buying a new dive computer

[stuartv]

Stuart, thanks for taking over the attempt at educating dmaziuk, I had given up. Though believing he/she knows everything, posts say otherwise. One of the dangers of the internet, separating the wheat from the chaff

Ha ha! It's only fair. I know I'm the same way often enough....

 

[dmaziuk]

Ha ha! It's only fair. I know I'm the same way often enough....

There is a difference: in this case I am arguing from the position I don't know squat. I am not qualified to evaluate the implementation of different algorithms in different dive computers. I am not going to base my choice of a product on a feature I cannot evaluate.

 

[stuartv]

There is a difference: in this case I am arguing from the position I don't know squat. I am not qualified to evaluate the implementation of different algorithms in different dive computers. I am not going to base my choice of a product on a feature I cannot evaluate.

Ahhh, false modesty. I know better. You are a sharp person. If you wanted to, you could read the lab test reports (I posted links to two different lab tests, I think) and draw reasonable conclusions from the data they provide.

Or maybe you have a Cressi computer and you don't want to admit - even to yourself - that you would enjoy longer NDLs (and safely) if you had bought a different computer?

 

[dmaziuk]

If you wanted to, you could read the lab test reports (I posted links to two different lab tests, I think) and draw reasonable conclusions from the data they provide.

I am familiar with software testing methodologies, data analysis, and all that. Where I come from "evaluate the performance of an algorithm" is not a clear statement of the goals of the test. The methodology is "we put a bunch of computers in a chamber and pressurized them to various extents": unjustified is putting it mildly. There is no baseline despite the fact that ZH-L16 is well documented and available is several software implementations. Yeah, I've read the reports. My conclusion is, I don't understand them.

I could familiarize myself with ZH-L16+GF and potentially: VPM-B. Everything else is various degrees of "based on" "modified" proprietary, so no understanding those. Then based on my understanding of the algorithms I'd have to buy a petrel or an ostc. Well, I'm not buying: they cost more than a ticket to Bonaire and don't even come with AI at that price. Meh.

(And don't say H3: is costs even more and still comes with no AI.)

 

[Freewillow]

who said that one needs an AI.................