background image
suffice to refer to them simply by a quote and restrict oneself to short-title information for the
sake of retrieval. Since emblem books themselves are well catalogued, it seems more appropriate
to place the emphasis on the description of the individual emblem itself. Differences in national
cataloguing rules, usages, and habits will not endanger data consistency. In the German union cat-
alogue Gemeinsamer Bibliotheksverbund (GBV),
11
both Library of Congress and German data
records enjoy a peaceful coexistence.There is, however, one decisive exception: authority-files are
treated differently. German rules prescribe writing names according to the original spelling, LOC
names are sometimes English translations of the original names.This leads to the question of how
to deal with names. It would be highly recommendable to have unified access to names across
the various resources. A concordance might help to achieve this to some extent, but in view of
reduced budgets it is unlikely that any one project will undertake the arduous task of developing
name concordances.Therefore, it seems advisable to rely on national authority files as much as
possible, in the future there may be mechanisms which are capable of mapping local to foreign
name records provided by the national libraries and other central institutions responsible for
authority files.
Another issue is the language code. Rawles suggested using language codes according to ISO
639-2, which is the alpha-3 code published in 1998 having 3 characters.The HAB used ISO 639-
1, that is, alpha-2 code.
12
According to ISO 639-2 German is ger, but de according to 639-1.
Although alpha-3 is a more recent development, some features speak against it. First of all the
XML schema offers a special language data type called xs:language, which accepts every language
code that is conformant to RFC 1766. This recommendation is based on alpha-2. Country infor-
mation may be added in capitals, e.g., for US-American English: en-US. ger is an invalid code for
this specific data type.These aspects therefore speak against the alpha-3 code as it is not compat-
ible with the broadly applied XML schema data type xs:language. Moreover, all languages in
emblem books are covered by alpha-2 encoding.
The first approach of the HAB emblem project was to draw up an emblem DTD to offer a
proposal for a standardized structure, but as OAI requires an XML schema definition we have
now converted the DTD to an XML schema.
14
The completed XML schema, which integrate s
the outcomes of this working conference, is obtainable from the HAB project webpage at
http://www.hab.de/bibliothek/wdb/emblematica/emblem.xsd. The advantage of having an
XML schema is that contrary to a DTD, an XML schema definition allows for control not only
of structure, but also of content. By parsing the XML file one can check whether documents
comply with particular rules. For instance, if there is a rule that an emblem document must have
a transcribed motto, the validation fails when the motto is missing. If a certain sequence of num-
bers and characters is expected, making up an Iconclass notation, or if a particular language code
is required, the validation will fail, if the code does not comply with the rule.The schema also
permits validating documents before they are integrated into a database, which is important for
harvesting distributed Web resources in order to ensure data consistency.The development of the
first draft of our XML schema relied on Rawles' first list of data fields.The language code, how-
ever, is now modified, and some elements and attributes, as well as indications as to whether the
data field, or rather XML element, should be mandatory or optional have been added.
Employing OAI in connection with a metadata set especially designed for the description of
emblems will provide a powerful means to exchange emblem data and build up joint databases.
It will facilitate the research on emblems tremendously.
11
http://www.gbv.de/.
12
ISO 639-2 with corresponding
ISO 639-1 code available at: http://
lcweb.loc.gov/standards/iso639-
2/langhome.html.
13
http://www.faqs.org/rfcs/
rfc1766.html; RFC stands for
"Request for Comments."
14
http://www.w3c.org/XML/
Schema.
95
DC_Emblemsbook_180204 19.02.2004 11:26 Uhr Seite 95