I'm not sure that enough tagging has yet been done to be able to refer to
"acceptable practice," but if you want yet another approach--the use of
recursive, unnumbered <c>s, look at some of the longer finding aids at the
Harvard site (, particularly those from the
Law School and Schlesinger.

I've found the use of _numbered_ <c>s helpful in doing the actual tagging,
because the way some of the older Houghton finding aids were done there's
a lot of nesting, and having numbered tags makes it easier to keep track
of where I am.  However, I then go through and "search and replace" the
numbered <c>s with "plain" <c>s, when I put it up on the web site.

Our reason for choosing to use unnumbered <c>s was simple: we are
attempting to present all Harvard finding aids in a uniform way, and found
that a <c03> in one finding aid might be a series, and in another a <c03>
might be a sub-series.  Since we do not yet have much experience in how
the EAD DTD is handled by a variety of style sheets and/or browsers, we
were concerned that having <c03> "meaning" differing levels in different
finding aids might cause us problems with display in the long term.  Using
<c> alone seemed more prudent, and using the nesting of the <c>s to
indicate levels of hierarchy seems to accomplish the same thing.

The practice of using <drow> and <dentry> does seem to be the prevalent
approach, but I agree with those who feel that such practice focusses on
the traditional paper layout of information, rather than utilizing the
strength of SGML, which is to mark up _content_.

Leslie A. Morris
Curator of Manuscripts in the Harvard College Library
Houghton Library, Harvard University
Cambridge, MA 02138

e-mail: [log in to unmask]
phone:  617.495.2449
fax:    617.495.1376