Re: [Information Gathering] next steps: syndication, good weblocation

On 31/03/07, Lee Feigenbaum <feigenbl@us.ibm.com> wrote:

> I must apologize for the fact that I haven't found the time and energy to
> speak up before now (especially after volunteering for the IG task), but I
> do think it would be a big mistake for SWEO not to proceed further along
> this path. I think that the aggregated data sources is a great first step,
> but without augmenting it with some way to identify the best-of-breed
> resources and to attach facets to various resources (what type of
> resource, what level of fmailiarity is assumed, what type of audience, is
> the resource slanted to a particular industry) and then without exposing
> this in a user-friendly fashion, I think we've actually accomplished very
> little.

I don't really disagree, but do feel that ideally the user-friendly
presentation should be left out of scope until there's good data to
present. (Sure, work on the presentation is likely to throw up issues
with the data, calling for further cycles. But whichever way you look
at it, there has to be something to present...)

> I will try my best to put my time where my words are, and--if anyone else
> agrees with me on this, of course--I'll be happy to spearhead an effort to
> make use of this data in the way that I see fit. While a single individual
> can produce the type of user-interface I'm picturing (and even generate
> the classification data), the ranking bit of my vision requires the
> agreement and participation of a larger group of us.

Ok, my attitude on this has been that given sufficiently interesting
data, a single individual could produce a reasonable UI in a
relatively short time. (And in the unlikely case that no-one else
emerged to do it I'd darn well do it myself). But if there are things
like ranking which call for wider participation, that's different.
Care to provide some specifics on the kind of issues you have in mind?
(In another thread I s'pose).

> Big thumbs-up to this point. SPARQL + exhibit is exactly the type of
> approach I'm picturing to building an interface for this data. (Though
> I've been playing with exhibit recently, and it may not be exactly what
> we're looking for. It's very close, though.)

Have you played with Longwell? [I get deja vu asking, so I expect your
answer will be "yes, but not recently"] That'd be my own first
candidate for a UI - does similar stuff to Exhibit but can scale up,
is server-based. You get not bad rendering & nav by default just
dumping RDF/XML in a dir for it, with a bit of config tweaking to
match it up with the vocabs, I think it could get really sweet.

Cheers,
Danny.

-- 

http://dannyayers.com

Received on Saturday, 31 March 2007 08:49:39 UTC