RDFMap, a proposed RSS 1.0 module for mapping
- I've put together a draft of a simple RSS 1.0 module for mapping. It defines classes and properties which express the locations of resources, and how they should be depicted on a map. The spec is called "RDFMap" and can be found at www.mapbureau.com/rdfmapAn applicaton based on RDFMap for mapping the contents of weblogs can be found at www.blogmapper.com This app is a stab at bringing the ability not only to view but to create online maps to the web public via a simple extension to weblogging technology. Blogmapper was concieved and implemented in collaboration with Jason Harlan (harlan@...)Comments on the spec (or the application, for that matter) are solicited.Chris Goad
- Doesn't the Dublin Core qualifiers already cover geographic positions?
I've suggested that feeds embed metadata into their 'main' HTML page that
indicates their geographic positioning. This as opposed to stuffing it into the
feeds themselves. Which, if they desired doing so, could be well implemented
with DCMI Point elements using DC modifiers. Not to say you're this isn't a
good suggestion; I like the idea and believe the more the merrier.
I'm open to discussion on how to expand the use of geographic data into more
sites/feeds/blogs. I've been dragging my feet on implementing it on Syndic8 but
the infrastructure for it is already present. I've applied geographic server
location metadata for the feeds already. It's based on NetGeo detection but
it's better than nothing. Official, user supplied, geographic positioning seems
like a better idea but I've got only a small portion of the work completed on a
user metadata builder.
----- Original Message -----
I've put together a draft of a simple RSS 1.0 module for mapping. It defines
classes and properties which express the locations of resources, and how they
should be depicted on a map. The spec is called "RDFMap" and can be found at
An applicaton based on RDFMap for mapping the contents of weblogs can be found
at www.blogmapper.com This app is a stab at bringing the ability not only to
view but to create online maps to the web public via a simple extension to
weblogging technology. Blogmapper was concieved and implemented in
collaboration with Jason Harlan (harlan@...)
Comments on the spec (or the application, for that matter) are solicited.
- -----BEGIN PGP SIGNED MESSAGE-----
"Chris Goad" <cg@...> writes:
> I've put together a draft of a simple RSS 1.0 module for mapping. It
> defines classes and properties which express the locations of resources,
> and how they should be depicted on a map. The spec is called "RDFMap" and
> can be found at www.mapbureau.com/rdfmap
Have you thought about using this data to provide warchalking information but
without having to graffiti the street? I notice that all my warchalkings in SF
have been erased within 24 hours.
If you made these into RSS entries one could aggregate the data and build a
piece of hardware (or with a PDA) to view close hotspots.
Kevin A. Burton ( burton@..., burton@..., burton@... )
Location - San Francisco, CA, Cell - 415.595.9965
Jabber - burtonator@..., Web - http://www.peerfear.org/
GPG fingerprint: 4D20 40A0 C734 307E C7B4 DCAA 0303 3AC5 BD9D 7C4D
IRC - openprojects.net #infoanarchy | #p2p-hackers | #reptile
I know not what course others may take but as for me: give me liberty or give
- Patrick Henry
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.7 (GNU/Linux)
Comment: Get my public key at: http://relativity.yi.org/pgpkey.txt
-----END PGP SIGNATURE-----
> If you made these into RSS entries one could aggregate the data and build aOr make an Web Services call into Syndic8 with your current lat/long and request
> piece of hardware (or with a PDA) to view close hotspots.
a list of feeds within a given radius.
Sort of like this preliminary page:
It's only based on the NetGeo server IP lat/long.
On Thu 25-Jul-2002 at 05:36:37PM -0400, Bill Kearney wrote:
> Doesn't the Dublin Core qualifiers already cover geographic
Sure the qualifier for dc:coverage, dcterms:spatial  can be used
with 4 types of spatial encoding  including DCMI Box  and DCMI
- --- In rss-dev@y..., "Bill Kearney" <wkearney99@h...> wrote:
> Doesn't the Dublin Core qualifiers already cover geographic
>Yes. rdfmap:location is indeed a refinement of dc:coverage, and
rdfmap:Point and rdfmap:Box obviously have similar intent to DCMI
Point and Box. However, there are a couple of issues. DCMI Point
and Box define positions via the properties "east" and "north"
instead of "x" and "y". This actually can cause real confusion in
some cases, because the coordinates of a point derive their physical
meaning from the projection specified, and not all projections
permit the coordinates to be interpreted as latitude and longitude.
For example, the various projections for the polar regions (gnomonic,
polar steregraphic ..) don't allow this interpretation. A more common
instance of this is the use of coordinates relative to maps of
unknown orientation (which our spec permits). For example, I can
use a Jpeg of my garden as a map this way, without saying or perhaps
knowing where north points. It is not just a little confusing to be
forced to call the coordinates "east" and "north" in these cases.
Another issue is that the dominant open spec for geography, GML, uses
x and y for coordinates, and lets the projection carry orientation
information. It might be more important to adhere to GMLs
conventions, which are widely used in the relevant discipline, than
to Dublin Core.
Hence the decision to define Points and Boxes using differently named
properties. Is this enough to put us out of compliance with DCMI
recommendations? Should we use the tags dc:Point, dc:Box if our
properties are different?
Also, RDFMap covers one additional class needed to actually go ahead
and do mapping of web resources (as at www.blogmapper.com): the class
rdfmap:Map, which specifies the step from projected coordinates to
coordinates on a map in SVG, Swf, Jpeg or whatever.
A more general point/question:
I think that mapping web resources in a way that fits in with RSS is
a good idea for many applications. (I'm glad to learn that syndic8 is
doing some things in this area too). It seemed to me that a minimal
vocabulary supporting this should be put together and formalized.
Parts of such a vocabulary are of course scattered about (eg in GML
Dublin core, Geo-tags), but not in a form that could quite serve the
purpose. Proposing an RSS module seemed a reasonable approach.
Scraping other kinds of data (I'm thinking particularly of geo-tags)
into the module format, whatever it is, will be easy and useful.
Comments? Anyone out there with ideas about what's needed to make a
module of this kind work well for your current or anticipated
applications? Is a module of this kind even the right approach?
Meanwhile, I'm working on an RSS 0.9x variant, which will be ready
- I've changed RDFMap spec into a form with two variants: a "plain" XML variant compatible with RSS 0.9x, and an RDF variant compatible with RSS 1.0 (the latter is identical to the spec mentioned in earlier posts). Unfortunately, this change means a new name is needed (already!), since RDF is now optional. RSSMap is the new name. It can be found at http://www.mapbureau.com/rssmap The implementation at http://www.blogmapper.com supports both variants.Comments?Chris Goad