[rdfweb-dev] tube map

[jo walsh]

hello,

> > (if anyone comes across data describing the connections & distances between
> > station, please let me know - I want to play with graph algorithms too...)
> 
> There's some interesting stuff on Jo Walsh's spacenamespace site
> (http://space.frot.org/), including a description of the entire network in
> ... can advise - Jo?

yes i did a couple different RDF renderings of the tube map. the site is
on slightly flaky DSL, sorry; i copied the tube stuff here:

http://uk.space.frot.org/rdf/tube.rdf
is the original simple model where all the stations are connected to a
'line' node so you can see what is going on but it is no use for graph
traversal routefinding

http://uk.space.frot.org/rdf/tube_model2.rdf
is the second attempt where all the stations are connected to the nearest
one either side of it on each line that passes through it; and the
information about the lines is stated about the reification of each
connection. that's probably an overcomplex approach i wouldnt use again

http://uk.space.frot.org/rdf/tubemap.svg
map of the tube generated from the second model, with HREFs to rdf uris
still using the first model...

http://uk.space.frot.org/rdf/tube.pl
some snippets of prolog which may or may not perform graph pathfinding on
that tube_model2.rdf. the depth-first traversal which prolog is built out
of too easily goes off on random irrelevant routes. i experiemented with
a few breadth-first shortest-path goals but my prolog is not strong enough
to get them working, i am still a learner shame shame. 
     
fwiw i used the same technique for foaf codepiction and foaf
friend-to-friend routefinding fairly successfully, it is nice to see
problem domains collapse like that

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