I have always found a fundamental problem with EAD at item level, which
I have raised on a few occasions, and which I think may have partly led
to the <abstract> tag coming about, although it did not do so in a way
that resolved my problem. As Michael said recently, the problem is not
in EAD but in the structure of our finding aids, but we have enough
unlisted material here to make revising a few hundred years worth of
finding aids an unlikely event, so a pragmatic approach to retrospective
conversion was all I could aspire to.

The problem with our finding aids is that at item level the
scope/content is the largest piece of description, and is normally
located amidst the elements available in the <did>, viz.
<unitid><unitdate><scopecontent><physdesc>. With calendars, this becomes
even more complex and can require many other elements embedded within
this structure. While it would have been possible to just move the
scope/content to the end of each item description, this would have been
yet another task and would have required re-writing of finding aids and
abandoning a perfectly sensible structure (as you may have noticed by
their absence from discussions on the subject, there are no UK content
description rules, and certainly there were none used on our lists).

To start with, I used a <note encodinganalog="520b"> element, which
worked fine internally, but meant that no other system could locate our
scope/content. As I was also using the tabular version of EAD for
another series of finding aids, I looked more closely at this option.
This gives you an alternative in that you can use the <did> or non <did>
structure, and the method I am now working on uses the tabular DTD
settings. At collection, series etc. level, use the <did> and at item
level use <drow><dentry>. All of the elements required are available
within <dentry> in any order, giving the flexibility required for item
level listing and retrospective conversion work. It abuses the
<drow><dentry> tags, but these are trivial as their only purpose is text
layout. If you set no attributes in these tags, and do not provide
<*spec> elements no formatting system will have any information on how
to process these elements, so it is less likely to fall foul of external
retrieval incompatabilities.

From what I remember, the point of the <did> was to provide an area
marked as containing the most important information about the
collection, so that this could be retrieved from the morass of our
finding aids. This aim would probably assume that item level information
is not relevant, so a system of placing series and collection summaries
in <did>s but not item level information would seem reasonable. I also
suspect that the majority of EAD finding aids do not contain item
information anyway.

--
# Richard Higgins
# Durham University Library
# Archives & Special Collections
# Palace Green
# Durham
# DH1 3RN
# E-Mail: [log in to unmask]