NLA Party Infrastructure
Basil Dewhurst
Thought I'd post some info and links about the NLA Party
Infrastructure being created by the Australian Research Data Commons
(ARDC) Party Infrastructure Project to grease the wheels so-to-
speak ...
NLA Party Infrastructure Overview: "The ARDC Party Infrastructure
Project is a National Library of Australia (NLA) project funded by the
Australian National Data Service (ANDS). The project is adapting the
NLA Party Infrastructure to enable the distributed management and
persistent identification of records about Australian researchers and
research organisations. In doing so it will encourage the better
management and curation of data generated by Australian research and
supports the long-term access, use and re-use of this data. Detailed
information about the project is available via the project wiki
located at http://wiki.nla.gov.au/display/ardcpip."
A detailed (slightly technical) overview of the infrastructure agreed
with ANDS and key stakeholders in the research sector is available at
http://bit.ly/9CpX7X.
Regarding party data and the Australian collections sector, late last
year I ingested all Collections Australia Network institution records
into the NLA Party Infrastructure and you can find these on Trove
(http://trove.nla.gov.au/people/result?q=au+nmus+can). This means that
all collecting institutions that CAN knows about have an NLA party
identifier (eg http://nla.gov.au/nla.party-783826).
In terms of the ARDC, parties (people and organisations) that relate
to collections/activities/services should be represented where
possible. This greatly improves discovery and access in the ARDC and
more broadly in Australia's Distributed National Information
Infrastructure - benefits exist here for both data contributors and
also end-users. The specific relationship of a party to the
collection/access/service record should be asserted where possible eg
'isManagedBy', 'isOwnedBy' or 'hasCollector' are very useful but the
blanket use of 'hasAssociationWith' not so much. Yes this takes a
little effort (but not much more that creating a party record or
locating an identifier) but there are significant benefits in
asserting specific relationships between things (think Semantic Web/
Linked Data here) that will accrue over time.
For individuals and organisations, often-times an identifiers/records
exist in the NLA Party Infrastructure (eg Hedda Morrison (http://
nla.gov.au/nla.party-615689), Charles Kerry (http://nla.gov.au/
nla.party-469651), Bureau of Meteorology (http://nla.gov.au/
nla.party-593013)) and these identifiers can be used in collection/
service/access records. The NLA Party Infrastructure provides a
number of web service interfaces that would allow the MME to obtain
stub records (thereby reducing 'cataloguing' effort) for parties that
we know about already. For those we don't, a party record would need
to be provided to the Party Infrastructure for an NLA Party Identifier
to be created. The process is described in the overview I mentioned
above (http://bit.ly/9CpX7X) and all components of the Party
Infrastructure are functional and available now. The rest of the ARDC
Party Infrastructure Project is focused on improving these tools and
working with some early implementers. We already have records from
QUT and Griffith University and are currently working with Monash
University, University of Melbourne and CSIRO to set up feeds.
So ... Ingrid, some early inter-project discussion might be useful !
Happy to answer any questions/provide advice.
Regards,
Basil
Basil Dewhurst | Project Manager, ARDC Party Infrastructure Project |
National Library of Australia
p: +61 2 6262 1046 | e: bdew...@nla.gov.au | w: wiki.nla.gov.au/
display/ardcpip
Mason, Ingrid
go astray!
I haven't managed to catch the email exchange I'll probably look at it
tomorrow and discover you've solved all this... ;-)
We're looking to avert scope creep but also to set in blocks to enable
"next steps". In my mind that means forging the first and most obvious
link between the Collection and the Party that manages it and describing
that Party Relation Type as "isManagerof" in the Party object and
"isManagedBy in the Collection object in the Collection Relation Type
using URI. After that is where the water gets pretty choppy - we're
looking at ways to enable richer relationship creation, and also not
overwhelm the Data Analyts either in the timeframe we've got.
How about several scenarios to test our thoughts a little:
The Hedda and Alastair Morrison collection
The Jenny Kee and Linda Jackson collection.
The Lawrence Hargraves collection
Your thoughts welcomed...
Ingrid
----------------------------------------------------------------------------------------------
This email and attachments are for the use of the intended recipient(s) only and may contain confidential or legally privileged information or material that is copyright of Powerhouse Museum or a third party. If you have received this email in error, please notify the sender immediately and then delete it. If you are not the intended recipient, you must not use, disclose or distribute this e-mail without the author's prior permission. Any views expressed in this message and attachments are those of the individual sender and the Powerhouse Museum accepts no liability for the content of this message.
Basil Dewhurst
Let's try this again with an attachment rather than hoping an html table will render properly ...
Forging an early link between collections and parties seems a really good way to go - especially from our perspective - and describing managers/owners of a collection early would be a good first step.
Re: the mappings in your PDF file, I think there are a few things that weren't quite right. Below as an example I've taken the statements about the Hargreaves collection and popped them into a table. If you read these statements like a sentence subject-predicate-object then this will help with which relation types to assign (eg row 6 the party http://nla.gov.au/nla.party-573769 [="Powerhouse Museum"]) is the manager of mme:LHC [="Lawrence Hargreaves Collection"]). I've included your mapping as the Existing mapping and proposed mappings for rows with issues.
Hope this helps and note that the response is non-normative ie I think you'd want to check it with people like Tim Sherratt and ANDS before nailing it down !
Cheers,
Basil
Basil Dewhurst | Project Manager, ARDC Party Infrastructure Project | National Library of Australia
p: +61 2 6262 1046 | f: +61 2 6273 1180 | e: bdew...@nla.gov.au | w: wiki.nla.gov.au/display/ardcpip
Tim Sherratt
Yep, Basil's reworking seems to make sense to me. 'hasAssociationWith'
is for an 'unspecified relationship' so it makes sense to avoid it
where more specific claims can be made.
In the end this is about discovery. The possibility of being able to
explore a network of people and the collections they made seems to me
to be one of the most exciting aspects of the project.
BTW I agree with Alex that the relationships currently being used in
RIF-CS are limited and it would be good to be using an external
vocabulary. As Robina noted, the vocabs can change, so it'd be good to
think more about what we need.
cheers, Tim
Mason, Ingrid
The examples online on the MME project site are aimed at introducing the scheme to conceptualising and describing people...we need a tech version too. Lynne has been beavering away at this and feeding our approach to Robina. She has also made a good case for recognising the value of human reading (rather than machine reading) of data.
So I've now got nagging thoughts about data labels for dates, locations and relationships aside the mechanics of computation.
Ingrid
Ps. I'm also thinking it might be beneficial to have several people talking about this in a room for a couple of hours in Canberra. Would that be useful?
All,
cheers, Tim
> ------------------------ This email and attachments are for the use of