Janice,
Thanks for the tips. They answer my questions exactly. In
the Harriman fragment I assumed level="series" represented
the top level series and level="subseries" was a more generic
term representing any level of the series hierarchy below
the top-most level, thus sub-subseries and sub-sub-subseries
were all really types of "subseries". I had mistakenly tagged
the Harriman as:
<c01 type="series">...Special Files
<c02 type="subseries">...Truman Administration
<c03 type="subseries">...Marshall Plan
<c04 type="subseries">...Subject File
I see now that I must omit type attributes for <c03> and
<c04>, e.g.,
<c01 type="series">...Special Files
<c02 type="subseries">...Truman Administration
<c03>...Marshall Plan
<c04>...Subject File
Alvin Pollock
Electronic Text Unit
UC Berkeley Library
[log in to unmask]
>Alvin:
>
> Thanks for your questions. Some of my colleagues at LC have
>begun to tag a few of the Manuscript Division finding aids, so we
>are also grappling with various application issues. In response
>to your questions, I offer the following suggestions:
>
>(1) The use of the level attribute is optional. The Bentley
> group thought that it would be useful to indicate certain
> hierarchical levels within a collection, especially the
> series and subseries levels. We also thought that an "item"
> level might be used in instances where the finding aid
> describes information about an individual photograph,
> manuscript, or other document. Although the container list
> in a finding aid may contain successively finer levels of
> description about the collection, each of those levels does
> not necessarily correlate to a meaningful archival term that
> would be useful for administrative or reference purposes.
> Consequently, I see little value in supplying "level"
> attributes for what would probably be arbitrary
> designations, and I would advise you not to assign level
> attributes to the <c04> and <c05> components in the example
> you have given.
>
> In reviewing your example, it occurred to me that you may
> have mistagged the series and subseries levels. I can
> appreciate how this could have happened, as there is much
> about the Harriman finding aid that is unusual. The series
> should be "Special Files: Public Service" and the subseries
> should be "Truman Administration." "Marshall Plan," if it
> was identified as anything, would be a sub-subseries, and
> "Subject File" would be a sub-sub-subseries. The fact that
> we described the "Marshall Plan" and "Subject File" in the
> Description of Series undoubtedly suggested to you that they
> were the series and subseries, but if you tag them as such,
> you omit the two levels above them.
>
> The <c06> components, for which you have assigned "item"
> level attributes, are not "items" as defined by the Bentley
> group. I would probably not assign them a level attribute
> at all, but if I did, the choice would be "file." I would
> reserve the use of "item" for a single document or
> manuscript within a given file or folder.
>
>(2) You raise a good point about the use of <unitdate> and
> <unittitle>. I suppose I (and the other Bentley members)
> originally assumed that <unitdate> would follow <unittitle>,
> but I think I now prefer to enclose <unitdate> within
> <unittitle> for the simple reason that some repositories may
> choose not to tag each date separately. If you don't wish
> to tag the date, the date would either float untagged inside
> the <c> or could be incorporated within the <unittitle>.
> The latter seems cleaner and more reflective of the fact
> that the date *is* indeed part of the title.
>
> Throughout the DTD, we tried to provide for a range of
> tagging choices--minimal, moderate, extensive, etc. The
> tags were constructed (hopefully) to allow the user to
> select first a broad tag and then, as desired, refine the
> contents within the broad tag by tagging at progressively
> more granular levels. By enclosing the date within the
> title, the repository can elect to add (either initially or
> at a later time) separate <unitdate> tags to enhance
> searching, retrieval, or display of dates in a finding aid.
>
>
> Keep us posted on your progress with the Harriman guide.
>Mary Lacy, of the Manuscript Division, is pulling together a few
>questions which have arisen during our initial tagging efforts,
>and she will be posting them shortly. I also hope to have a
>chance soon to look more carefully at the other examples you made
>available on the Berkeley site. I was surprised at your choice
>of tags in a few instances, and I think it would be fruitful to
>get a dialog started among all the early implementors regarding
>tag selection.
>
>Janice
>
>
>Janice E. Ruth
>Manuscript Division
>Library of Congress
>
>
|