In response to several ideas:

   I cannot reinforce too strongly Kris' statement that EAD is markup of
the intellectual and not physical arrangement of archives.  The Stanford
example violates not only the ARUP of their cooperative project but more
generally the spirit of EAD.  As evidence, I note that the type attribute
for <c> includes only terms denoting intellectual or structural
relationships and not physical ones like box and folder.  Those are found
under <unitloc>.  Richard Higgins' second example clearly reflects the
"proper" approach.

    As to tables and drows, Bill Landis has correctly identified the
issues to my mind, ones that we need to stretch and bend and rip (is
that what you said Steve?  Awfully violent for archivists).  In terms
of creating styles for display (yes, we do have to think about such
issues- SGML software cannot imagine what we have in mind for presentation
if we don't tell it and unfortunately style sheets need some hooks to work
on).   The question is this.  Within a single complex finding aid or
across many finding aids in one or many institutions, will the display
of a given element, and particularly components, always/sometimes/never be
absolute (depending completely on its intellectual designation as a
series, fond, sub-collection, etc.) or relative to its parent irrespective
of the level of the parent?   I think that we haven't had enough
experience to answer that collectively and definitively.  We may
never.  And I'm not entirely sanguine that we could ever get agreement on
the level of a particular component my colleagues here at MHS let alone
across multiple institutions.  My initial personal reaction is that it may
not matter since designations like series, meaning nothing to our users,
are unlikely to the basis for any searching mechanism.  But then I've been
wrong before.

   RE drows, we have found that it is possible to handle simple columnar
displays using the Panorama stylesheet alone, without resort to layering
the drow structure over the components.  At least if you have what I
suspect to be very common layout like this- three columns for box, folder
and contents, reading from left to right where the contents column might
wrap at the end of a line but only back to the beginning of its column.
Thusly-
     Box        Folder     Contents
      1          2         Correspondence, 1957-1962.  12 folders


Michael Fox

Processing Department
Minnesota Historical Society
[log in to unmask] or [log in to unmask]
fax  612.296.9961