-----BEGIN PGP SIGNED MESSAGE-----
Am 27.08.2015 um 03:26 schrieb Karen Coyle:
> On 8/26/15 3:17 PM, Thomas Berger wrote:
>> LC records may be lacking explicit (especially death) dates, but
>> they usually contain a wealth of textual references which more often
>> than not reveal exact birth dates and other important contextual
>> information (affiliations, notable works, places of birth, death
>> and activity).
> You say "usually," but I see "occasionally." As an example, here's the
> LC authority record for me:
> http://id.loc.gov/authorities/names/n89613425.marcxml.xml, and if you
> can find "other important contextual information" you are a better
> reader than I am.
Sure I do. I "found: FRBR, before and after, 2016: ECIP t.p. (Karen
Coyle) data view (Karen Coyle is a librarian with over 30 years'
experience with library technology)" (probably you weren't aware the
record was changed last week and human readers should prefer the
text view < http://id.loc.gov/authorities/names/n89613425 > to grok
in the information more easily.
- From my personal experience I can tell that this exactly is the kind
of information needed when e.g. having to decide whether this person
could be identical with some E. K. Coyle, b. 1977 with a speciality in
constitutional law. Much of it works over exclusion and having data
(as opposed to text snippets as above) definitely will assist with
preliminary filtering. In case of exact matching data sometimes
doubts arise whether the records you're looking at already confuse
different persons. Especially beware of other authority records
citing LCauth records as source - they may have gotten it completely
> Plus, when that information is available it's in a
> note field and can be quite cryptic:
> "found:Phone call to Curbstone Press, 07-13-94(Michael H. Cooper is a
> pseudonym of Michael Clark; b. 6-11-51; res. in Alaska)"
> "found:Message from J. Baker, 11/08/88(Michael D. Cooper; b. 10/30/41)
Fortunately not in the same record...
> Thomas, I think you are being overly optimistic about the state of
> authority data today. I agree that it can get better, but I don't agre
> that VIAF extracts much "data" -- it basically gets what you can get
> from a MARC name heading field, and adds in titles of works (which it
> gets from the WorldCat database). There may be more information in
> non-US authority files, but rather than relying on impressions it woul
> be better to have a good study of available data.
Maybe I had been a bit sloppy but my intention was not talking about
authority data but on information contained in authority records. From
personal experience, and statements made by those involved with VIAF
(like Thom Hickeys blog < http://outgoing.typepad.com/ >) I've gathered
the firm conviction that in the Cooper example
VIAF /does/ extract exact to the day birth dates from the variant
heading "Clark, Michael, 1951 June 11-" in order to perform tie-breaking
when needed. I mean, VIAF is all about data mining applied to library
generated information and unfolds its use exactly in situations where
suitable categorized data elements are lacking. In a world where
authority records are in a better shape they would almost immediately
find to each other based on the data they contain and we perhaps
wouldn't even need something like VIAF...
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1
Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/
-----END PGP SIGNATURE-----