Diane,
You're pointing to distinctions that I think are good to make for names
of responsible parties and other data elements as well. After all, most
records are going to be first visible as part of a "brief display" in a
hit list or something. And bibliographies are important. There are a
lot of context where, for space reasons or whatever, the computer needs
to "pick one" or "pick some".
Before machine processing that re-sorted records by tag number came
along, the order in which notes and subjects appeared in a record was
considered useful and we followed a pattern there. Now, the order of
what displays (and what appears in that critical "above the fold" part
of the screen in a record display ) is often less controlled. We can't
rely on position to give us information because things move.
In MODS we have now a "primary display" usage attribute for the
<location><url> element. A selection like that for name, title,
subject, location and maybe other elements makes sense. Maybe even two
levels of "designating precedence" indicating the strength or importance
of the relationship of that entity to the work or whatever we're
describing, with the highest level assigned to one element, would be a
useful option. Could probably help relevance ranking as well.
Laura
--
Laura Akerman, Technology and Metadata Librarian
Robert W. Woodruff Library, Room 128
Emory University
Atlanta, Ga. 30322
ph: (404) 727-6888 / email: liblna_at_emory.edu
Diane I. Hillmann wrote:
> Suzanne:
>
> I think part of the problem is that the distinctions between a 100
> and a 700 made sense when what we were doing was creating catalog
> cards, but those distinctions are probably no longer the ones we want
> or need to make. We want to know which are authors, which are
> illustrators, which are editors, etc. We may want to distinguish the
> first author from the other 5 (or 50) for citation purposes but
> different tags may not be necessary to do that. If we want computers
> to make the distinctions that we and our users want, we have to give
> them the data upon which to act. Encoding relator information is
> probably a good start, but determining a finer level of
> responsibility may be more difficult (and may indeed be more than our
> users want and more than we can afford to provide).
>
> There's a lot to sort out, no doubt about it.
>
> Diane
>
>
>> And the comment on one of these threads mentions keeping the user in
>> mind
>> not the rules.
>>
>> More and more my user is another computer. Making my work needing more
>> granularity for a computer to know what the data means.
>>
>> Is the main entry used for display mainly or is there something else
>> that
>> the main entry provides? Does it help identify a work? If the
>> computer find
>> the name string in the 100 vs in the 700 is there something the program
>> "knows" about that name string? I guess I can tell a computer would
>> appreciate the name string being in a 600 vs a 100 or 700. But with
>> out, say
>> the relator code with the name in the 700, is the computer have some
>> knowledge that that name is "less" responsible? And would that even be
>> correct!? No. 100 and 700 can both be essentially responsible.
>>
>>
>> On 7/11/07, Sharon Foster <fostersm1_at_gmail.com> wrote:
>>
>>>
>>> And unfortunately cataloging is still being taught like this. My
>>> instructor, who I consider one of the better ones at my school, kept
>>> referring to "main entry" and "added entry," and how long it takes to
>>> search 10,000 records. But, said I, we're not really searching each
>>> record, are we? We're really looking up in an index of all the
>>> records. And what is the significance of a "main entry," when any
>>> field may be defined as an access point in the database? It seems to
>>> me that we are retaining an awful lot of concepts and terminology that
>>> meant something in the card catalog days, but are somewhat less than
>>> useful now.
>>>
>>> On 7/11/07, Suzanne Pilsk <suzanne.pilsk_at_gmail.com> wrote:
>>>
>>>> I'm always flummoxed when it comes to the decision : is the
>>>> information
>>>
>>> I'm
>>>
>>>> putting in just detail vs is it actually providing access.
>>>>
>>>> It comes out when something "interesting" is wanting to be done
>>>> with the
>>>> data and it turns out that details were skipped because they weren't
>>>
>>> seen as
>>>
>>>> important access points.
>>>>
>>>> Anyone who tries to repurpose metadata sees the inconsistencies
>>>> and can
>>>
>>> get
>>>
>>>> very frustrated when trying to pull out or pull together data
>>>
>>> differently
>>>
>>>> than, say, the established index of an ILS.
>>>>
>>>> Case: Whining recently about old old MARC records that had very
>>>> sketch
>>>> titles and often didn't include authors in 100 fields because the
>>>> rules
>>>> didn't make you do it. It was good enough then. Now it isn't
>>>>
>>>> Case: Georgraphic grouping of data but it turns out the fixed
>>>> field for
>>>
>>> ctry
>>>
>>>> was skipped because the ILS at the time didn't do anything with it
>>>
>>> anyway.
>>>
>>>> Wasn't an access point - why look up those pesky three letters?
>>>>
>>>> Case: Grouping publishers together with dates to approach for copy
>>>> right
>>>> permissions; whoops, abbreviated publishers get us nowhere. The
>>>
>>> Association
>>>
>>>> in 260 $b ugh! Well, it isn't an "access point" why bother.
>>>>
>>>> Case: Grouping personal names together - without relator codes -
>>>> is the
>>>> person in the 700 an author of the work or some other "role"? A
>>>> detail
>>>
>>> that
>>>
>>>> was not used because access was for the name not the role
>>>>
>>>> Anyway, you all get my point. The problem I have is when is a detail
>>>> actually turn into an access point and when is "good enough" now turn
>>>
>>> out to
>>>
>>>> be useless down the road?
>>>>
>>>> I think what I took away from the meeting at LC was that I need to
>>>> give
>>>
>>> up
>>>
>>>> on the touch the record once and walk away. Instead, the record
>>>> might
>>>
>>> be
>>>
>>>> touched and retouched over and over as more data becomes
>>>> available. But
>>>> that sounds very expensive.
>>>>
>>>> Yet still, I go back to my cubical and I do.... what? Go fast and
>>>> hope
>>>
>>> that
>>>
>>>> what I've decided was good enough is actually okay and that later I
>>>
>>> won't be
>>>
>>>> cursing myself because I can't provide subset of my data by something
>>>
>>> that
>>>
>>>> seems so obvious to that future me?
>>>>
>>>
>>>
>>> --
>>> Sharon M. Foster
>>> [affiliations omitted]
>>>
>>> Any opinions expressed here are entirely my own.
>>
>
>
Received on Fri Jul 13 2007 - 07:13:40 EDT