> As Michael Fox said last week, the problem with this route is that
> searching EAD-encoded finding aids shouldn't be restricted to just Dublin
> Core or MARC-type elements (author, title, subject, etc.) but often can
> include structural elements of the document like "inventory" or "summary
> collection description". You can't map that stuff to BIB-1 attributes.
> Furthermore, the mapping of the traditional access points gets difficult
> with finding aids.
This point is well taken. It's not simple. And Z39.50 access to EAD should
not be restricted to mediation through MARC or DC, but I would think it
very useful to have those routes tested out too, since of course some
larger systems will use either of those to obtain access to single or
multi-format resources and of course for talking between systems
(especially those in which EAD is one but not the only format). But besides
knitting EAD into a broader information universe, it seems without saying
worthwhile to have a straight EAD<-->Z39.50 profile. In fact Paul Watry
(Liverpool), us and some other folks here at Berkeley had a JISC
application in to work on just that - it did not get approved, but I think
all might be willing to participate or lend a hand if something is cooking
up? Just a thought.
It seems one of the best ways to start exploring is to have some testbeds
going on for the various access methods, and see what works and what can be
learned. That's why I'll be very anxious to get our own Z-system up and
running and publish the various semantic mappings so I can see how they
actually play out, and can get feedback from "those in the know" like
people on this list. I know I made some decisions in doing mapping that
definitely take the access into a particular conceptual direction; for
instance do you provide access and mappings to the primary materials
(author=equals the author of the collection) or to the secondary
(author=author of finding aid)? I'm eager to get input on the results and
will post to the list.
Thanks for pointing out your project, I think it sounds very interesting. I
wonder if the Online Archive of California project could be another likely
venue for testing some of these multi-institutional access issues as well,
perhaps even in coordination? We are participating on the "Museums and the
Online Archive of California" project, but perhaps there are others from
the OAC on this list who might speak to that idea? Just another thought....
Richard Rinehart | Berkeley Art Museum/Pacific Film Archive
Systems Manager & Education | University of California
Technology Specialist | 2625 Durant, Berkeley, CA 94720-2250
[log in to unmask] | http://www.bampfa.berkeley.edu/
& President-Elect, Museum Computer Network, http://www.mcn.edu/
> As Michael Fox said last week, the problem with this route is that
> searching EAD-encoded finding aids shouldn't be restricted to just Dublin
> Core or MARC-type elements (author, title, subject, etc.) but often can
> include structural elements of the document like "inventory" or "summary
> collection description". You can't map that stuff to BIB-1 attributes.
> Furthermore, the mapping of the traditional access points gets difficult
> with finding aids. For example, "author" is usually defined as the
> repository which authored the finding aid, not the author who is the
> subject of the collection. And "title" could be the official title of the
> finding aid in some cases, and the <title> elements in other cases.
> "Subject" includes, in our case, eight different elements with
> subject-like content including the <subject> element, but not the same
> elements choosen at other sites, I'm sure.
>
> As far as I can see, there's no way to enforce consistency of indexing
> decisions for EAD, even if the markup is internally consistent across
> different repositories.
>
> I think this is a very complex and difficult problem with searching
> finding aids. For this reason we will be participating in a small project
> with some other universities to try to develop a set of common, useful
> access points for cross-collection search criteria on finding aids. The
> project will be lead by the University of Michigan and minimally include
> Harvard, Columbia, and Oxford. Our hope is that if we can figure out what
> kinds of access are of common utility we can help to develop the Z39.50
> Collections profile (or EAD profile, or whatever makes the most sense by
> then).
>
> If you're only using Z39.50 to search your own collection then it's not as
> much of a problem and there are some short-term options, as others have
> pointed out. But this mapping isn't quite as simple as it may have
> sounded...
>
> _____________________________________________________________________________
> MacKenzie Smith [log in to unmask]
> Digital Library Projects Manager phone: (617)495-3724
> Office for Information Systems fax: (617)495-0491
> Harvard University Library %\%\%\%\%\%\%\%\%\%\%\%\%\%
> -----------------------------------------------------------------------------
Richard Rinehart | Berkeley Art Museum/Pacific Film Archive
Systems Manager & Education | University of California
Technology Specialist | 2625 Durant, Berkeley, CA 94720-2250
[log in to unmask] | http://www.bampfa.berkeley.edu/
& President-Elect, Museum Computer Network, http://www.mcn.edu/
|