Re: Wikidata book schema

Pascal, what I am finding, perhaps because it's what I hope to find ;-), 
is that most functioning systems create a graph design with few 
properties exclusive to a particular entity. So I agree that Works can 
have subjects, but so can Editions. However, much depends not on the 
defined ontology but how one "munges" the data. In Open Library, the 
incoming data has subjects associated with editions, and these are 
gathered by the Work in a similar way in which VIAF [1] gathers name 
variants from the many authority records in its merged set. So I think 
it's best to think about a dynamic graph, not a flat ontology. (And, no, 
I'm not entirely sure what I mean by that, but it sounds right. I'll 
think about this some more.)

kc


[1] http://viaf.org

On 9/16/13 8:44 AM, Christoph, Pascal wrote:
> Am 15.09.2013 17:15 schrieb Antoine Isaac :
>
>> The round-up of sites sounds like a great idea, Karen!
>
> The comparison may find that the wikidata book task force[0] misses a
> "dcterms:subject" on the work level. This is what at least Etienne and me is
> thinking.[1]
> Not sure if my proposal "main category topic"[0] is correct - what do you think?
>
> pascal
>
> [0]https://www.wikidata.org/wiki/Wikidata:Books_task_force
> [1]https://twitter.com/dr0ide/status/379496420870328320
>
>> A.
>>
>>> Thanks, Antoine. I hadn't seen this. I note that they refer to "Work" and "Edition" which is also the terminology used by Open Library. (Plus they have "item" for individual books, like rare books.) I've begun a (hopefully short) round-up of bibliographic sites to see what levels of abstraction they use, and what they call them. This fits into that nicely.
>>>
>>> kc
>>>
>>> On 9/15/13 2:58 AM, Antoine Isaac wrote:
>>>> Dear all,
>>>>
>>>> This may have been sent to the list before, but in case...
>>>> https://www.wikidata.org/wiki/Wikidata:Books_task_force
>>>>
>>>> I believe this can be a quite useful reference in terms of use case.
>>>> These are properties that somehow reflect user needs, it's likely that
>>>> it would end expressed in schema.org one day.
>>>>
>>>> Would it be a task for this group to have a look at this schema, and
>>>> flag any missing properties to schema.org?
>>>>
>>>> Note that it could also bring input for our 'work' debate. They have
>>>> only two levels, work and edition. Apparently they regard the edition to
>>>> be either the expression or manifestion (or both of them in fact), and
>>>> the link between the edition and the work is simply 'edition of'.
>>>>
>>>> Cheers,
>>>>
>>>> Antoine
>>>>
>>>>
>>>
>>
>>
>
>
>

-- 
Karen Coyle
kcoyle@kcoyle.net http://kcoyle.net
m: 1-510-435-8234
skype: kcoylenet

Received on Monday, 16 September 2013 16:06:10 UTC