About Me

My photo
Web person at the Imperial War Museum, just completed PhD about digital sustainability in museums (the original motivation for this blog was as my research diary). Posting occasionally, and usually museum tech stuff but prone to stray. I welcome comments if you want to take anything further. These are my opinions and should not be attributed to my employer or anyone else (unless they thought of them too). Twitter: @jottevanger

Thursday, April 30, 2009


Jim O'Donnell organised a talk on Tuesday at the National Maritime Museum from Christian Heilmann of Yahoo! Mia wrote up her notes already and I've not got much to add, but it was a very enjoyable presentation, and when he reached the juicy bit about YQL and BOSS, both of which I'd left for another day's exploration, I learned a lot. Clearly there's a lot of potential there (especially now it's augmented by YQL Execute, announced yesterday), and it looks like it will let you do a bunch of things that Pipes can't do, or is a pain to do (the GUI is great and yet infuriating with Pipes). YQL gives a common API meta-interface (I guess that's the word) for loads of other APIs and for things with no API; it also handles all the crap with authentication, tokens etc; and it will act as the gatekeeper for your API so you don't get hammered by unreasonable numbers of requests.

As with similar tools/services (Pipes, Dapper, dbpedia, and various things nearer the surface like GMaps), YQL is clearly a blessing from both ends of the telescope: we get to use it for its intended purpose - to be "select * from Internet" is the grandiose ambition - knitting together data sources from Yahoo! and beyond; and we also get to offer our data in a developer-friendly way to encourage its reuse by creating OpenTables [note that these are purely a machine-friendly description of how to access data: no data is handed over as such]. Jim has already been busy creating Open Tables and experimenting with YQL.

Following the talk we headed for a pint (and one of themost jaw-dropping jokes I've heard, from Chris), and it was good to talk to Tristan from Cogapp. When I stopped raving incoherently about the marvel that is Solr (yes, still in love even as I gradually find out more about it), Tristan cleared up some questions for me about Cogapp's COBOAT app. They recently open-sourced this (as far as possible), in the context of the Museum Data Exchange project with OCLC (see Gunter Waibel's recent post), where it plays the role of connecting various collections management systems to an OAI Gateway-in-a-box, OAICatMuseum (well seems like it's only used with TMS in the project, but the point of COBOAT is that it just makes life easier for mapping one data structure to another, and another CollMS would slot in just fine).

For me, both COBOAT and OAICatMuseum are of interest for the role they could play in our the revamped Collections Online Delivery System* we'll build this year, resources allowing (in other words, don't hold your breath. Mission critical, yeah, but worth paying for? I await the answer with interest). Integrating and re-mapping data sources, an OAI gateway, and sophisticated and fast search are key requirements, as is a good clean API, and taking these two applications along with Solr I feel like I may have identified candidates for achieving all of these aims. We're a long way from a decision, of course, at least on the architecture as a whole, but I have some tasty stuff to investigate, and I'm already well down the track in my tests of Solr.

Thanks again to Jim for arranging the talk. He's got another great guest coming up, hopefully I can make it to that one too.

*I'm resigned to this thing being called CODS but still hoping for something less, well, shit

No comments: