Multiple tanks from divelogs

[Dirk Hohndel]

On Wed, Jan 28, 2015 at 02:44:19PM +0100, Rainer Mohr wrote:
> 
> Here's a point where I would suggest to generally handle this differently...
> IMHO, the gas change event is not suitable for deciding if a gas has been
> used or not.
> 
> Think of the growing number of sidemount divers. They use 2 separate tanks
> with a reg each, mostly filled with the same gas. They switch regs several
> times throughout the dive, but don't do the gas switch in their computers,
> because there is no need for that (same gas).
> 
> Or take the people with computer models, that don't support multiple gasses.
> They might still do their deco with a second nitrox tank or even pure oxygen
> for safety reasons, but don't have a possibility to document the switch in
> their computer.... but they will still want to document this in subsurface.
> 
> All these examples would get a wrong SAC, because the other tanks have not
> been considered, although they breathed from them.
> 
> Why not simply check if the endpressure of a tank is lower than the
> startpressure to decide if a tank has been used? If the pressure changes by
> more than say 3% (to compensate pressure changes from temperature), it must
> have been used... No error I can see using this approach, as pressures per
> tank are entered individually for each dive.

Here's the problem. Shearwater and a few other dive computers always give
us ALL gases that are configured. And since they aren't air integrated, we
don't know which ones were used based on pressure. So THE ONLY WAY we can
guess which ones were used is by looking at gas change events

> >I assume, you have not selected the "show unused cylinders in equipment
> >tab" as you are not seeing the EAN95.
> 
> OK, i see them now. Thanks. Was not aware of this
> 
> But I would still suggest to use another approach, see above. Your thoughts?
> Dirk? Linus?

If I knew of a solution that works, I'd have implemented it a long time
ago.

I guess we could change our logic like this:

- get rid of the "show unused cylinders" thing
- on download from dive computer, only include those cylinders that are in
  use on the dive, based on whatever information we get from the dive
  computer
- on all other methods of input, don't mess with the cylinders that are
  being given

That means that for all existing dives downloaded from one of the
"offending" dive computers you are out of luck and need to manually remove
them if you don't want to see them. But for other scenarios we are doing
something "somewhat intuitive and consistent".

Thought? Counter suggestions?

/D