote
          login (gopher/Web/telnet).  The version/access relationship is
          quite varied as some Web versions are ASCII and some gopher
          versions are formatted (eg., PostScript).

          When one actually compares the different versions one can find
          content differences, such as graphs and figures included in a
          formatted-text version but not found in the plain-text (ASCII)
          version.  And some ASCII files also have hypertext versions that
          include links to graphics or image files.  But far more
          frequently I did NOT find *substantial differences* between
          versions.  Even hypertext versions often do not link to files
          outside the basic text file but typically take you to the
          footnotes and back (and back to the home-page).

          Caplan states:  *What _AACR2_ may or may not require is a
          different issue.*  Yes, cataloging is still a rules-driven
          operation and standards are fundamental to cooperative programs.
          _AACR2_ does have a specific definition for computer file
          editions:

             **All copies embodying *essentially the same content*
             (emphasis supplied) and issued by the same entity.**

          Also, the 9.2 Edition Area is rather fully developed and has
          virtually no LCRIs associated with it.  Several examples for
          edition statements use *version.*  A reference back to 1.2B
          reminds us that in 1.2B3 (as well as in 9.2B2) you see:

             **In case of doubt ... take the presence of such words as
             edition, issue, or *version* (emphasis supplied) as evidence
             that such a statement is an edition statement**

          9.2B4 is interesting in that it suggests that the cataloger
          distinguish between minor changes in files (e.g., spelling
          corrections, output format, display medium) from *significant
          differences.*  A reference to 9.7B7 indicates that minor changes
          can simply be noted.  9.2B3 (optional addition) suggests that
          when one finds *significant changes from other editions (e.g.,
          changes in the data involving content ... the addition of sound
          or graphics ...)* an edition statement may be supplied.

          It would seem to me that as the rules now stand the cataloger
          would generally interpret a statement of *version* to indicate a
          separate edition, but that the cataloger is expected to
          differentiate between *significant changes* and *essentially the
          same content.*

          If the vast majority of e-serial versions do have *essentially
          the same content* then perhaps we can generally rely on one
          record for different versions.  Caplan explained that USMARC does
          allow for this quite adequately.  On the other hand *significant
          differences* are debatable as Jones and Lamp indicated.  I would
          emphasize *differences in content* rather than *differences in
          access/technical means.*  I think the rules do not adequately
          address differences in navigational means (i.e., hypertext
          links).  But there does not seem to be a **requirement** in the
          rules to create separate records for files that are
          differentiated only in the way you navigate within or between
          them, or to external documents.  The only **requirement** I can
          detect involves versions with significant changes or differences
          from existing files.

          In today's cataloging world of cooperative programs, cataloger's
          judgement, and the core standard, I think flexibility is crucial
          when dealing with complex bibliographic issues.