There are several concepts to be explored here- the notion that the
files we have created (as housing units) do not correspond to the
original housing units.

There is also a related problem that I want to interject though Jackie
doesn't, namely the concept of filing units.  There are often layers
between the series or subseries and the file.  In organizational
records, these may have been chronological or topical headers used to
organize files.  When arranging units of personal papers, we archivists
often create such arbitrary headings as a way of creating organizational
subunits that really have no exact warrant but are useful for grouping
materials into more manageable chunks.  We have no terms to describe
these intermediary levels.

   As a solution, one can always assign the "otherlevel" value to the
Level attribute and define a suitable local value for the Otherlevel
attribute.  But it would be nice to add some terms to the "closed list"
for Level.  Jackie suggests the word "portfolio."    "Original file
unit" might also apply here.  For the other type of level I might
suggest "file organization unit" or some such mouthful.

Michael

>----------
>From:  Jackie Dooley[SMTP:[log in to unmask]]
>Sent:  Tuesday, September 16, 1997 9:30 PM
>To:    Multiple recipients of list EAD
>Subject:       "file" attribute
>
>Is there a generally agreed-upon definition of the level attribute "file"?
>We're trying to decide whether it would be appropriate in a circumstance
>in which we want to nest multiple folder titles under a broader
><unittitle> indicating the the creator's original file units, e.g.
>(example includes only partial tagging, obviously):
>
> <c01 level=file>Portfolio 1: its title.</c01>
> <c02>First item from portfolio 1.</c02>
> <c02>Second item from portfolio 1.</c02>
> <c03>Third ...
>
> <c01 level=file>Portfolio 2: its title.
> <c02>First ...
>
>The portfolios provide the only hierarchy in the entire container list
>(ah, legacy data ...), so we would like to get the portfolio titles into
>our table of contents by assigning a level, and the "file" attribute seems
>like the only reasonable alternative other than "otherlevel".  All
>comments and criticisms will be appreciated!
>
>Thanks--
>
>  Jackie M. Dooley, Head of Special Collections and University Archives
>UCI Libraries, P.O. Box 19557, Univ. of California, Irvine, CA 92623-9557
>   Internet: [log in to unmask]  Phone: 714/824-4935  Fax: 714/824-2472
>