Thomas,

Thanks. That was a great description of why I tend to think of authorities as about a person, place, event, etc. Whether an authority structure started out with a strong effort to record information about the thing it was describing or evolved to include it, there is no denying that most do at this point in our history.

I, too, am not advocating for catalogers to create triples for all knowledge in the universe. Yet, if we model our data in a way that allows us to create entities that the rest of the word can understand and link to, we will be able to harvest/link to data they generate. Together (with some creativity, strong semantics, and more open data pipelines) we can fill in the gaps.

Sincerely,
Steven

Forgive any typos, sent while on the run.

> On Aug 26, 2015, at 6:18 PM, Thomas Berger <[log in to unmask]> wrote:
>
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
>> Am 26.08.2015 um 22:48 schrieb Karen Coyle:
>>
>> What I'm getting at is that the reason that we struggle with tradition
> al
>> authorities in RDF is that they do not have the qualities that an RDF
>> graph about a person might have. We might use the existing name
>> authorities record identifiers as URIs, but the thing that is today an
>> authority record will not be adequate as a graph describing a person o
> r
>> corporate body. The other option is to leave name authorities alone an
> d
>> create an entire new set of identifiers (parallel, perhaps) that are
>> representative of the entity, not just of the label.
>
> I might humbly point out that RDA mandates the collection of /data/
> for identification purposes and leaves the crafting of individualized
> headings merely as an option. DACS (Describing Archives - A Content
> Standard) has a chapter on "/describing/ creators" (emphasis mine).
> The archivist's vision of "authority" is communicated by EAC - Encoded
> Archival Context without "authority" even in its name, i.e. providing
> rich biographical etc. descriptions tailored to the actual corpus
> (admittedly, establishing identity and rudimentary normalization of
> name forms is only a tiny albeit important portion of the functions
> of these records)
>
> So I am confident that currently most authority files are under way
> of being transformed from collections of headings and variant
> headings into something more easily recognizable as a databases of
> core identifiying information. And of course there are authority
> files besides LCAuth which - possibly for a bunch of different
> reasons - adopted a more data-centric view on their contents some
> decades earlier.
>
> - From the beginning on VIAF has (among other things) been relying on
> the ability to extract /data/ from authority records - comparison
> of headings can bring you only so far and birth and death dates
> as such (and increasingly with granularity to the day, not the
> year) are essential to the process, especially if there is no
> bibliographical information to back the process, e.g. when trying
> to bridge the gap to ULAN or Wikipedia/Wikidata and other datasets
> with only loose ties to library-land.
>
> 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). So as information resources to be consulted by
> humans for identification purposes these records are absolutely up
> to most needs, they are just lacking data-ness. As do biographic
> encyclopedias by the way. Other information resources on the
> web may be richer in data (and poorer in information at the same
> time) and can be used as providers of complimentary data.
>
> In a sense, the actual contents of an authority record don't even
> matter that much, the more connections to other resources are
> established, the more the identity is determined by these
> relations and not by information recorded *in* the record. Any
> single data element might be wrong or contrived and still the
> identity holds true. Think of pseudepigraphic authors of the
> Late Antiquity - not even their name is known and yet they
> have a very distinct identity! Eventually not only a name
> like "Pseudo Galenus" but also an identifier like no2011008661
> by itself serves as a cipher and the act of spelling it out
> already evoces (or at least encompasses) everything that
> can possibly be known, thought or felt about that entity.
> (Semantically. I'm not advocating to encode the complete
> information of our universe into triples and pour a substantial
> part of them out on the unfortunate soul chanting the magic
> no2011008661)
>
> viele Gruesse
> Thomas Berger
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1
> Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/
>
> iJwEAQECAAYFAlXeOw0ACgkQYhMlmJ6W47OTJwP/TQimrys1JKCtLEfoVvrfl7rE
> jkFKLHO7WKSz3c5300/DoLHKvio5uaFzdXPdymdbs+H3aKgg59i1BfGYnpKIyJMk
> gBtQowCSCpbh13bbmL03TwMEjU3u31r3j1pBeN2QdM+RXSOBX7vGFP9sxnvZx4SF
> CRmTztRfQfw/+XZ8Odk=
> =Dr+F
> -----END PGP SIGNATURE-----