To all the list,
many thanks for your replies so far,  rather than reply to each
individually I have picked out several of the issues raised by various
people and will try to answer them here.

1) Audio systems.
There are quite a few good solid state audio systems already available
both those using broadcast media and those with built in data.  It would
be quite difficult for us to compete with these from scratch.
Our device would be capable of playing MP3 and WAV files for both sound
clips and narration but using a lot of sound will seriously deplete the
storage available for text. Ultimately this would be the content designers
choice.
The real problem on this front however is the ergonomic problems of using
headphones or speakers in the museum environment.  I personally think that
these degrade interaction between visitors.

For those with reading difficulties we may also be able to provide a text
to speech system - i.e the unit can read out the written material. This
avoids the storage problems associated with large amounts of recorded
material.  Current speech systems tend to sound rather robotic and North
American but these systems are improving rapidly.

2) display quality.
In my opinion this is critical we would probably trade off quite a lot of
battery life in order to be able to run a very clear high contrast
display, black on white rather than dark green on light green. IYSWIM
- we would definitely put in some kind of text enlargement feature.

3) reliability and robustness
I take everyones point about the system having to be people proof.
units would be automatically recharged when placed in their docking
receptical.

4) storage and CD's
The unit supports a USB connector which will allow the addition of
peripherals - in theory you could plug in a cd-rom player to boost to
content up to full multimedia levels. however there may be some
difficulties in carrying everything around.  Normally you would archive
your content onto disk or tape from your PC. you can easily change the
content in the handheld when you synchronise it with the PC.

5) labelling lots of items.
we are looking at several methods of identifying the exhibit - different
ones will be appropriate in different places.  where you have a great many
small items it would be most appropriate to use a short range passive tag.
the tags would have to be cheap and the unit would have to get pretty
close to compensate so the effect is of having to point the unit within a
few inches of the object.  Where you have got big objects - say an art
gallery its more appropriate to use the 'spatial awareness' built into the
unit.

---------------
please keep the comments coming, they will all help to improve the system.
if anyone is particularly interested in talking about this project please
feel free to call me directly ( intl +44 1926 640073, uk 01926 640073 )

We are also collecting financial backers at the moment if anyone really
wants to help <grin>

Regards
Andrew Watkins.