Bruce D'Arcus wrote:

> On Feb 20, 2005, at 12:37 PM, Karen Coyle wrote:
>
>>> This shows exactly the problem I would like MADS (and I guess later
>>> MODS) to solve :-)
>>
>>
>> This is exactly what MADS does solve.
>
>
> Not exactly; at least not the last I looked.
>
> First, I should clarify and say that I would not put the two forms of
> Boyle in a MODS record; I'd only ever put in the form that allowed me
> to get the correct formatted output, whatever that may be.

Then that's a different problem to the one that MADS solves. MADS is
about the authoritative form, not formatting the output. So I think
we're talking about different things here.

>
> As I explained earlier, the issue is really names in different scripts
> (or perhaps names and pseudonyms).

Those are two different things. Very different. And not the whole range
of issues, either. There are also variant forms of names (as used on
different publications or in different databases). And for some persons
and many corporate bodies there are changes in name over time. Each of
these may need a different solution.

>
> With respect to my point about MADS and solving my bigger problems:
>
> There's no way to encode names in MADS to be able reliably format a
> name like J. Edgar Hoover; where you can easily handle not just one
> representation (the full name), but also sort-order stuff like:
>
>        Hoover, J. Edgar.
>        Hoover, J.E.
>        Hoover, J. E.
>        etc.
>
> Likewise, there's no way to unambiguously code a name like T. C. Boyle.
>
> The reason is simple: MODS/MADS only allow distinctions between family
> and given names.  There no explicit coding that allows you distinguish
> between "J" (a given initial that is of secondary importance) from
> "Edgar" (a primary given name).

An initial is not necessarily of secondary importance. You are making
assumptions about names and their formatting that may not be universal.
And I think you are asking for markup for something that can easily be
done algorithmically and may not need markup. I know that we did this
algorithmically in the first version of the UC online catalog and it was
just part of the coding that we did for indexing and display. (And it
was based on the MARC forms of the names which have less markup than the
MODS forms do.)  It sounds like you want to include these derived forms
in the record itself, which is not what I'm accustomed to in any systems
I've been involved with.

But this is also a very different task to determining that the
authoritative form of the name is J Edgar Hoover and not John E Hoover.
That is something that you have to decide based on knowledge about the
person, their publishing history, changes that they've gone through,
etc. So I think that you've misunderstood the rationale behind the
concept of "authorities" as we mean it in libraries. It's not a matter
of listing the variant ways that you could index or sort a name, or even
ways that you could display it. We've always left that to the systems
that process the data, and those derived forms are not stored in either
the bibliographic nor the authority record.

kc

>
> But I suppose we're getting off-topic.  I'd be happy if we could settle
> on a way to address my issue of multiple variants (and titles; I'm
> running into refer journal article records from online databases that
> try to fake this sort of markup by adding stuff like <i>some
> title</i>).
>
> Bruce
>


--
-----------------------------------
Karen Coyle / Digital Library Consultant
[log in to unmask] http://www.kcoyle.net
ph.: 510-540-7596
fx.: 510-848-3913
mo.: 510-435-8234
------------------------------------