Presence and IntComm WG

[Jeremy George <>]

Deke.

  Since Jamey may not be in today and I know he's somewhat pressed for
time, I'll try a non-authoritative answer.

> How many skiffs will we need to do the job at the conference hotel?

  Given the time contraints for this first demo we may need to limit
the area in which we provide location services.  As I understand it,
all the beacons (skiff & iPAQ) that exist are in the CRL.  So, there's
a research quantity only.

  The larger question depends on how fine a granularity you need to hit
good enough.  In a large track session room you could put skiffs in
each corner and, based on relative signal strength, figure out roughly
where people are sitting within the room.  Because infrared beacons are
directional you can tell what side of a 10 X 10 office the laptop is on.
Bluetooth hangs around the edges but nobody seems enthusiastic.  In
seems to come down to figuring the trade offs based on good-idea notions.

> Is position determination in 3 dimensions, not just 2, available?

  Not sure about this question.  The beacons are purposely range limited
to limit location granularity.  I wouldn't be completely surprised to
see production models allowing users to set the power.  If you can
fairly reliably limit the scope to a room or hall on a particular floor,
then by definition you have location in 3-space.

> How hard is it to go from coordinates information to a room?

  As I understand it, in Indianapolis this will be a hard association.
If you are seen by beacon 3, then you are in Salon B.

  It's a little more interesting to think how the beacon associates your
laptop signal with you.  The 802.11 scan packet contains your MAC
address.  The central server then needs either to do a lookup in some sort
of static table or hit DHCP for the association to L3 ip address.  From
there the SIP registrar can be queried for user information.

- Jeremy

-- 

----------------------------------------------------------------wg-pic-+
For list utilities, archives, subscribe, unsubscribe, etc. please visit the
ListProc web interface at 

    http://archives.internet2.edu/

----------------------------------------------------------------wg-pic--