[standards-jig] JEP for Location Based Services

Matt Mankins mankins at media.mit.edu
Thu Feb 6 22:17:47 UTC 2003


I've been writing an application that links Geography with Information
Nuggets.  I call it Location Linked Information.

Here's a presentation I did about it.  It's got some extra junk in it, but
might give you an idea of what I'm doing with XMPP and Location Based
Services.  I should have the server component available for downloading


(perhaps this is the wrong list for this...but I can give you more info
if you want. )

[asynchronous discovery]

Oh, since I'm writing, I might as well mention that it uses asynchronous
discovery, a combination of the JEP 0004 and JEP 0060.  I don't know if
it's totally "legal", but I find it quite useful...It works like this:

1) C: I want a search form.

2) S: Ok, here's a search form.

3) C: Here's my completed search form.

4) S: Ok, I've created a generic node "123" for the results of the search,
and you're subscribed to that node.  Others will publish results to that

5) Then the server (component) either finds out the answer to the search
and publishes the answer itself to the generic node it created (which then
gets sent to the client who asked for the info in the first place), or
sends the search out to other clients/servers who can then publish their
results back to the generic node.  These other helpers in the search don't
need to subscribe to the generic node, but rather the generic node is open
for all to publish to.

6) After some period of time the server reaps the generic node.

So the result is that a search can get sent around the net and information
gets sent to the client that wants to know about it, but rather than
having to come in all at once, it can trickle in as it's available.


Matt Mankins

On Thu, 6 Feb 2003, Stephen Pendleton wrote:

> Has anyone been working on a JEP for the integration Location Based Services
> with Jabber? In a general simplest sense, this means giving Jabber users
> information on the geographical location of other users.
> Thanks

More information about the Standards mailing list