[OAI-implementers] Sets in and subjects in OAI-PMH

Hussein Suleman hussein@cs.uct.ac.za
Tue, 21 Oct 2003 23:10:44 +0200


while everyone says their piece on standardisation of sets, can we 
please think carefully on where this falls in the spectrum between data 
provision and service provision.

it seems this is a first-tier service rather than a data provider 
function. as such, it can be implemented (and has been in the ODL 
project and NDLTD Union Catalog) as a component that analyses metadata 
and dynamically organises records into sets for further 
access/harvesting. of course this means thinking of digital libraries as 
collections of service-oriented components ... very much in line with 
the Web Services philosophy.

so why aren't other service providers doing this?


p.s. is LCSH international? is it used in South Africa? Ethiopia?

Thomas G. Habing wrote:

> Jody DeRidder wrote:
>> Seems to me that search, selection and sortation would *all* be much more
>> useful at the service provider level if we had standardized lists of 
>> sets AND
>> subjects (LCSH?  SEARS?)  from which to select when creating records 
>> at the
>> repository level.
>>   --jody
>> _______________________________________________
>> OAI-implementers mailing list
>> List information, archives, preferences and to unsubscribe:
>> http://oaisrv.nsdl.cornell.edu/mailman/listinfo/oai-implementers
> For an interesting view of set usage across different OAI repositories 
> look at my Experimental OAI Registry at 
> http://oai.grainger.uiuc.edu/registry/, specifically the miscellaneous 
> reports:
>   Repositories Sorted by Number of Sets
>   Frequently Occurring Distinct Set Specs
>   Browse All Distinct Set Specs (hex encoded set are decoded)
>   Frequently Occurring Distinct Set Names
>   Browse All Distinct Set Names
>  From this it appears that set usage is all over the place.  There are 
> some repositories that have more sets than they do individual records.  
> There are also many that list sets for which they have no records.  And, 
> while I understand why it was done, I am finding the setSpecs that are 
> hex encoded strings to be quite annoying -- my OAI pet peeve of the 
> moment. :-)
> Anyway, I guess I am seconding the call for some set standardization as 
> long as it stays simple.  I also like Sarah Shreeves idea that this 
> should be driven in large part by service providers and not just 
> metadata providers.
> Kind regards,
>     Tom

hussein suleman ~ hussein@cs.uct.ac.za ~ http://www.husseinsspace.com