I'm going to give you one more go and then I have to get some real work done.
However that is not what Erik Baker described.
Do you accept that the GF low prescribes the first stop? I think that was the original point you disputed and you went down the 'it cannot be because my shearwater ignores it for NDL' route to argue that.
Baker's Fortran code implementation of GF does indeed use GF Lo to determine the depth of the first stop, and that is set regardless of what the GF Hi is. It is quite simplistic, actually, in that the first stop is actually the ceiling. The calculation does not factor in ascent time. So, it could easily give you a list of stops to make that you would never actually need to make because you will off-gas enough during the ascent that your leading compartment tension would be less than the target for that depth before you ever get there.
I have never disputed that GF Lo prescribes the first stop. What I have disputed is whether it is used in the case that a direct ascent would not exceed GF Hi at the surface. Baker's code does use it. Other implementations of his concept do not.
If GF low is used to calculate the first stop and the presence or absence of stops determines whether a dive is a deco dive or not then surely the GF low is involved in determining the No deco limit?
So, one more time, some implementations (e.g. Shearwater's) use the GF Hi to determine at what time you would no longer be able to do a direct ascent without exceeding GF Hi at the surface. That is your NDL time. Once you exceed it, then you will have to make 1 or more deco stops and GF Lo is used to calculate where that first stop is. So, no, GF Lo is NOT NECESSARILY involved in determining the No Deco Limit. It depends on how the implementer chose to implement the algorithm.
Another choice for the implementer is whether to take the ceiling, based on GF Lo, as the first stop, or whether to calculate the time during the ascent where GF Lo would be exceeded and use that as the first stop. Baker's actual code does it the first way.
My opinion is just my opinion and is completely meaningless until such time as I choose to implement a deco algorithm into some code myself. That said, my opinion is that:
- if I can do a direct ascent to the surface without exceeding GF Hi at the surface, then that's what I want my computer to tell me, as far as NDLs go. If I'm not going to exceed GF Hi at the surface (after a 30ft/min ascent), then I really don't care what GF Lo is. Virtually every dive more than 5 or 10 minutes to deeper than 40 or 50 feet is likely to result in surfacing with tissue tensions higher than most GF Lo values (that people actually use) would allow. So what? If I'm diving GF30/70 and a direct ascent will result in having 40% of the M value in my leading compartment at the surface, should I have to do a deco stop at 10'? That seems ridiculous. And it's even more ridiculous if I'm at 60' depth and the 30% ceiling is at 10', so I'm prescribed a 10' stop, even though my leading compartment is only at 25% by the time my ascent gets me to that depth (a la Baker's Fortran implementation).
- Baker's implementation calculates the first stop as the current ceiling, using GF Lo, at the end of the last bottom time dive segment. Again, let's say you're diving GF30/70. You're at 150' and his implementation says that your first stop is at 70'. It has determined that by, essentially, saying that if you magically transported from 150' to 70' instantly, you would need to hold at 70' for some amount of time in order to avoid having your leading compartment exceed 30% of the M value. In reality, by the time you actually ascend that far, you may very well off-gas enough that you could continue to ascend to 60' or higher without having your leading compartment exceed 30% of the M value. So, how much sense does it really make to stop at 70' if you could proceed non-stop to 60' without exceeding 30% of the M-value in your leading compartment?
Now, all that I have said is based on my reading of Baker's code. I have not compiled it and run it to prove that it works the way my reading of it indicates. I could be wrong. If I am, I hope someone will explain to me why.
One thing I am not wrong about is that the Shearwater Dive Planner on my Perdix AI does NOT use the GF Lo value in determining an NDL. A dive to 150' on GF90/90 has exactly the same NDL as a dive to 150' on GF10/90. Of course, using the Planner is not the same as actually doing the dive. And, as SW does not allow changing GF Lo during a dive, I guess there's no way I can prove that it works the same way during a dive as what the Planner says, short of diving 2 Shearwaters together, with different GF settings. In the meantime, I have faith in Shearwater, that the computer will perform the same during a dive as what the Planner predicts. And, I am completely comfortable diving a GF implementation that does not use GF Lo when calculating NDL. YMMV
Lastly, this is a lot of debate about how to do something that Baker's algorithm wasn't really even intended for. I could be wrong, but I believe his algorithm is/was specifically intended to calculate ascents for decompression dives. It was not designed or intended to be used to calculate NDLs. So, how his implementation works when you try to use it for NDL dives may not really be the "best" way to use his algorithm or the best way to calculate NDLs. In other words, even though his implementation always calculates the first stop using GF Lo, that doesn't mean that that is the best way to calculate an NDL. I personally think the way Shearwater and others are doing it (using only GF Hi) is a better way of calculating an NDL.