- From: Yves Raimond <yves.raimond@gmail.com>
- Date: Wed, 9 Jul 2008 12:42:11 +0100
- To: "Dickinson, Ian J. (HP Labs, Bristol, UK)" <ian.dickinson@hp.com>
- Cc: "public-lod@w3.org" <public-lod@w3.org>
Hello! >> Actually, I often use this one along with other >> "things-that-we-can-do-now" music-related use-cases, as I >> find people tends to like it: anyone would have struggled >> with its iTunes (or >> whatever) library at least once. I often find that use cases >> work better when it relates to things the public experienced >> in the past. > Agreed. I think that finding music I might like *is* a good use case, and one that people can relate to. What I don't see is how geographical location helps. Where I live there are a bunch of local musicians living, and doubtless some famous-ish ones were born nearby, but I don't like or dislike their music because of that. I also like Arthur C. Clarke's fiction, which has nothing to do with the fact that Minehead, where he was born, is less than 100 miles from my current location. So yes to music selection as a use case, yes to helping find stuff I might like through semantic annotations/LOD, but I'm not yet convinced by the geo location angle. > I totally agree with you. I was actually more thinking of these geolocation queries against an aggregation describing my personal music collection (using the sort of stuff in the motools sourceforge project), so I kind of know this is music I like already. I typically use such geo-location queries for a "plot my music collection on a map" use-case (as in http://dbtune.org/facet-demo/), which I found people tend to like as a semweb use-case (much easier to create a playlist of Cuban music, for example, by directly seeing the part of my music collection which was recorded by Cuban artists). I think the fact that you can also categorise your collection between "rural" or "urban" by using associated statistics is also quite interesting. Not on a really "find me music i like" use-case, but more on a "give me more information about the music i like". Also, I forgot to mention DBpediaMobile. I found this has a really great "wow" factor for people not knowing anything about the semweb. Associated with last.fm recommended events... yummy :-) >> Big sci-fi use-cases tend to work a bit less. > For sure. > >> I had the feeling the journalist *was* actually impressed by >> this use-case, btw? > Well, I heard, after Tim explained the music<->geohash mashup idea: > > "Ha ha ha. Slightly recherché things to want to do!" > > re·cher·ché > 1. sought out with care. > 2. very rare, exotic, or choice; arcane; obscure. > 3. of studied refinement or elegance; precious; affected; pretentious. > > [1] > Heh - sounds like a French word :-) But you're right, I didn't notice it. I still don't really think it is bad, actually. I guess that, with such a "recherché" example involving such a simple linkage, most people are left wondering what else could be possible by interlinking data in different domains? Cheers! y
Received on Wednesday, 9 July 2008 11:42:48 UTC