metadata about a big repository of geodata

[Jo Walsh]

dear geodata discussion,

I have felt really psyched to hear of the resources that John Graham
is offering for use by OSGeo which in the meantime will be at
http://osgeo.telascience.org/ . 
https://mail.osgeo.org/servlets/ReadMsg?list=discuss&msgNo=350 
is Mark Lucas' description of what is going on with this.

John is already hosting a mirror of the data service at
http://onearth.jpl.nasa.gov/ , see also
http://onearth.jpl.nasa.gov/WK/bmng/ , Landsat7, Blue Marble, has some
ASTER data too. He has WMS interfaces to a lot of public geodata which
are available but not very discoverable.

Norm Vine raised the subject of metadata, indexes, searches, for
geodata, on IRC today.

I would be interested to hear conversation about what peoples' needs are to 
express metadata about geodata, and when they are searching for information
how would they like to find it, what they are offering now and what interfaces 
are missing/incomplete, what kind of open standards there are that
exist that people actually use. 

Over the next couple of weeks this group should come to offer a set of
collective suggestions about what kinds of data can be hosted at
telascience and what kinds of access / discovery services can support
that offering, for discussion on the discuss list and then for Norm Vine 
to offer to John Graham as a statement of requests and ideals from OSGeo.

I have an RDF shaped hammer and i wanted to share some related
material in the RDF domain. Dublin Core is one vocabulary which gets
touted, gets use but it's very loose and woolly, doesn't allow you to
predict and reuse very much about what information you will get in its
properties. Other more project specific RDF vocabularies have
'subclassed' terms from it and from online lexicons like
http://xmlns.com/2001/08/wordnet/ 

The other day i heard a interesting suggestion on #openguides on
irc.perl.org,
15:18 <@ivorw> Unless anyone objects, I will use <foaf:depiction> for
map links

http://openguides.org/ is a "spatial wiki" project that some friends
of mine started in London; the busiest one now i think is
http://boston.openguides.org/ which is being run and enhanced by Chris
Schmidt with a lot of Google Maps based, for now, extensions.

Anyway my point in raising this, is that a feature of openguides is
that it provides fairly detailed RDF feeds of metadata about places:
http://boston.openguides.org/?id=The_Field;format=rdf for example.

So FOAF is the 'Friend of a Friend' vocabulary that defines a lot of
common statements about properties of people, especially in how they
are connected to information.
http://xmlns.com/foaf/0.1/#term_depiction
states simply, "The foaf:depiction property is a relationship between
a thing and an foaf:Image that depicts it." I wondered how this fits
into the ontology of maps; as Rich says, (I think):
"Maps shape and create reality".

http://space.frot.org/ontology.html has a short collection of links to
work on RDF "ontologies" to describe things in space. But i don't
really know of work in that field to provide metadata about spatial
things - descriptions of the descriptions. When i started in on openstreetmap
I wrote http://uo.space.frot.org/?MapOntology when thinking about
describing properties of vector data (in the end OSM went for a very
open, 'tag' based approach and encourage the development of a
vocabulary through advice on the wiki. We never thought about
descriptions of raster data done in the same way, but that could make
a lot of sense.
 

jo