TypePad foaf:knows Agent? (was RE: [rdfweb-dev] Plink)

[Victor Lindesay]

Danny Ayers wrote: 
>With due respect to the good Dr Ont, I don't think TypePad's FOAF is
all
> that bad, I certainly don't think it'll take a team to fix it.
> ..... 
> *But*, currently the spec says the range and domain of foaf:knows is
> foaf:Person. I think perhaps this is too narrow - we've already taken
> "knows" down to a generic (human) relationship, would it be 
> too much of a stretch to make it apply more generally to agents?

I have discussed the TypePad FOAF issue with Dr Ont and he would as
always welcome any adjustment to the FOAF schema resulting from valid
use cases and the requirements of developers producing or working
against this data format. He is easily pleased and is only concerned
that instance documents, especially those originating from influential
producers, reflect the current spec to avoid the whole thing going
pear-shaped. A typical schema head!

As a hands-on developer who dabbles in the FOAFware space, I would not
like the domain or range of foaf:knows to be changed to foaf:Agent
because of the many extra lines of code (even using an RDF parser) to
accommodate this major semantic shift, that people can 'know' things or
that things can 'know' things or people. 

I vote for keeping the current foaf:knows definition the same and do
blogrolls with extensions. Let's keep the semantics simple and cost of
entry low (in hours and brain power) for FOAFware in order to promote
practical implementation by RDF new wavers.

Perhaps the man (sorry person) of the moment Dom Ramsey, creator of the
magnificent and soon to be Slashdotted Plink would care to comment. Is
he on the list?

And respect to Morten with his equally valid and interesting approach of
mining data within <foaf:Person/> elements to connect people and extract
value giving information. This demonstrates the beauty of webby data
perfectly. Data delivered in a decent RDF/XML or XML format can be used
in ways that are un-imaginable to the producers.