Graph documentation [was: Re: Meaning of GF settings]

[Dirk Hohndel]

"Robert C. Helling" <helling at atdotde.de> writes:
>> I hear you. I learned FORTRAN in high school and forgot it as quickly as
>> I could. I actually looked at those same sources before Heinrichs and
>> Weikamp kindly offered to let me start from their C sources…
>
> You mean you have a VPM implementation that is more readable than the original FORTRAN code? That would be interesting to see.

No, I have seen a FORTRAN implementation and a Java implementation,
neither of which I would consider "readable"...

>> I was simply wondering if this would be something that long term
>> you'd be interested to play with.
>
> That is quite a precise description of the situation.

Cool.

>> But what I like about VPM(B) and RGBM is the abstract idea behind it. To
>> me it feels more "right" to try to estimate the bubble build-up then to
>> use some magic formulas to make 16 (or fewer in many cases) compartments
>> that simulate my tissues match empirical data. Call that naive (because
>> frankly, I realize it is), but I would love a working, well tested
>> bubble model :-)
>
> My understanding is you have the bubbles in addition to the 16
> compartments. You first compute tissue loadings for the compartments
> just like Buehlmann. Then only as a second step you don't calculate
> M-values (or their GF corrected version) but you pretend these
> compartments now feed their inert gases to bubbles.

That shows you how little I must have understood when reading the
Sources. I did not realize that this saddles on top of Buehlmann.

>> Having said that - questions about this keep coming up and allowing
>> people to play with this in the rather nice visual representation that
>> Subsurface can do - and do so on top of their own dive profiles... I
>> think that does have value.
>
> I would love to do that myself.
>
>> I certainly have spent way more time than I'm willing to openly admit
>> using different GFlow/high pairs against my existing deco dives... :-)
>
> That's exactly what you are supposed to do. I did the same (feature
> request: It would be great to have sliders for the GF's rather than
> having to type numbers).

Yeah - I think Linus was very vocal that he doesn't like the way the GF
settings are hidden in the preferences right now - so we should find a
better way to do this.

I had intentionally hidden them as my thought had been that much of this
isn't interesting for most recreational divers. But I guess once someone
selects "show me deco overlays" then it makes sense to have those
controls in the main UI instead.

> I realized only quite late that even though the green area changes a
> lot when you play with GF_low so the effect seems huge it is not since
> the ceiling only indicates the first stop and that typically is quite
> short so if you ascent slow enough you don't do a stop at all. You
> have to look at the full decompression profile so see how significant
> the effect is (like total time to surface).

That is the other change that I want to make. Extend the graph out to
the end of deco - as we already discussed.

/D