jo walsh zool at f...
Wed Oct 30 13:54:20 UTC 2002

hey rdfwebbers,

state51, the people i used to work for, run a site called,
which is like a collaborative guide to britain, it was originally a
skateboarding magazine, it has been on the web in a form rather like its
current since 1994. it has a few thousand active users out of some 30
thousand registered, with message boards and chat rooms and the like. but
for legal reasons the postings to the town guides have to be moderated,
and there's an immense backlog for one admin person.

there's a central clique of trusted users who provide a lot of feedback,
though; and we'd thought of using a foaf network to express both
decentralised moderation, and filtered views. so state51 could offer a
dozen or two users moderation rights for particular locations, and they
could offer such rights to their friends, etc, in webs of trust 

the one could choose to filter the view of a site by the recommendations
of particular people or foaf networks of people, and then tentatively 
offer an unfiltered view of the 100,000 odd back entries with heavy
provisos, given the acceptable alternatives.

it could be good to offer a web-based service to people who might choose
to publicise or otherwise use their foaf data but don't have shell
accounts, clue regarding gpg, etc. i dont know how desirable it would be
to centralise too much of this data; a few thousand active entries. i'd
imagined something like a keyserver, for foaf data. i'd feel nervous about
storing what amounted to peoples identities, though.

it any case, we'd had thoughts of providing a foaf-based backend service
for the site and some kind of RESTlike interface to the data, which is all
location-based so i suppose that goes in a separate namespace. any
suggestions there would be appreciated. 

the reviews are all very free text though and there is little structure or
precision in the whereabouts of things. i have been making a little rdf
model of london which is all connection- and grid-location- based but the
vaguer and more descriptive information in knowhere - implicit local
areas, 'hook-up spots', are a lot harder to model. 

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

----- End forwarded message -----

More information about the foaf-dev mailing list