Re: [Geowanking] Open Source Geocoder Released Today

Tue, 25 Apr 2006 13:44:41 -0700

We have a Java-based Tiger/LINE geocoder (operating within our EJB3/J2EE based environment, but it could be refactored to be non-J2EE), and I have long thought that it might be better placed into open-source.  It has some internationalization hooks, but we would like to take that further.  Please let me know if you are currently working on or contemplating a Java geocoder, and would be able to contribute to this work.

Dan R. Greening, Ph.D.,  CEO BigTribe Corporation,  http://dan.greening.name/contact.htm



On Apr 25, 2006, at 1:22 PM, David Bitner wrote:

Can't we all get along :)

Here are the main distinctions that I see between different geocoders:

TIGER vs. Other usually Shapefile derived roads
Database vs. Flat File Indexes
Batch processing vs. One off or web
Interpolation between segment ends vs. Along the line
Hand parsed addresses vs. automatic parsing

What pieces can we build in common?  What can we do to cross some of
these lines?  This is clearly something that a lot of people need and
there are a number of folks pulling in different directions.

There is a mailing list at the aforementioned http://avoir.uwc.ac.za/
that has been silent for quite a while that could be used for this
discussion.

Would anyone be interested in having a mini-forum for geocoding over IRC?

There will of course be different needs that will necessitate
different approaches due to speed/scalability/accuracy, but there are
a number of pieces (ie address parsing) that it would definitely
behoove us to collaborate and find some good solutions that we can
refine together.

David

On 4/25/06, Paul Ramsey <[EMAIL PROTECTED]> wrote:
Also there is <http://dotproject.sauder.ubc.ca/pagc/> this one from UBC.

P

On Apr 25, 2006, at 7:22 AM, David Bitner wrote:

There is also the geocoder project at http://avoir.uwc.ac.za/
projects/geocoder
that has stalled for quite a while now that hopes to create a
configurable geocoder to use any appropriate user inputed data.

On 4/24/06, Bill Thoen <[EMAIL PROTECTED]> wrote:
On Mon, Apr 24, 2006 at 08:17:42PM -0700, Rich Gibson wrote:
geocoder.us (which Schuyler wrote, and I run) is also open source,
written in Perl, and using the Tiger data.

If you get SRC's up on a public web service let me know, I'd love to
do some tests to see if they are handling things that geocoder.us
doesn't.

You mean the one at http://geocoder.us? Yes, that's pretty good
too. I
whined about it not being able to find my address (on 2308 South
St.) and
within a week, somebody (you?) fixed the code and then it could
find me. Gotta
love open source! Problems gets fixed fast. We should definitely
try to
test the SRC tool in the forge of public scrutiny, too!

But there's others too. Daniel Egnor's C-based geocoder (see
http://ofb.net/~egnor/google.html) is worth a look if you can't
deal with
Perl (which I got into before I learned some Perl), and Philip
Holmstrad's
"The Geocoding Blog" (http://batchgeocode.blogspot.com/), the
"voice" of
the (free) Batch Geocoding site at www.batchgeocode.com/ needs a
mention
too. Are there any others in the geocoding game?

I'm sort of curious to determine where the state of the art is at
this
point. I use a fairly expensive commercial geocoder now (MapInfo
MapMarker)
which is rather good, and scales well, but I'd be happy as a pig in a
wallow to find an open source solution just as good that didn't cost
quite so much.
_______________________________________________
Geowanking mailing list

_______________________________________________
Geowanking mailing list

_______________________________________________
Geowanking mailing list

_______________________________________________
Geowanking mailing list

_______________________________________________
Geowanking mailing list
[email protected]
http://lists.burri.to/mailman/listinfo/geowanking

Reply via email to