Re: ProductDB

On Fri, 2009-08-14 at 15:17 +0200, Martin Hepp (UniBW) wrote:
> Note that the GoodRelations "seeks" patterns allows using the full 
> amount of details and the same vocabulary for specifying wish lists
> you can say that you are interested in TV sets with at least 11 inches
> of screen size etc. So it is not only a simple wish list, but allows
> using all features of GoodRelations for the buy side - with the simple
> difference of using gr:seeks instead of gr:offers between the 
> gr:BusinessEntity and the gr:Offering.

Aha. I'm not surprised that GoodRelations provides such a relation -
it's fairly comprehensive.

I'm be a little concerned over it's complexity for common FOAF-like use
cases though. There's quite a lot of indirection to get from the
foaf:Person to the product they want - in GoodRelations, a person would
seek an Offering, which includes a TypeAndQuantityNode which has a type
of good which is the end product.

So:

<#me> gr:seeks [ gr:includesObject [ gr:typeOfGood <#product> ] ] .

Really I'd just want something like:

<#me> ex:seeks <#product> .

Yes, that's not quite as expressive - I can't say how much I'd be
willing to pay for it, or how many I want, or in what timeframe I want
to buy it, but it's enough for a simple wishlist. ex:seeks could of
course be defined in terms of gr:seeks, so that mapping could be done in
both directions, but I do think that a simpler term is needed.

Then a wishlist becomes as easy as:

	<> a sioctypes:WishList ;
	   foaf:maker <#me> .
	<#me> ex:seeks <#product1> , <#product2> , <#product3> .

Ian's barter vocab is closer to this, though I think it could be
improved by specifying domains and ranges for the terms and generally
defining them a little less loosely.

-- 
Toby A Inkster
<mailto:mail@tobyinkster.co.uk>
<http://tobyinkster.co.uk>

Received on Friday, 14 August 2009 15:33:52 UTC