On Mon, 29 Sep 1997, Richard Higgins wrote:
> This is a very good idea where there is an existing model for finding aids,
> such as APPM, and for
> those created from scratch. However, as someone who is using EAD for
> retrospective conversion
> work, I have found that you must rely on the structure etc of your
> existing material. It is not feasible
> to discard over a century's worth of listings and start from scratch,
> or even to refer to the original
> material except for in the most extreme cases of bafflement. Thus all
> you can reflect in the EAD
> file is what you have already got in your finding aid.
Richard Higgins makes a great point. One thing that has concerned me
about this collective endeavor is that we seem poised on the brink of
basing EAD 1.0 and all its documentation on the retrospective conversion
of existing finding aid information, which seems like a big mistake to me.
I would argue that we need some way in EAD to denote that an instance of
an encoded file is "retrospective", but that we need more input from
people who are using EAD from scratch to encode finding aids that are
compliant with some descriptive standard like ISAD(G). Future encoding
practice and the construction of systems to deliver encoded finding aids
need to be informed by "best practice" and not "best practice with legacy
data conversion". Bringing legacy data up to "best practice" standards is
something that each institution will have to grapple with and address as
time/money/other driving needs allow or require.
Maybe the EAD application guidelines need to cover conversion of legacy
data separately from the encoding of new finding aids?
> While the <unittitle> is needed at the higher levels of description -
> for series, collections etc, when
> you get down to item level, items do not tend to have titles - if this
> were imposed the <unittitle> at
> item level would have to become the description of the item.
Well, items do indeed have "titles" as far as rules for describing them
go. If your item is described in some other format (MARC, for example),
then perhaps EAD should accommodate your using a pointer to your MARC
record in lieu of a <did> for a <c0x level="item">. But to repeat what I
said above, I think it is a mistake to let the conversion of legacy data
be the primary guide to developing implementation guidelines for EAD.
________ [Bill Landis] ___________________________ [[log in to unmask]]
Graduate School of Education & Information Studies
University of California, Los Angeles