It’s 1973 and you’re not sure whether to make burnt orange curtains or sunshine yellow ones to go with your lime green kitchen benchtops. You need information - taste information , opinion information.
You ask Marie over the back fence what she thinks. She’s just bought new curtains, so she’ll understand. You ask Aunt Helen who has worked as an interior designer. You phone up your second best friend because you always love what she does with her house.
You don’t ask Jilly - she’s into those off-beat pastel colours. You don’t ask Uncle Jim - he always tries to tell you about trainspotting.
You definitely don’t look in Encyclopedia Britannica. That is full of great objective information - but you are after subjective information. An authoritative source in this case is someone with similar tastes, or whose taste you admire, maybe someone with professional skills.
In 1998, we thought we were clever mounting our objective information up on the web. What we could store on paper was transferred to the web and it became a set of interlinking brochureware. Then came Web 2.0.
Web 2.0 is heralded as the era of user created content. But more is happening than amateurs creating their own brochureware. We are creating and exposing huge amounts of social data. Data that allows us to discover subjective information and sources that in 1973 could really only be obtained from people with whom you had traditional social relationships, like Marie, Aunt Helen and your second best friend.
In the Horizon Report 2008 , one of the trends mentioned as ” likely to have a large impact on teaching, learning, or creative expression within learning-focused organizations”is social operating systems. This is data stored and retrieved not according to what the file contains, but how the creators and users of the data relate socially.
Take last.fm, for example. It promotes itself as:
Last.fm connects you with your favorite music, and uses your unique taste to find new music, people, and concerts you’ll like.
When I am logged into last.fm, if I play music tracks from any source on my PC, they are “scrobbled” - or collected by last.fm. It builds up a set of data that associates me as a user with an awful lot of Crowded House, Elvis Costello, Peggy Seeger and the Beatles. With enough users doing this it creates a collaborative filtering - people who like Madness generally like Squeeze. When I am after something to listen to, I can see what other people with similar musical tastes listen to and take that as a recommendation.
I understand this bit vaguely, but last.fm can take it further than that. Using a Friend of a Friend (FOAF) URI, your last.fm identity, your relationships to friends there, your favourite music tracks can all be exposed to other applications like Facebook and Flickr…allowing all sorts of re-mixing and mashing up.
The type of interoperability being offered by last.fm, the creation of new relationships between data in different systems is what underpins the semantic web.
As I understand it, the semantic web seeks to harvest all this subjective data, and objective data we put up on the web, find out patterns and relationships between them and then use these patterns and relationships to better answer your query.
In all this, one thing becomes essential - some nodes of stability. Places to datamatch. How do I know that the Guns N’ Roses track featured in last.fm is the same as the Guns N’ Roses track mentioned in the Internet Movie Database as a soundtrack ? With way over 100 ways to write the track Guns N’ Roses - Knocking on Heaven’s Door , where could we find authoritative source data for this track?
Yup - it’s library core business, our authority files. It’s the pride and joy of cataloguers. And it’s been used as an in-house resource forever. What if we got it out there and opened it up to other applications as data that can anchor datamatches across social operating systems?
This is happening to author authority data via sites like OCLC’s Worldcat Identities and the National Library of Australia’s People Australia .
At the recent VALA conference, Stuart Weibel from OCLC announced that they were going to attempt to anchor works (in the FRBR sense) to have a static, identifiable web address via a project called Global Library Manifestation Identities - Glimir - another way library work can provide anchors within a social operating system.
There’s a problem. Library standards are not same as the standards being developed internationally to allow interchange of data. We speak MARC, they speak RDF.
If you want to follow up moves to bring these standards together, read the blogs of some people who actually understand it all - Ed Summers, Karen Coyle or John Phipps .
The image below shows datasets published in the Linking Open Data community project , based in the W3C consortium, headed by Sir Tim Berners-Lee . These are datasets from organisations that undertake to allow them to be opened up and linked out. More about the project is here, How to publish linked data on the web.
I wonder whether one day we’ll see library data as one of the nodes? I wish.
Source: The Linking Open Data datasets cloud maintained by Richard Cyganiak. .
