Date: Thursday, December 16, 6pm-8pm.
Venue: the bar of the Hotel Grosvenor Kensington, 2-10 Harrington Road, London SW7 3ER
http://www.google.co.uk/maps/place?cid=2019180852904133914&q=grosvenor+hotel+south+kensington&fb=1&gl=uk&hq=grosvenor+hotel&hnear=South+Kensington,+Greater+London
Why? There's a 'working bee' earlier in the day to try and mark up some biography records, so the meetup is a chance for developers and content providers to reality check the work by providing feedback on the usability and utility of the records as marked up, contributing use cases, examples of working practice, etc.
To quote Richard Light discussing it, "What we're hoping for is movement towards an agreed "design pattern" for encoding historical information as Linked Data. If we were to use the CIDOC CRM, for example, there are lots of ways in which you can make the same assertion, "full CRM" vs. "short-cuts", "names" vs. "appellations", etc.
If there were a published template to follow when recording life events, objects in their historical context, etc., then e.g. SPARQL queries against the resulting Linked Data would work more effectively. It also means that there would be some point in pooling such Linked Data from multiple sources."
Who? This event is for you if you're interested in publishing, sharing or using data about museums and collections through technologies such as linked data and microformats.
Here's a draft agenda for the working bee - feel free to edit or comment to add specific things you think should be addressed.
What kinds of data?
Mia: I'm looking at content like http://www.sciencemuseum.org.uk/broughttolife/people/alexanderfleming.aspx
http://www.sciencemuseum.org.uk/broughttolife/people/typhoidmary.aspx
http://www.sciencemuseum.org.uk/broughttolife/people/onesimus.aspx
Sign up or register your interest now!
I'm coming [edit and add your name here]
Mia @mia_out
Richard @richardofsussex
Joshan @thebritishmuseum
Maybe [edit and add your name here]
Jonty @jonty
Regrets [edit and add your name here]
Greg @GregHadfield - Sorry...I've now got to be in Brighton :-(
James (Kew) @jamesinealing - afarid I've gone from a Maybe to a Regrets due to illness (and not even Christmas party induced!)
Jim @pekingspring – work xmas party that night.
Jo (TNA) @mentionthewar - also Christmas party. I'll see if someone else can make it.
Comments (5)
Mia said
at 1:04 pm on Dec 8, 2010
Jim, Jo - sorry you can't make it - pesky Christmas parties!
Joshan Mahmud said
at 10:57 am on Dec 16, 2010
Hi Mia
Having read the blurb for this meeting, I was just wondering about the main intention of this meeting, as it seems that you wish to find a 'standard' way in which to map cultural heritage records to something like the CRM. However, isn't the standard the CRM itself and the way in which records are mapped to it specific to each organisation and dependant on the original schema? It is important that CRM implementations directly relate to the original units of information in the source dataset.
We have done some work with the CRM and have consulted with the original creator of the ontology who have reached a point with it that attempts to cater for almost, all scearios and variations of cultural heritage records beyond which it becomes far more difficult to standardise and other semnatic techniques are then required such as: ontology extension and mapping or co-referencing.
What are your thoughts?
Josh & Dominic
(PS - there is a possibility we may not be able to make it tonight hence this posting)
Richard Light said
at 11:27 am on Dec 16, 2010
Even if you accept the CIDOC CRM as a basis for publishing as Linked Data (and at least one delegate this afternoon will be suggesting other ontologies), there are still choices to make: native CRM; Erlangen CRM; or the rather neglected CRM Core? Then, when you come to use your chosen flavour of CRM, there are lots of ways to express any given fact (full vs. "short-cuts"; string values vs. URLs; etc.).
All of these variations will have an effect on the form of the RDF which you generate. Inconsistently-structured RDF is harder to query effectively, and therefore there is less merit in pooling such RDF resources from multiple institutions.
At present, very few (if any!) institutions are actually publishing Linked Data. Apart from issues of the exact syntax to use, there are broader questions of _what_ data should be published in this way. If we can put a suggested "pattern book" into the public domain, there is a chance that at least some institutions will follow it, and the resulting published data will be that much more consistent.
The volunteers coming this afternoon represent the archives and TEI communities as well as museums, so it is also an exciting opportunity to get a cross-sectoral view (and hopefully consensus) on this issue. This is why, Dominic, I mentioned the meeting to you at the UKMW10, and why I was hoping you would be able to come along this afternoon to share your experience from the BM mapping work.
Mia said
at 1:48 pm on Dec 16, 2010
Hi Josh,
I hope you guys can make it still!
I've started an agenda at http://museum-api.pbworks.com/w/page/33902115/Linking-Museums-III-%27people%27-writeup to help clarify some of the questions we should address before cracking on with it. If there's stuff you'd like to see, please feel free to edit it.
As to the goals - with any markup, the devil is in the details, and it's really only by sitting down and working through real data and potential user requirements that progress is made. I think it's going to be important to differentiate between different uses of the data - mapping to a common standard for external developers to use in hack days and mashups is an important goal, as is providing records to aggregators like Europeana and Culture Grid, but they might in turn be served by an institutionally-specific schema and internal service.
Which reminds me, is the work you're doing at the British Museum mostly designed for internal use or is there a public-facing output too?
cheers, Mia
Dominic said
at 3:39 pm on Dec 16, 2010
Richard
I completely agree with you on both points. On the first you will never get everybody to agree on everything whether it is sql schemas, taxonomies or ontologies (I note and I am more than content with your CRM dissenter). On the second point, that doesn't mean that you shouldn't try to get some consensus where possible and I think that these meeting look very useful and potentially very important and I am very grateful that you have invited me and Josh. It is only because of work commitments that we are unable to contribute this afternoon, but Josh (who will be more useful in any case) and I are hoping to stop by at six o'clock. However, I do think that it is probably better to talk about guidelines and best practice rather than pattern books, although these may be the same thing? See you later.
You don't have permission to comment on this page.