[rdfweb-dev] FOAF-a-matic Mark 2

[jo walsh]

> So I was thinking of this cool RDF representation of UK and London by Jo
> Walsh:
> 
> http://space.frot.org/mudengland.html

aw thanks ;) in terms of useful xml namespace info and schema design there
is not anything there really. i had two sets of geographical data at
different scales, with different kinds of concepts - abstract places and
concrete physical spaces - perhaps that is difficult to resolve
philosophically and it would be better to keep everything in the same
model. the UK model is huge though, 300000+ triples in a big sql database

anyway i made up two different xml namespaces rather slapdash, as i wanted
to express structurally quite different things about connection and
containment. but authoritative stuff like latitude mearurement is
duplicated in both namespaces and even see the same kind of value named
differently: space:wgs_lat and place:dd_lat are the same measurement of
latitude, varying because that was the description at the source. 

there is a lot of uk specifics, i have data for landranger grid, and
ordnance survey x and y co-ordinates where i can get 'em. then different
countries' postal codes can be quite semantically different - a uk or
dutch postcode refers to a house or block, a spanish one covers a
small town or city district. 

this might suggest a locale-like scheme for different countries' metadata 
could be appropriate. i don't know if such a thing exists or how
important it is to be standardised. i would like my bot and your bot to be
able to reach a shared consensus that dd_lat and wgt_lat point to the
same thing. i suppose this is expressible in daml, maybe looking for a lot
of complexity. 

i would probably rather use a reliable third-party namespace to express
basic geodata if there was one available, or set one up on xmlns.com maybe

zx
--
"Common sense won't tell you. We have to tell each other." -DNA