> *With XML files* when using <eadid type=> with the fully declared SGML
> Open type of identifier -
> - do I keep the type="SGML Catalog" - instead of changing it to something
> like type="XML Catalog"?
> Here's an example:
> <eadid type="SGML catalog">PUBLIC "-//Yale University::Beinecke Rare Book
> and Manuscript Library//TEXT(US::CtY-BR::YCALMSS134:Glenway Wescott
> Papers)//EN" "Wescott.xml"</eadid>
> Can anyone confirm this?
Crumbs, this isn't really an answer to the question, and it
is going to sound like awful pedantry (but my friends are
used to it)....
Firstly, I think that in certain circles "XML Catalogs" and
"SGML Catalogs" are perceived as different beasts, i.e.
two different syntaxes to provide the same functionality:
John Cowan's XML Catalog syntax
OASIS Technical Resolution TR9401 (previously SGML
Leaving this aside, I would argue that your example (and I
see a similar one in the eadid entry in the EAD Tag
Library) actually has as the content of the eadid element
something which, strictly speaking, is not an "identifier",
but rather a catalog entry conforming to OASIS Technical
That catalog entry itself contains _two_ distinct identifiers
of different types: a Formal Public Identifier and a
I think this raises many interesting issues. Judging by
recent discussions on XML-DEV (not all of which I fully
understood!), there still seems to be a degree of
uncertainty as to the role and value of FPIs in XML.
In any case, I think it is true to say that many XML-based
tools have not provided means of resolving public
identifiers, though many XML parsers do provide the
"hooks" to provide the code to do so, and Norman Walsh
of ArborText recently made available some Java classes
to provide this functionality, which I have experimented
with to a very limited extent.
I wonder whether what EAD really requires is to allow the
eadid element to be repeatable (which is not permitted by
the current DTD) so that your example would be
transformed into something like:
<eadid type="FPI">-//Yale University::Beinecke Rare
Book and Manuscript Library//TEXT(US::CtY
(I guess the same effect could be achieved by allowing a
new repeatable sub-element within eadid)
I can imagine that in the future there may be a
requirement to extend that list of identifiers of different
types. Each identifier remains a unique instance of its
type but the same object can have multiple identifiers.
Given this possibility that an EAD document may acquire
new identifiers of new types at some time (perhaps a long
time) after its creation, I do wonder whether perhaps this
information is best recorded in some form of metadata
"package" which is maintained quite separately from the
EAD document itself.
More questions than answers, really....
Glasgow University Archives & Business Records Centre