On 6 February 2012 16:58, Karen Coyle <[email protected]> wrote:
> Quoting Ben Companjen <[email protected]>:
> ).
>> [snip]
>
> Have you discovered the type lists, e.g.:
>  http://openlibrary.org/type/author
>
> http://openlibrary.org/type/ shows the types that exist.
>
Yes, I have. What I meant is that there are no descriptions for any of
these types. I can only guess how these types are used. By the way, it
took me a while to find out this list of types is the data model.
Searching the website I first encountered several tables and schemas
that were outdated.

E.g. /type/edition/distributors - there is no field for distributor(s)
in the edit form, and I don't know about the existence of a field for
distributor in MARC. So why is this field there?

And what is the uris[] field in /type/author for? Is that where VIAF
URIs (will) go?

>>
>> On a side note: I just had a wild idea: don't show the "by statement"
>> field on the edit form, not even in the librarian mode, if it is
>> empty, so that no one is tempted to put anything in it.
>
>
> That's a possibility. It's mainly there because it comes in on the library
> records, not because it's likely that anyone will fill it in.
>
The label for the By statement field in the edit form kind of invited
me to enter something, but I wasn't sure what to put in that field, so
I left it empty.

>
>> OL Edition contributors are
>> just saved and treated as strings (which I find a little
>> dissapointing), so entering a corporate identity there isn't a
>> problem. The RDF template outputs these contributors as foaf:Persons,
>> though.
>
>
> Too bad those are just strings. It would be ideal to have all parties
> identified. Many people will be authors, contributors and subjects.
>
And the roles can be identified by URIs too.

The British Library's Linked Data even has links from Person to
Person-as-subject:
http://consulting.talis.com/wp-content/uploads/2011/07/British-Library-Data-Model1.pdf

If they can do it, a community like OL's can do it, IMO.

>
>>
>> Your proposed field for "responsible organization" could help here.
>> But should it be part of OL's Work or Edition type?
>
>
> I haven't thought that through.
>
In any case: if it were created, the rationale behind this decision
should be made clear.

>>
>> If we (users) want it, and if the data model supports it, I think we
>> can make the data support it, programmatically or manually. I'm from
>> the Discogs world, in which everything is done manually by people
>> mostly smart enough to match imprints (a.k.a. labels) to
>> entities/publishers behind those labels. Even recording, mixing and
>> mastering studios and the companies behind the labels (mentioned as
>> copyright holders) are matched. Guidelines and some rules are needed
>> there, but they seem to work.
>> Sure, Open Library is not Discogs and the world of books is different
>> from the world of recorded music (one important reason, I guess, is
>> that the former is much older), but goals of OL and Discogs are
>> similar (one page for every book/music record) and the means
>> (collaborative editing) are too. But I'm not here to just promote
>> Discogs - I like MusicBrainz very much too ;-)
>
>
> This is also being talked about in the library world. Identifying the
> publishers is tricky because of the various mergers, buyouts, etc. Keeping
> track of name changes isn't easy. But it would still be better, IMO, to take
> a stab at it.
>
I think it's possible, even though I'm not always sure which city to
enter in OL if a book is published by a publisher that is active in
many cities in many countries... :)
Is there data about mergers, buyouts etc somewhere online? Chambers of
Commerce perhaps? Or could OL be the first to try to collect that for
the library world?

>>
>> Just to make sure I understand you correctly: by "collaborator" you
>> mean "contributor" in the OL Edition? Or a role at the "creator"
>> level?
>
>
> sorry, "contributor"
>
I haven't come across organizations in the contributors, but that
would mean that the RDF shouldn't output these as foaf:Persons either.
>
>>>
>> Sindice.com does some simple inferencing: it can add the superclasses
>> of resources in a search result. Not all RDF software will do
>> reasoning of course, but I believe in the world of catalogs there will
>> be useful software that does do it.
>
>
> And responding to Ross here, I think we agree, though, that identifying
> actual person data from agent data might have some advantages, so we should
> do that where possible.
>
> kc
>
>
>>
>> Ben
>>
>>> kc
>>>
>>>
>>>
>>> --
>>> Karen Coyle
>>> [email protected] http://kcoyle.net
>>> ph: 1-510-540-7596
>>> m: 1-510-435-8234
>>> skype: kcoylenet
>>
>>
>

Ben
_______________________________________________
Ol-tech mailing list
[email protected]
http://mail.archive.org/cgi-bin/mailman/listinfo/ol-tech
To unsubscribe from this mailing list, send email to 
[email protected]

Reply via email to