On Wed, 8 Jun 2011, Peter Campbell Smith wrote:

Offramp wrote in news:3ebb8c47-d448-4ed1-967a-
:

You may have seen Bus Inspectors using PDAs to check addresses and
issue penalty fares. They use what looks like an XDA II. The Rapid
project was to create a very similar system for use by Tube RIs*. They
did not have to work underground, as such, they would send PFN
information telephonically once above ground. Electoral roll info was
to be stored on the device itself. The project was in development for
about two years but the company which had been asked to develop the
system produced nothing physical and the project was abandoned.

As I recall, the original spec did ask for the device to transmit the PF
information underground. Apparently experience with parking enforcement
is that miscreants grab and make off with the PDA, so the principle is
that it saves the data elsewhere within a few seconds.

At the time I read it I could only envisage saving the data initially on
some device in the train, and then downloading it at a station or depot.
It seemed a lot of cost and a lot that could go wrong for perhaps not
very much return.

The obvious move would be Bluetooth, i think - the cheapo text-entry
device in the hand, and a more expensive recording and uploading device
concealed the pocket, satchel, codpiece, etc, with data being sent
locally. I'd even suggest using a physically unmodified mobile phone for
the handheld part; as long as it has Bluetooth and can run J2ME, a bit of
custom software would be all that's needed. Throw in a camera, and you can
even snap a picture of the perpetrator for the records. Such things are
available to the man in the street for $42.70 from Hong Kong:

http://www.dealextreme.com/p/jc-3t-2...fm-black-59773

And could perhaps be got more cheaply in bulk. If the software was written
using J2ME, then any cheap phone could be used instead - the Bluetooth API
is standardised. Oh, and users could make phone calls on it. And watch TV.

The device in the pocket could be more expensive - but it actually needn't
be a lot more. That handset has GPRS, so it could do the upload work.
Something in the pocket could have 3G or WiFi, but that's a marginal
benefit. Perhaps it could just be another JC-3T, running in server rather
than client mode. Perhaps the software should be symmetric, and just
mirror the data between both phones (or rather, as many phones as the
gripper happens to be carrying).

Anyway, i reckon i could do it myself in less than two years for less than
1.6 million. Probably six months for development, another six to twelve
for rollout and troubleshooting, half a million all in. TfL can contact me
at this address if they're interested.

tom

--
Baby got a masterplan. A foolproof masterplan.