My guest for this week’s Innovators show is Mike Dunn, a veteran media technologist who recently attended, and spoke at, the 2009 Semantic Technology. Mike and I were both impressed by Tom Tague’s keynote talk, which avoided theory and focused on practical ways that here-and-now semantic technologies are helping media businesses work smarter and more profitably. In this conversation, Mike describes some of the ways that his company, Hearst Media Interactive, is proving that point.
Search engine optimization is currently one of the best ways to profit from data-enabled content. Meanwhile, one of the expected benefits of semantic technology — better search recall and precision — hasn’t materialized. But although most users may not care about querying archives more comprehensively and more precisely, writers and editors should. And not only because it helps automate the assembly of context around a current story. If you can review an archive in a precise and comprehensive way, you can do a better job of planning future stories that acknowledge — and advance — the ones you’ve already done.
Hi Jon, you might be interested in this discussion of the merits of the precision/recall metrics.
http://thenoisychannel.com/2009/07/17/in-defense-of-recall/
You write:
“the idea that a single search will be used to fully address an information seeking problem is simplistic.”
Agreed. When I reflect on my own search strategies (http://delicious.com/judell/searchstrategy+jonudell) they are invariably multi-step procedures.
That’s why, actually, I think there’s a ton of headroom to improve online editorial products — and maybe even turn them into services I’d pay for.
Just to clarify, those are my words, though I hope Neil agrees with them too! I certainly appreciate the link!
Semantic technology is a big umbrella – not a stretch to suggest Google/LinkRank sits under it. But regarding Semantic Web technology (in the W3C sense), I’d suggest it implies a shift in emphasis regarding search. Rather than waste cycles hunting for stuff, it’s better not to lose it in the first place :)
This guest would rock.
http://www.wicn.org/audio/inquiry-melanie-mitchell
looking through this book now.
Thanks so much for the pointer! I’ve queued up that MP3, really looking forward to it.
Her Copycat project (with Douglas Hofstadter, http://portal.acm.org/citation.cfm?id=218767) remains fresh in my mind 15 years after I first read about it in Fluid Concepts and Creative Analogies.
> it’s better not to lose it in the first
> place
That’s a great way to put it!
Of course if I learn that search will reliably find a thing, even if I do not mark up the thing, then it is arguably never lost, and never needs to be found, it only needs to be retrieved.
I wonder how many items in your personal archive, or mine, or anyone’s, have this property: the combination of one’s name, plus a set of search terms, reliably yields a singular result.
from search perspective, marking up (semtech) should allow more granularity in search retrieval down to entities within pages generated systemically – the hope at least :)