New Bug Reports/Feature Requests

[Dirk Hohndel]

On Tue, Feb 23, 2016 at 05:18:52PM -0500, Richard Houser wrote:
> > The weight of the metal changes at least for different manufacturers so
> unless you weight your cylinder, which value are you going to use? The
> weight of the gas is of course always the same.
> 
> Well, the weight of the metal itself isn't so important, but rather the
> bouyancy when switching between tanks.  These are published for each tank,
> however, and they will be in a similar ballpark at least between
> manufacturers.  For example with valve, an empty HP 120 PST is about -1.3,
> Worthington is about -2 and a Faber is -0.65.  Pick a center point and that
> whole range is less than 1lb +/- for the range of tanks.
> 
> My LP72 comes in about neutral, and my AL80 is around +4.4lb.  I just think
> it would be really helpful to have that information on the same equipment
> page with the rest of the weighting, so it's less mental math and manual
> notes to figure out what all went into a dive (including which pony bottle,
> etc.) to adapt that portion of the weighting back to something usable for
> other dives with different details.

And the relative buoyancy is different depending on the manufacturer and
even the year when it was manufactured. So no, you can't just assume that
all AL80 are +4.4lbs (I assume that you are quoting empty buoyancy). E.g.,
the very popular XS Scuba AL80 right now is +3.4lbs empty and -1.4lbs
full.

> > As you see, we indeed use the nominal sizes even though they are not
> exact. We had a discussion in the past (but I am not able to find it either
> in the mailing list archives or trac) if we should use the „real“ values as
> shown on that web page but ended up the conclusion that that would actually
> be even more confusing.
> 
> In my classes back in Michigan, and during checkout dives I witnessed as a
> bystander in Ohio and Florida, instructors with at least 5 different shops
> (SSI/PADI/NAUI) have explicitly reminded students that when doing their
> logs, air capacity in an AL80 is 77cuft, NOT 80.  So, the software's

Just because many people are saying this doesn't make it right.
The XS Scuba tank above is indeed 77.4cuft.
There's a Luxfer that's 78.2cuft. There's an older Luxfer that's 79.8cuft.

Frankly, if you are worried about 3% "incorrect" data - remove the valve,
fill the tank with water, measure the amount of water and use it like a
metric tank.

And oh, btw: don't use ideal gas laws much beyond 200bar because those no
longer apply (and we do take that into account for HP cylinders).

> > A field is only changed when the keyboard focus leaves it (changing it
> with every keystroke would have strange side effects as you would end up
> with strange values in the middle of typing a number). But you can for
> example use the TAB key if you don’t want to change to the mouse.
> 
> This is exactly what I experienced, but I argue that is an aspect of bad UI
> design.  The workaround of clicking or tabbing to another field shouldn't
> be neccessary, IMO.  As a user, when I select save, I expect it to save.

You are confusing "UI design" and "shortcoming in the UI toolkit that we
haven't figured out a way to reliably work around". No, we didn't DESIGN
it that way. It turns out it works that way because no one has figured out
how to fix it. Patches are welcome.

> > My guess is that this is a rounding error: divelogs.de uses metric units
> internally so converting from imperial to metric and back could change the
> least significant digit.
> 
> I agree, but this worked in December with 4.4.2 and divelogs.de at that
> time.  It certainly looks like a regression, but I can go back and confirm
> it's on the Subsurface side by re-exporting old dives in 4.5.3 and 4.4.2
> once I return from my trip and match it up to divelogs.de, if others don't
> see this same behaviour.  If divelogs.de is working in all metric
> internally though, and subsurface is working internally in metric, doesn't
> the fact that divelogs.de still displays the old values correctly indicate
> this is a problem between the subsurface export to divelogs.de?  Perhaps
> one or both sides started using a different precision or something?

We store all data in metric units because the imperial units are
completely braindead, anyway. It's interesting that something changed
there between 4.4.2 and 4.5.3. Something one of the developers could look
into to figure out what might have triggered that. Thanks for the report.

> > Do you realize that the color of the depth graph is already a
> representation of the momentary SAC? Given the resolution of the pressure
> sensor I guess it would make little sense to present finer granularity than
> a color coding (there is only very limited numerical precision).
> 
> I did not (thanks for pointing it out), but my logs (Aeris A300 CS - aka
> Oceanic VTX) seem to imply I have at least 1psi resolution.  I think this
> would be plenty to give a useful graph of the SAC calculations as well as a
> smoothed out rolling SAC average over the last 4 samples or something (with
> presentation similar to the average depth of the dive thus far
> calculation).  If I had a less accurate computer, I would certainly expect
> a little more noise in any calculations, like I currently do with the 1
> degree resolution of the temperature sensor.

We already do all that. It's in the color of the graph as Robert
suggested. Including the smoothing, of course. And the smoothing is time
based not sample based because otherwise on a dive computer with 2 second
resolution you'd still get very very "jumpy" data.


/D