Reading, again (and a metalesson in its perils)

A very odd thing happened to me on my way to writing what I now realise to be a a completely unrelated post about user-generated content…

Two exciting things came to my attention; Nate Hill linked me to Aaron Schmidt, specifically his social database mockup, my reaction to which got bogged down in first hearing about, and then waiting to hear more about, Talis’ Project Xiphos, now finally better explained better over on their Panlibus blog under the explanatory title “Silos Silos Silos”. There’s a lot to both of them, but broadly the former is a social information site that creates a set off tools for users to databases that, in turn, creates a level of metadata for other users, and the latter is a semweb set of technologies for opening up multiple datasets to, amongst other things, radically alter the currently insanely annoying experience of federated information access regimes.

What’s surprising to me is that the initiative to make data interoperable is explicitly courseware, and the one to make it social is implicitly (and slightly reluctantly) so. You’d except that to be the other way around, which to me firms up the idea that libraries host a specific kind of information use, that they served and encouraged by topic-based physical browsing, that they need to now better support with new, non-physical (well, non-linearly-physical, or something) technologies. (TANGENTALLY RELATED ASIDE: last night Stan Cohen, a bright button, said that Google held 338,000 entries on “climate change denial” as if this meant it contained 338,000 items denying climate change. Of course, it does not. It contains 338,000 items discussing the denial of climate change. There’s something about search as the primary form of access that implies you’re finding exactly that encourages a bias in favour of certain desired biases, I feel. He wouldn’t go to 324.243 and become outraged at the quantity of fascist literature encroaching on academia, would he?)

Final point, (WARNING: .pdf) the JISC & SCONUL Library Management Study reports that “[t]he ability to aggregate user behaviour has significant potential for discovery services, based on click streams, context and personalisation. Nevertheless libraries are not yet exploiting intelligence about user habits to enhance their position in the information value chain.” Well, yes.

Filed under: classification systems, Information use, scholarly communication, Semantic Web, User experiences, data silos, databases, social software

What’s happening with all this “What’s happening with the Semantic Web?”?

Blimey! The Semantic Web’s made it all the way up my RSS reader and in to Svia Vaidhyanathan’s The Googlization of Everything, repping a Digital Eccentric post which is functionally identical to this one on Fiona Bradley’s The Semantic Library.

They’re two nice articles, and seem to be pretty good at telling people what’s going on (I assume), but it was still a surprise to see it on Googlization. I’d like to know what Vaidhyanathan meant by it: it might be that the tech is going to power the next go-to search engine and eat the big-G’s market-share before cleanly replicating it, but the more exciting prospect is, as Danny Ayers says in the Nodalities post, it’s not going to refine search, really. With enough sources, your terms probably can brute-force relevant results, because the question “How do I bake bread?” is always going to return results with the words “how” “bake” and “bread” in them, if the engine knows what “baking” means or not. The real thing about search is you only look at so many results before you actually read them, and from there you go and look at things those people are talking about.

Vaidhyanathan’s Critical Information Studies has a healthy dose of scholarly communication theorising bound up in it, and what with the Vaidhyanathan-affiliated Institute for the Future of the Book moving formally into research in this area, I was kind of hoping he’d tackle something about the actual nature of search and what it means for information retrieval.

Searching immediately brings to mind finding the right thing, which is presumably why there’s an I’m Feeling Lucky button. But how many people use that? How many people are taken to the best possible result through doing that? In how many instances is there a best possible result? The real joy of the almost-no-finite-resources-used non-tree model of publishing means single scraps of data can be produced, found and consumed (that’s basically the form of the Googlization blog, and many others), the ease with which they can be linked to other comparable resources making the assemblage as a whole at least if not more valuable then a single resource, no matter how detailed or well-referenced.

Browsing around from one resource to another, as opposed to reading every entry off a single search results page or making constant new ones is what Semantic Web tech like RDF promises, introducing interoperability between information resources will allow reading and writing in a networked environment to finally realise what reading and writing in academic (and other discursive, albeit largely non-conversational) forms has always striven for, is probably the real reason why people who read things other people say should be excited.

At the very least, it might give us hyperlink with tones of voice. Nothing is worth anything if it’s not a vector for sarcasm.

Filed under: e-text, Semantic Web