RE: simple text format ID

[EXT Tim Salo <salo@networkcs.com>]

> Date: Mon, 10 Jul 2000 21:33:22 -0500
> From: "EXT James M. Polk" <jmpolk@cisco.com>
> Subject: RE: simple text format ID
> 
> At 01:59 PM 7/7/2000 -0700, EXT Pradhan, Salil wrote:
> >>This can be a derived value.  Take a couple of readings &
> >>determine the direction.  I don't think we want to load
> >>the protocol too heavily.
> >
> >Usually not very accurate. A built-in compass (if the device has one) can
> >give a much more accurate reading. I would argue that direction is very
> >valuable for location-based services.
> 
> Does your IP Cell phone have one of these? Does your PDA? Does your Notebook?

The general answer is "yes".

GPS receivers are the most common technology used to determine the location
of mobile devices.  Single-chip GPS receivers are available and are starting
to be embedded in a wide range of mobile devices.

o	I have heard proposals to put GPS receivers in the helmets of
	forest firefighters and the microphones of police officers.  In
	both cases, the location of the individual is periodically
	beaconed using his or her two-way radio.

o	Transceivers are available with ports to which GPS receivers can
	be connected, enabling the transceiver to beacon its position
	periodically.

o	People have integrated Palm Pilots, GPS receivers, maps and
	transceivers into wireless, mobile location units.

So, while my examples don't include a compass in an IP cell phone, it would
be silly to bet against far more powerful technology (namely a GPS receiver)
appearing in these devices in the near future.

Having said that, a few other comments follow:

o	GPS is probably the most common method of determining the
	location of mobile devices.  Important questions to ask
	probably include:

	-	What is the relationship between this protocol and
		NMEA-0183 (which defines the sentences generated by
		most GPS receivers)?
	-	What data generated by GPS receivers should the
		protocol carry?  What data generated by GPS
		receivers should the protocol _not_ carry?

o	I don't believe that text-based protocols are particularly
	easy to parse.  On the other hand, I don't this is particularly
	important.  I suspect that what you really want is a protocol for
	which the packets are easy to _create_.  That is, the place
	where you want to save power/processing power/... is in the
	mobile device that must transmit its position.  Presumably,
	these devices (at least the really dumb ones that appear to
	be a justification for this protocol) won't be receiving
	these packets, merely creating them.

o	Your protocol appears to assume that the location sensors
	in the mobile devices will be provide location telemetry
	in the form of text strings.  Of course, the most common
	text string for specifying location information is NMEA-0183.

o	(By the way, are you expecting this really dumb device to compute
	the day of the week, given a date?  I don't have a copy of
	NMEA in front of me, but I believe that GPS receivers will
	indicate the time and date, but not the day of the week.)

I strongly suggest:

o	You look at the NMEA-0183 specification and the information
	included in NMEA sentences.

o	You look the capability of several manufacturers of
	GPS receivers.  They generate a tremendous amount of information,
	including course and heading (another recent topic).  I have
	wondered whether chip-form-factor GPS receivers with the
	"most math per cubic inch prize"...

o	You decide and document what role GPS technology and NMEA-0183 will
	have in you spec.  At a minimum, I suggest that the mapping between
	at least certain specified NMEA-0183 sentences and your protocol
	be trivial.

o	You consider documenting why you don't simply use NMEA-0183
	strings.  This would permit really dumb devices to simply transmit
	the NMEA-0183 string, rather than parse it.  Again, this presumes
	that your location sensing device is likely to be a GPS receiver.

I am all for simple solutions, but describing a mobile device's location
is probably not a simple problem, (as may be evident if one were to look
at the NMEA-0183 spec).

-tjs