On 6/21/12 12:10 PM, Ben Companjen wrote:

> That is not a property, but a class.


Aargh! Sorry, I thought I was in the property list. Obviously not awake 
yet. (Just did a fairly smooth but long 24-hour door-to-door Europe to 
California trip. I want the transporter to be invented NOW!) This could 
be another reason why we didn't use series. However, as for the series 
property in the RDA ontology:

http://metadataregistry.org/schemaprop/show/id/296.rdf

This is "series statement" from the ontology for RDA, Resource 
Description and Access, the upcoming new cataloging rules that will be 
used in the US and other countries. It is called "series statement" 
because it may consist of a complex data statement, including author, 
series title, and enumeration. It can be used with literals. (The term 
"statement" in RDA is not the same as "statement" in RDF, which can be 
confusing.) but....

> where oledition is a prefix for<http://openlibrary.org/editions/>  and
> I disregard that we would probably want to link Works and Editions
> (are they both bibo:Books? One of them? Neither?) :)

>
> I couldn't find a suitable property in the ontologies at
> http://purl.org/spar (although it has the FRBR aligned bibliographic
> ontology, which may be of interest for OL for other properties). There
> is http://rdvocab.info/RDARelationshipsWEMI/inSeriesWork to relate an
> RDA Work to a series, but the property's range is undefined. I don't
> think it is supposed to be used for Literals, though.

Note that very few properties in RDVocab have been given a range... yet. 
In traditional cataloging, all of the date elements defined in the 
cataloging rules take literals as values; in RDA there is still work 
ongoing to determine how (if it is possible) to define a set of 
properties that can have both literal and non-literal values. In the 
rules themselves it is often stated that you can use either a textual 
description *or* an identifier, but there's nothing specific about how 
that would be done, or even of what might be found when the identifier 
is resolved. (Heck, the identifiers do not even have to be URIs.) So 
what you see there at rdvocab.info is somewhat embryonic because it's a 
first attempt to define traditional library data in RDF. (Some of us 
think that defining traditional library data in RDF is not only not 
feasible, but it isn't even a good idea.)

Another example is from ISBD (International Standard Bibliographic 
Description) which has "has series and multipart monographic resource area"
http://metadataregistry.org/schemaprop/show/id/2146.rdf

defined as:

"Relates a resource to a statement including the title proper of a 
series or multipart monographic resource, the parallel title of a series 
or multipart monographic resource, the other title information of a 
series or multipart monographic resource, the parallel other title 
information of a series or multipart monographic resource, the statement 
of responsibility relating to a series or multipart monographic 
resource, the parallel statement of responsibility relating to a series 
or multipart monographic resource, the international standard number of 
a series or multipart monographic resource, and the numbering within a 
series or multipart monographic resource."

Just from this, it might be best to avoid this one for its conceptual 
semantics.

One reason to have a series property that is a literal is that at the 
time you create the bibliographic description you may not have handy an 
identifier for the series that is listed on the book. Requiring the 
creation of an identifier for that series could be impractical, so it 
makes sense to me to allow people to record the series as a literal, and 
that that is preferable to not recording it at all if a URI is required. 
This supports the suggestion (was it Alan's?) that these literal series 
statements might be considered notes; additional information about the 
book, but not really expressed as a relationship between resources.

As for "Work series" in RDA -- I have no idea what is meant, and will 
have to find where in the rules it is described. It seems odd to me. As 
does much of what appears in library cataloging rules.

kc


>
> Perhaps it is time to create http://openlibrary.org/type/series and
> let (bot) users link works or editions to series objects.
> The other possibilities I see are disregarding the series information
> in RDF, using a more general property like dc:relation and creating a
> new property ol:series.
>
> Ben
>>
>> kc
>

-- 
Karen Coyle
[email protected] http://kcoyle.net
ph: 1-510-540-7596
m: 1-510-435-8234
skype: kcoylenet
_______________________________________________
Ol-discuss mailing list
[email protected]
http://mail.archive.org/cgi-bin/mailman/listinfo/ol-discuss
To unsubscribe from this mailing list, send email to 
[email protected]

Reply via email to