On 13/02/2014 09:43, James Grimster wrote:
> All,
>
> a recent effect has been that a lot of regional and national aggregators have been using variants of Apache Lucene / SOLR or Elasticsearch indexing fronted with common forms of facet refining REST APIs; this is the case for CultureGrid, Europeana, DigitalNZ, Trove, our CollectionsBase and Rob's CIIM (he'll correct me if I'm wrong) . With the work being done at TNA and EH , a self organising distributed 'grid' as it were is emerging. Creating a federating meta, meta search on top should be reasonably trivial.
It's interesting, though not that surprising, that these APIs are
defining our expectations in terms of cross-collection search. However,
both the nature of the data provided - typically search-oriented
metadata - and the means of access - those lovely facets - represent a
dumbing-down from what we might ideally aspire to. Shades of Z39.50 [1]
application profiles. It is certainly a more simplistic search
environment than, say, ResearchSpace [1]. Conversely, it may be
achievable at scale.
If we /are /going in that direction, I think it behoves us as a
community to define an "application profile" which represents our
collective best guess at our cultural heritage search requirements:
something to which contributing aggregators can align their own access
points.
> The limiting factor in the UK will be the continuing lack of any cross sector (museums, archives, libraries, archaeology) vocabulary harmonisation. I've seen CIDOC CRM overlaps at the 'edges' of museums / archaeology , but without a distributed grid of harmonising vocabulary services, metadata will continue to be stuck in conceptual silos. The Finns have made a good stab at it with ONKI. Will the UK ever meet this (very difficult) challenge?
Three points to pick up here:
* why just the UK? Frameworks like SPECTRUM and the CIDOC CRM are
international in scope, as are standardised resources like Getty AAT
(now available as Linked Data [3])
* I see the CIDOC CRM as a framework for expressing relationships
between cultural heritage entities. There would be no role for it in
the "application profile" search environment you are sketching out.
Personally, I would regret this
* the term "vocabulary" is unhelpful, in that what we need is a more
rigorous alignment of concepts, not just words which look similar.
This requires more of a Linked Data approach
Richard
[1] http://en.wikipedia.org/wiki/Z39.50
[2] http://www.researchspace.org/
[3] http://www.getty.edu/research/tools/vocabularies/lod/
>
> all best
>
> James
>
> On 12 Feb 2014, at 10:35, Nick Poole wrote:
>
>> Culture Grid is a lot less like Europeana than people seem to think, though. Our ambition is not really to provide a search box for collections. The ambition since the beginning has been to provide an open platform as a service to the UK museum sector, supporting exactly the kinds of cross-organisational sharing of collections data that has been the subject of this thread.
> ****************************************************************
> website: http://museumscomputergroup.org.uk/
> Twitter: http://www.twitter.com/ukmcg
> Facebook: http://www.facebook.com/museumscomputergroup
> [un]subscribe: http://museumscomputergroup.org.uk/email-list/
> ****************************************************************
>
--
*Richard Light*
****************************************************************
website: http://museumscomputergroup.org.uk/
Twitter: http://www.twitter.com/ukmcg
Facebook: http://www.facebook.com/museumscomputergroup
[un]subscribe: http://museumscomputergroup.org.uk/email-list/
****************************************************************
|