All,
This is a comment for dc:title as literal.
The range of dc:title should not be "literal" if it means
a "sequence of characters".
I would like to point that a literal with language identification
is not just a "sequence of characters". Multiple script
representation of title (or any text) in Japanese is an example.
Even in English, "DUBLIN CORE METADATA INITIATIVE" is same name
as "Dublin Core Metadata Initiative", though they are not identical
as a "sequence of characters".
One may argue that well established tradition of indexing and
software techniques can cover this case issues. Yes, but, it
is a hard-coding of language dependent (implicit) knowledge.
I think our goal is to describe the knowledge explicitly.
What most people think as a literal is, in most cases, a text
in some language. In this sense, dc:title may a literal.
But, it is no just a "sequence of characters".
(It may be a subclass of "sequence of characters", though.)
I know my argument is not well organized. But I was contemplating
this, since this spring and still in this stage...
Best wishes,
Akira MIYAZAWA
National Institute of Informatics
2-1-2 Hitotsubashi, Chiyoda-ku, Tokyo 101-8430
Japan
Phone: +81-3-4212-2508
Fax: +81-3-3556-1916
email: [log in to unmask]
|