Well, the scan(index) operation returns the complete set of values used
by all records in the collection for that index, so a
scan("dc:subject") will return all the distinct keywords that had been
used within the target collection.
This isn't quite the same as discovering where those keywords had come
from (that is, an authority list), but having the complete array of
keywords used in a collection to date can be useful anyway, if nothing
else to check whether UK or US spellings are in vogue :-)
For (a trivial) example, if the collection contains:
dc:subject=ants, bees, wasps
dc:identifier=2 dc:subject=wasps, trees
... scan("dc:subject") will return "ants
[1],bees[1],wasps[2],trees[1]". (OK, its a bit more sophisticated than
that - see the bottom of this page for a 'real' example:
http://www.loc.gov/z3950/agency/zing/srw/sru-examples.html)
If something is being used as a keyword rather than a classification, I
would suggest the source of the keyword isn't that critical, but
consistency in choosing keywords definitely helps.
I guess in the EEVL example, a registry query against a collection
tagged by EEVL could turn up the complete set of EEVL authority lists.
A scan on the other hand would turn up all the EEVL keywords in current
usage in the collection regardless of which list they came from
originally.
Although I've used dc:subject as an example, the same applies to any
index term understood by a target collection.
More info on srw/u scan operations here:
http://www.loc.gov/z3950/agency/zing/srw/scan.html
There is a space in the scan response for "extraTermInfo" that could
identify the authority list URL(s) for any keyword, but I don't know if
anyone has actually implemented such a thing.
- S
On 24 Feb 2005, at 21:26, Phil Barker wrote:
> Sorry Scott, I'm not sure I understand you. I think what you suggest
> could only apply to discovering the source of the vocabulary that was
> being used. I don't see how the keywords that had been selected as
> applicable to a given resource could be discovered like this. Am I
> right in thinking that what you mean is "Instead of somehow including
> the SOURCE OF THE vocabulary terms directly ..."?
>
> Assuming I'm right, you're still stuck with the problem that a
> catalogue (or even an individual record) might contain keywords from
> several controlled sources. Consider EEVL, they catalogue Engineering,
> Maths and computing resources, they use different subject-specific
> controlled vocabularies for each subject (let's assume these are
> keywords rather than classification--even though that's not quite how
> they are using them). So a single record about a resource on maths for
> engineers will get a key words from the Engineering and Mathematics
> vocabularies. Will what you describe below be able to say which term
> comes from each source? (And would it matter if it couldn't?)
>
> Still, sounds better than nothing. Thanks for the suggestion.
>
> All the best, Phil.
>
> Scott Wilson wrote:
>> Hi gang,
>> Instead of somehow including the vocabulary terms directly within the
>> (ever expanding) metadata record, why not use z39.50/SRW/SRU
>> operations to dynamically discover this information from the target
>> when searching? e.g.. Scan(fieldname): for terms in current use for a
>> particular field, as a list of tokens in XML
>> Explain() doesn't seem to return the right sort of things in the
>> Zeerex record for discovering controlled vocabularies, although I'm
>> no expert (but I'm CCing one :-)
>> Alternatively, this information could be provided in a service
>> registry (e.g. the IESR) or metadata schema registry (e.g. IEMSR).
>> Just an idea :-)
>> - S
>> On 24 Feb 2005, at 02:38, Sarah Currier wrote:
>>> Hi Phil,
>>>
>>> Phil Barker wrote:
>>>
>>>> Just want to be clear what you're suggesting for UK LOM Core / IMS
>>>> Best
>>>> Practice Guide
>>>
>>>
>>> Cool.
>>>
>>>> Sarah Currier wrote:
>>>>
>>>>> 1. There are two things: keyword indexing, and classification.
>>>>> They are
>>>>> different, serve different functions, and are both important.
>>>>>
>>>> ...
>>>>
>>>>> 3. ... Where keyword
>>>>> indexing does involve a publicly available controlled vocabulary
>>>>> (as
>>>>> opposed to an internal authority list as HLSI use) I think it
>>>>> would be
>>>>> sensible to make this clear in the metadata as with other
>>>>> vocabularies
>>>>> in the LOM.
>>>>
>>>>
>>>> Bit difficult to "make this clear in the metadata as with other
>>>> vocabularies in the LOM" if you're using either of the keyword
>>>> elements
>>>> since they have langString rather than vocabulary datatype, so you
>>>> can't
>>>> put in source-value pairs. Are you suggesting using the taxonPath
>>>> element in the classification category? or does this conflict with
>>>> your
>>>> first point?
>>>
>>>
>>>
>>> The last bit about TaxonPath DOES conflict with my first point. In
>>> typical librarian fashion I didn't really get the technical ins and
>>> outs so it sounds like what you are saying is that it's not
>>> technically possible to declare the vocabulary of keywords within
>>> the LOM as it currently stands because of the different datatype.
>>> Therefore it will only be possible for someone to use a published
>>> controlled vocabulary, but not to declare it (although they could
>>> name it in their AP which they should make publicly available). So I
>>> guess if we're only looking at what the LOM Core / Best Practice
>>> Guide can say about this, I would suggest that they just make this
>>> point clear.
>>>
>>>>> I wonder if there is some way of also making this clear in the IMS
>>>>> Metadata Best Practice Guide?
>>>>>
>>>> I'll see whether we can look into that when moving from 1.3 Public
>>>> Draft
>>>> to Final.
>>>
>>>
>>> Cheers! :-)
>>>
>>> Sarah
>>>
>>> --
>>> *******************************************
>>> Ms. Sarah Currier
>>> Librarian, Stòr Cùram Project
>>> "A Storehouse of Learning Resources for Social Care"
>>> Dept. of Social Work, University of Strathclyde
>>> c/o: Centre for Academic Practice, University of Strathclyde
>>> Graham Hills Building, 50 George Street
>>> Glasgow G1 1QE, Scotland, United Kingdom
>>> Web: http://www.storcuram.ac.uk/
>>> Tel.: +44 (0)141 548 4573 Fax: +44 (0)141 553 2053
>>> E-mail: [log in to unmask] Mob.: +44 (0)7980 855 801
>>> Stòr Cùram is Gaelic for Storehouse of Care
>>> *******************************************
>
>
> --
> Phil Barker Learning Technology Adviser
> ICBL, School of Mathematical and Computer Sciences
> Mountbatten Building, Heriot-Watt University,
> Edinburgh, EH14 4AS
> Tel: 0131 451 3278 Fax: 0131 451 3327
> Web: http://www.icbl.hw.ac.uk/~philb/
>
|