AustLit [ http://www.austlit.edu.au, described here:
is basically a triple store of around 10 million statements
implemented as an extremely simple data model in conventional
Appropriate index creation and caching of "reconstructed" objects
in the application give acceptable performance running on a 4 yr old
Sun workstation (750Mhz, 1GB RAM). Typical search queries generate
between 5 and 20 relational joins on the triple store table.
The system consists of 2 layers:
- the triple store
- a logical FRBR view of objects reconstructed from the triple store
These FRBR objects are exposed as XML which are styled using XSLT to
produce the display view.
Unfortunately, AustLit has to recover ongoing costs as a not-for-profit
subscription service. But Kerry Kilner [[log in to unmask] ], the
executive manager of AustLit, is happy to arrange for temporary free
for people interested in exploring the system.
>From: Metadata Object Description Schema List
>[mailto:[log in to unmask]] On Behalf Of Kevin S. Clarke
>Sent: Wednesday, 13 July 2005 12:05 AM
>To: [log in to unmask]
>Subject: Re: [MODS] library metadata and RDF
>On 7/12/05, Leigh Dodds <[log in to unmask]> wrote:
>> In terms of graph complexity, I'm not aware of any limits or
>I'm sure there is no problem modelling very complex structures in the
>graph, my questions really are more about being able to pull a very
>normalized record back together into one piece rapidly.
>There was an article awhile back (in a lib sci journal... I can't
>remember the article or journal off the top of my head but I should
>look it up again) looking at the record as a document. Since library
>records are mostly read-only (infrequent updating) and we do want
>access to the whole more often than to discreet units, I thought this
>was an interesting approach (rather than approaching metadata more
>like business data). Anyway... I'm just rambling.
>Thanks for the answer, Kevin