Terry Allen writes:
>AUTHOR is only a label for a definition. (And on the issue of how to
>define an element, I much prefer definition by prose to definition
>by example.)
We definitely need prose definitions. But the definitions should not just
prescribe usage; they should also describe it. The Dublin Core should be
a little bit like a computer language, where its designers define terms
in a declarative fashion, but also something like a human language (or
pidgin), where meanings evolve through usage. Human language dictionaries
can have a more descriptive or prescriptive flavor, but in the good
ones, the definitions evolve to capture the trends their editors see
in the clippings their correspondants submit. And we shouldn't believe
that the DC definitions can be fixed once by decree and in English and
then frozen for the next fifty years, or even that this is desirable.
The definition and examples should evolve in tandem over time, where
examples will be the more dynamic part as the DC gets deployed in
different languages and fields. And it will be easier to manage a
global, distributed discussion if we focus in on concrete examples
("Does this example here belong in this element?") than if we haggle
legalistically over the wording of definitions.
>That is, if AUTHOR is #33, #33 still has to be defined (e.g., in English).
>One can imagine a regime in which the English definition is normative,
>and translations are only representations of it in other languages;
>if there is some dispute over what a translation means, you examine
>the English definition.
I can imagine translating English definitions as the first step in
creating, say, a Japanese Dublin Core. But then Japanese need to
edit them into concepts that make good practical sense in Japanese.
One wouldn't compile a Japanese dictionary by translating the words,
definitions, and examples from an English one.
>Or one can imagine a regime in which every translation of the English
>definition is normative and must be carefully coordinated with every
>other translation (as in diplomatic documents).
This is exactly what I think we must avoid, for both political and
linguistic reasons.
Tom
_______________________________________________________________________________
Dr. Thomas Baker [log in to unmask]
GMD -- Forschungszentrum Informationstechnik GmbH Tel. +49-2241-14-2171
Schloss Birlinghoven Fax. +49-2241-14-2071
53754 Sankt Augustin, Germany Secr. +49-2241-14-2911
GMD - German National Research Center for Information Technology
ERCIM - European Research Consortium for Informatics and Mathematics
|