On Sep 7, 2004, at 12:26 PM, Karen Coyle wrote:

> On Mon, 2004-09-06 at 13:52, Bruce D'Arcus wrote:
>
>> To leave it to authority records to solve the problem is not an
>> adequate solution, particularly if you're trying maintain
>> compatibility
>> across MODS and MADS.
>
> In fact, in libraries it is the authority records that solve this
> problem. So this may be a matter of differences in rules. The rule in
> libraries is that each author gets one and only one "name." The name is
> essentially a unique identifier for the author; as each new name is
> entered into the catalog, the authority file is used to determine if
> the
> name is unique (within that catalog, but since we all share now it
> needs
> to be unique within the whole community that share bib data). If there
> already is another instance of that name for a different author, then
> work is done to make the new name unique (adding date of birth, or some
> other info to it). This is why some names in library catalogs seem so
> ... unintuitive.
>
> The authority record keeps all of the other forms of the name, and
> those
> names are included in the library catalog as references to the unique
> form of the name (called "authoritative").

Yes, and I think this is a good thing.

I do believe there's a basic logical problem here, though, which is
also expressed in the MADS draft schema:

A name not a contributor to a work.  It is an identifier for a person
or organization who is that contributor, in the same way that a title
is an identifier for a work.  A person can have one or more names,
which is the problem in question.

In citation practice, the concrete issue here is that if you work in
multiple languages, you typically have to include both the original and
the translated/transliterated titles and author names on the final
output.  The example my friend gave for typical output using the
original script and transliterated titles and names is:

        Tanaka Ichiro KANJI. 1988. _Zengaku nyumon_ KANJI.

I can do this really easily if MODS allows for such coding of names,
but it doesn't.  I'm prone to just say I can't support it, because
while I would like to support MADS as an option, I do not want to
depend on it.

I have a hunch that the current name handling will introduce other
problems, though I can't put my finger on exactly what.

> You are preferring to allow equivalent forms of the name in a
> bibliographic record. I don't know if your case is only for
> transliterated vs. vernacular fields. If so, this is similar to what
> MARC does with its non-roman script fields, which are a special case.
> Those fields were not carried over to MODS, however.
>
> I don't know enough about the languages in question, but could the
> transliteration be an attribute on the namePart? i.e.:
>
>     <name type="personal">
>        <namePart type="given" trans="Ichiro"> $B0lO) (B</namePart>
>        <namePart type="family" trans="Suzuki"> $BNkLZ (B</namePart>
>       <role>
>          <roleTerm type="text">author</roleTerm>
>        </role>
>     </name>
>
> That would make it clear that they are the same name. And I would think
> that the vernacular would be the authoritative form of the name, since
> transliterations often lose some information.

That may be an option.  I'll see if I can get some feedback (I don't
myself work with multiple languages).

One problem is that it then introduces an inconsistency between
mods:name and mods:titleInfo.  If one allows more than one titleInfo
element, I see no logical reason why not to allow more than one name.

Bruce