Re: Tag ontology RFC

Ah good, I'm glad you're of the same opinion! That's a big 
encouragement. :)

I second your reservations about Items, which I've already expounded on 
in an off-list email to Seth. "Temporary inference" seems like quite a 
canny idea to forge this kind of interoperability! What I suggested to 
Seth this evening was sort of the other way around:

[x a rss:Item] && [x from <http://del.icio.us/>] => [x a tags:Tag] .

i.e. you exploit some additional knowledge about the sources you want 
to integrate (in this case that you know that del.icio.us's feed 
contains tags) to generate the integrating triples.

I really have to get some code done for this... perhaps a little Easter 
project for myself. Everyone seems to be releasing del.icio.us-alikes 
recently (I'm thinking of del.irio.us)!

-R

On Mar 25, 2005, at 21:18, Danny Ayers wrote:
> Whatever, looking again it not only makes sense, but also
> should lend itself to fairly straightforward implementation and
> support nice interop with class-oriented inference. With a bit of luck
> ;-)
>
> Seth:
>>> I would say that Tagging is a sub class of Items.
>
> Hmm, I've wondered about reusing Items in a few place, I suspect it's
> a mixed blessing - it makes syndication a doddle but then if you need
> to separate non-tag Items for whatever reason, it won't be such fun.
> For an idea I'm playing with at the moment I've got to-do items (class
> ToDo), which will mostly comprise a resource with a dc:description (or
> somesuch). I'm planning on leaving them more or less that until I want
> to do things like syndication. At which time I'm hoping to insert a
> triple into the local model (which will already contain Items) saying
> that ToDo is a subclass of Item, infer their itemsness and pull a
> combined feed out from there. Thing is, in every other context that
> schema-ing triple won't be around, so querying should be easier. Might
> work ;-)

Received on Saturday, 26 March 2005 00:27:33 UTC