Here at the Bodleian we've been trying out the EAD on a couple of finding aids - here
are some comments based on our findings so far:-
1) I think a reg attribute (rather like the TEI equivalent) is necessary for <persname>,
<geogname> and <famname> to allow the construction of name indexes using
a name authority file.
2) I think an attribute (similar to the general n attribute in TEI) would be useful for
the component elements to allow shelfmarks etc for individual items to be recorded,
more elegantly than is possible using the <did> and then <unitid> elements.
3) I don't think the current way of constructing indexes is at all useful - I think
the facility to generate indexes automatically is vital, and the tagging of a
hand-generated index is going to be laborious and less than useful.
4) On the whole I think the DTD tends to take the look of a printed finding aid
too much as a model - hence the extensive table formatting codes,etc. I think the
problem with taking this approach is that we could end up with something like
HTML, ie great for formatting but less useful for expressing descriptive content.
We are using the <did> element within components, rather than the table options,
and this seems the most useful approach so far.
5) Also I would suggest using attributes rather more, for the sake of greater
concision and readability.
Despite these minor quibbles, a tremendously useful tool - we're certainly going
to be using it for all our finding aids in future.
Bodleian Library, Oxford