Rutherford's message was very interesting... I see many of the same
concerns in the archival community at Harvard.  I have a couple of
additional points, which although they derive from local, internal system
needs, may illuminate some limitations in the communications format.

In discussions with archivists at Harvard it became clear that even among
themselves, they did not agree on what it meant for materials to be under
archival control.  If two libraries here have the same pre-1800 pamphlets,
one might choose to consider them to be under archival control, while the
other did not.  Should archival control be a bibliographic record-level
characteristic or a holdings level characteristic?

On the other hand, Harvard intends to use the archival control byte
essentially to circumvent the new rules for coding type of record,
and for that purpose it will be handy to have it at the bib level.
We've been asked to continue to support search qualification by
manuscripts format, and to include in the results of such a search
anything which

        -       is cataloged in the old AMC format
        -       is cataloged as the new mixed materials format
        -       has a manuscript type of record code in the LEADER or in
                        any 006 field (i.e. d, f, or t) ; or
        -       has the type of control byte set to 'a'.

If we do this programming, we'll be jumping through hoops to give users
the kind of access they had to archival materials BEFORE the MARC format
changed.  We would also be setting two types of search qualification
based on a single code -- e.g. a collection of correspondence (Type of
record 't') would be searchable in HOLLIS both as a manuscript and as a book.

Harvard also uses format information in index displays, in the form of a
three-letter code at the end (e.g. bks, ser).  These codes are admittedly
reductive, but users have in the past found them to be useful.
That will be less true now that individual broadsides or collections
of letters are to be coded as if they were books.  Let's face it,
"they're all language material" is not a useful way of explaining to
people why they can't tell manuscripts from published books at the index
level.  So we may end up deriving the display of format information from
some hierarchy (e.g. manuscriptness takes precedence over bookness),
which feels to me like circumventing the guidelines again.

We'll be better off if we completely chuck the old format concept and
redesign search qualification and display in terms of our own record type
groupings.  Unfortunately, format is buried throughout our system,
and as you all know, it's easier to manage 7 or 8 distinctions than 14 --
both for systems and for users.

It seems that for archival materials, we may have lost something in our
attempt at rationalization.  I know I'm several paces away from clarity
on this topic, and I'd love to hear what others are experiencing.

Robin Wendler  ........................     work  (617) 495-3724
Office for Info. Systems  .............     fax   (617) 495-0491
Harvard University Library  ...........     [log in to unmask]
Cambridge, MA, USA 02138  .............