Many people go on dive vacation trips. Being able to retain the complete trip worth of data would be useful (no need to carry downloading equipment). Also useful for forensic purposes if something bad happens. If you are doing lots of dives, you will still retain some nitrogen loading the next day. So you really want the complete trip worth of data to be available.
So a minimum of 2 weeks of diving would be good. Not sure how many minutes this is. But a minimum of 50 hours would be a starting place. More is better. Memory is almost free these days, so put lots in.
Also agree with the suggestion of wet contacts and water (not pressure) activation Instead of buttons (like early uwatec computers). Also consider filling it with oil so the case does not need to be so strong (again, like early uwatec computers).
Or if it is designed to discardable at the end of battery life, then just pot the whole thing as a solid mass. No moving parts (other than the pressure sensor) no serviceable parts, no need to design a case that can be opened for service. The potting compound can be the case.
---------- Post added July 3rd, 2014 at 05:47 PM ----------
going cheap likely means no rechargeable battery circuitry. Going real cheap likely means no user replaceable battery either. Setup cost to design complex case moulding and then include a watertight door can be significant.
This may mean a larger footprint in order to include sufficient battery power so that it becomes a disposable unit.
So I want a battery life of around 500 1 hour dives with at least a 5 (or 10) year shelf life.