On 6/2/2010 9:04 AM, Karen Coyle wrote:

[snip]

> The problem I see with RDA's Person is
> the problem I see with all of RDA which is that it is tied to FRBR
> entities as classes, so if you use it you can't really get far from
> FRBR.

I fail to see why this is a problem. I though OL was trying to move to a 
FRBR model in its database.

Of course, it is almost always a mistake to try to build a database 
model from a data transfer format. Whether the source, or the sink, of a 
data record is XML, MARC, XML:RDF, JSON, or any other arbitrary format 
should be largely irrelevant. As Bill Dueber recently pointed out, 
"[a]nyone advocating or dismissing a data model based on the data 
structure or serialization most-often associated with that model is 
missing the goddamn point."

Whether you want an author expressed as a foaf:Person, or a dc:creator, 
a generic rdf:description, or even a JSON text object depends on the 
consumer of the data. If the database has been designed correctly, any 
of these transfer formats can be easily generated.

So before any questions about how best to represent a person in RDF can 
be addressed, you should try to find out who will be consuming the data, 
and what their expectations are. If there are no existing applications 
out there designed to consume data in a know format, or which can easily 
be modified to consume data in a slightly altered format, then it 
doesn't matter how the data is serialized; a new application will have 
to be written to consume the data anyway--pick anything you like that 
carries all the data you need. If I were writing a new application to 
consume OL data, I would use the JSON object; it's all UTF-8 so it's 
easy to parse, and so far it's the only interface around that guaranteed 
to furnish /all/ the data currently held by OL.

Of course, there remains the question as to just how much effort should 
be put into defining this interface, as those who would be consuming it 
are not the primary target audience for OL. I would echo the suggestions 
of several others in saying OL should take all the time it's spending on 
creating Yet Another API, and spend it instead on measures to increase 
the quality of the existing data.
_______________________________________________
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