Physiology and Computer algorithms

[jagfish]

Hi
Regarding this quote from another thread:
"The Aeris, Oceanic and Sherwood are essentially the same computer and all manufactured by the same company (Pelagic)...they are the most liberal computers out there and that they are based on older algorithms that do not take free gas (microbubbles) into consideration.
...
The ScubaPro SmartCom is a different computer. It uses wet contacts, which make it a little harder to operate. It's algorithm takes free gas, workload, temperature and more into account and is not nearly as aggressive as the Pelagics."

I got to thinking about different computers and algorithms. From a medical perspecitve, how much research/evidence is there that any particular algorithm is inherently safer, or more importantly, riskier, than others? Not theory, but actual observations from actual occurences.

Many thanks in advance to those who can illuminate...

JAG

 

[BigJetDriver]

I got to thinking about different computers and algorithms. From a medical perspecitve, how much research/evidence is there that any particular algorithm is inherently safer, or more importantly, riskier, than others? Not theory, but actual observations from actual occurences.

Many thanks in advance to those who can illuminate...

JAG

JAG, et al,

If you remember from your diving beginnings the stories of the Haldane's, pere et fils, you will remember that their work was closely attached to actual observations of physiological effect---on goats, that is!

From there, their results were applied to "crash-test dummies", otherwise known as caisson workers and British Royal Navy Divers. Here the results were modified based upon actual observations. Such as: "I say, that looks bloody well painful!!! I should imagine we should modify the numbers a bit, eh?"

Lo, time has passed, and the Haldanian tables have been modified by folks in the Royal and U.S. Navies, Buhlmann, et al. Some of that work has been empirical in nature, (i.e. "Ouch, that hurts! Let's change it!") Some has been mathematical in nature, (i.e. "That looks like it should work.---Oops!--- Ouch, that hurts!---Let's change it!")

Some of the mathematical work DID work in practice, and was establlished as "gospel".

Now in the modern era, we take a theory based upon what we know of the physics and physiology of the problem, add to that current statistical data, and crunch the numbers with super-computers.

When we are ready, we head for the goat pen to find some test divers. The goats, however, are defended by animal rights terrorists from PETA. While the goats are happy, this makes our task much more difficult!

Fortunately, we find volunteer crash-test dummies such as Tom Mount, Tim O'Leary, and the boys and girls from NAUI's research team. Since the goats are NOT willing, and these folks ARE, we had best not get into comparative thoughts about who is being smarter about this.

Finally, with much diving, much cursing ("Damn, that hurts!"), much modifying, and much more diving, we come up with a full-blown and well-tested algorithm, such as Dr. Wienke's marvellous RGBM!!!! Thus, everyone is happy, including the goats!

It is truly an amazing process, n'est pas, mes amies???

BJD :anakinpod

P.S.---I am, in fact, a trained and certified DMT and so have a bit of a background in the theory and practice of----wait for it----
"goat-bending"!!! :doctor:

 

[jagfish]

Sorry BJD
It must be me, but I'm not sure I get the point of your post. Perhaps my post was not penned well, or the question is itself is flawed.

I'll try again. In reading posts by advocates of different computer manufacteres due to Algorithm preference, I am wondering if there is any observation in DCS incidence that suggests the use of certain computer algorithms is "riskier" than others...

In other words, has dive accident data shown that wearers of computers with certain algorithms are more prone to DCS?

JAG

 

[oharag]

Sorry BJD
It must be me, but I'm not sure I get the point of your post. Perhaps my post was not penned well, or the question is itself is flawed.

I'll try again. In reading posts by advocates of different computer manufacteres due to Algorithm preference, I am wondering if there is any observation in DCS incidence that suggests the use of certain computer algorithms is "riskier" than others...

In other words, has dive accident data shown that wearers of computers with certain algorithms are more prone to DCS?

JAG

jagfish, there was a lawsuit against Uwatec for their Aladin series (I do not remember the model) from a couple of divers claiming they had gotten bent because of faulty algorithms. Uwatec never acknowledge the flaw, or they acknowledged it too late. So I guess you could always question the accuracy of the models for every computer (brought on by bugs in the software).

I've heard of some computers (Vytec comes to mind) that give erroneous depths. So, again maybe times are wrong as well.

From what I've been reading on these boards most believe (ie. Tec Divers) that computers are tooooo conservative. The only method to maximize bottom time is to cut tables using what they believe to be more accurate algorithms using software. Whether you will get bent on a computer is a matter that I can not comment on. I would suppose that computers have been dove numerous times over the past decade (or whenever computers were introduced to diving) with incident.

I'm a new diver. I will be purchasing a computer initially to track timer, pressure and no deco limits. Eventually as I grow as a diver I will maybe evolve into cutting my own tables using software, and go to using the computer as a bottom timer.

oharag

 

[BigJetDriver]

Sorry BJD
It must be me, but I'm not sure I get the point of your post. JAG

Jag,

Sorry. I think I misunderstood your request. The only repository for that information would be DAN. For obvious legal reasons, they could never publish it. Sorry.

BJD :anakinpod

 

[jagfish]

It's intersting for me
On the computer forum, people were talking about how some comtuer algorithms are inhereently safer than others (more conservative) and yet as all of you pointed out, I think it would be difficult to confirm that from actual incidents.

As someone also pointed out, it would be in the interest of manufacturers to not look into that information too carefully...

JAg

 

[Scuba]

It's intersting for me
On the computer forum, people were talking about how some comtuer algorithms are inhereently safer than others (more conservative) and yet as all of you pointed out, I think it would be difficult to confirm that from actual incidents.

As someone also pointed out, it would be in the interest of manufacturers to not look into that information too carefully...

JAg

Safer in the following terms: If you reduce your nitrogen load, you reduce your risk of DCS.

Computer A allows 30 min at 80fsw.
Computer B allows 40 min at 80fsw.

I doubt anyone has the information necessary for an accurate assessment. Many variables affect DCS: temperature, hydration, physical exertion level, fitness, pfo, ascent rate, etc. Unless you have this information its difficult to prove algorithm A or B is safer with certainty. What can and is done is to use the example above to try and predict safety. Also, if you take a large enough and varied sample, such as 1 million dives using A algorithm and 1 million dives using B algorithm. There should be a fairly similiar representative sample containing all the various traits affecting DCS within each sample base. This gives us probability of outcome. If divers using algorithm A get DCS at a much higher rate than those using B, it "probably' is the algorithm. But, this type of study doesn't 'prove" it. There is a chance the two samples are biased in some aspects, as opposed to containing a similarly even representatiion.

This type of sampling is used to predict incidence of DCS. The often quoted number of 1,2.3.4 cases per 10,000 dives for the general population. There are cold water statistics showing a higher rate. The increase in use of computers which tract and store actual dive profiles is leading to research that will help to isolate and compare some of the variables involved more precisely. I believe someone, don't remember who'm, is collecting computer dive profiles which divers can upload on the web. Don't know any specifics. There are probably treatment chambers and others doing this as well.

.

 

[jagfish]

Interesting observations, Scuba
I've heard that contary to what one might naturally think, the incidence of DCS has on the whole been influenced upwards with computer use. I did not personally see the literature on this, but an instructor I know said he did.

Apparenly, the gist of the finding was that computers are naturally less convervative that pure table use on the whole, so computers bring people closer to their limits and thus suffer more hits on average.

JAG

 

[DEEPLOU]

Interesting observations, Scuba
I've heard that contary to what one might naturally think, the incidence of DCS has on the whole been influenced upwards with computer use. I did not personally see the literature on this, but an instructor I know said he did.

Apparenly, the gist of the finding was that computers are naturally less convervative that pure table use on the whole, so computers bring people closer to their limits and thus suffer more hits on average.

JAG

When you use a table, everything gets rounded to the next larger number. (depth to the deeper depth and time to the next time group) and the table is done based on a square profile at the max dive depth planned.
This rounding always keeps adding "padding" to the numbers from a table.

The computer keeps track of the exact time and depth and gives exact calculations based on the actual profile of the dive.
So while the table and computer may use the same algorithm the table has built in "room for error" or conservatism.

 

[DEEPLOU]

It's intersting for me
On the computer forum, people were talking about how some comtuer algorithms are inhereently safer than others (more conservative) and yet as all of you pointed out, I think it would be difficult to confirm that from actual incidents.

As someone also pointed out, it would be in the interest of manufacturers to not look into that information too carefully...

JAg

I think if there were any pattern of incidents from other computers, we would see about it. Some lawyer would get ahold of it and we would see other lawsuits like the one against Uwatec.