Notes from the Linking Museums II meetup on September 27, 2010, London.
Initial (vague) notes by Mia but please feel free to edit and add your notes, or comment on the page if that's easier for you.
Who was there?
James (Kew), @jamesinealing
Jonty @Jonty
Paul Miller @PaulMiller
Greg Hadfield @greghadfield
Richard Boulton @rboulton
Joe Padfield (National Gallery) @JoePadfield
Jo Pugh (Nat. Archives) @mentionthewar
Jim O'Donnell @pekingspring
Mia (Science Museum), @mia_out
What was discussed?
- @leifuss couldn't make it but had been in contact earlier to suggest Antiquist could offer a prize to the first cultural heritage organisation to offer actual linked, open data (as his initial survey findings were that no cultural heritage organisation was there yet). What do you think?
- Paul mentioned a relevant JISC call, out in a few weeks
- We did a round of introductions (partly because I am selfish and wanted a chance to eat my dinner). Quite a few people from cultural heritage organisations were there because they are interested in publishing their data through APIs or as structured data and weren't sure where to start. And of course some developers were there because they want access to our data, hoorah!
- People at the start of the process of publishing data mentioned some of the issues they were encountering, including data sharing concerns from staff - attribution. Also, fear of/perceived penalties for being wrong (well, more precisely - being wrong in public).
- First question from developers when someone mentioned releasing data - what licence would it be released under?
- Joe has been charging ahead and is close to having entire collection online and linked to dbpedia. Web with SPARQL endpoints. Queries dbpedia to see if artists are in it. Worked on date range data for paintings/artists. CIDOC-CRM, has-possible, has-probable (fuzziness). Dealing with 'artist active' dates vs 'artist lived' dates - sometimes only know when they were active, not when they were born or died. [Apologies for the mangled notes, please feel free to correct them]
- Mia is looking to put subject authorities (people, places, events, topics) online so would love to work with people on testing the usability of the data for developers (alongside the user experience for the web visitor).
- Comment: GLAM data needs pictures to be compelling.
- Comment: 'a hack day with dull data is no fun for anyone'.
- 'labs' areas are really useful for working on projects in public for feedback and input and seems to be a way for people to understand 'beta', pilot-ish nature of the work and manage expectations around that
- Joe: D2RQ - can use to map collections database to live SPARQL endpoint or produce a dump file. [Cool!]
Where should this stuff live?
Is the wiki the best place for people to share stuff, think aloud, get feedback and work out what 'best practice' looks like? What do you think?
Related: what's the best way to push notifications at people - twitter, mailing list? What do you think?
In more detail - where should these discussions and beta work live? I know the wiki could do with a tidy-up generally, but we could also move to another platform if that'd be more usable. Joe has stuff on another, password-protected wiki about work he's done that he could put somewhere public, I'd love to share bits as I'm working on them and there are already some pages like that on the wiki. Ideally whatever platform we use should be discoverable by search engines, and content should include attribution.
Concrete issues for discussion include: dealing with dates, dealing with subject authorities (as discussed); a 'how to' for content providers (look also at what spacetimecamp and other hack days, open/linked data projects have produced, also Good APIs project, info from Europeana, etc?); a list of resources people have found useful; lists of questions to ask when assessing formats, standards; how to choose a licence that matches your existing terms and test it against proposed usage, etc...
Provocative questions
Some really useful questions were asked or emerged as issues over the evening - I'd love to know what you think.
- James: can you provide tools to help non-tech people use your data? Further, on discussion, what would an interface that helped people help themselves be like?
- Mia: as developers, what do you want from a subject authority? [a museum-y term, I know, but it's people, organisations, places, events, etc. e.g. biographies of scientists, events or places relevant to our collections]
- Is there any value in a survey of developers that would help museums understand the type and level of demand for machine-readable data? Ask developers their top wish request from GLAM (galleries, museums, libraries and archives). It could help prove demand (or not) and help museums understand what users need and better direct efforts. It could be as simple as: 1. what do you want to do with the data? 2. How do you want to access it? 3. what rights do you need? (variation: do you want to use it commercially?)
- Mia: geek accessibility - is the mention of RDF/SPARQL to a mid-level or tinkerer developer what stairs are to Daleks?
- [can't remember who]: what if the British Museum was a start-up?
- What does a good request for data look like? AKA, moving beyond FOI to sustainable (for the org) reusable (for the dev) provision of data.
Other stuff
Other events - http://historyhackday.pbworks.com
Next meetup
Possibly at spacetimecamp meetup on November 5th, and then maybe one later in November? Jonty to organise (yay!).