This is a pretty tuff issue for software development.
I workded for a DBMS company that provided DBMS servers and client software. the servers ran on windows and unix/linux and the client software on windows/unix/linux and Mac os.
After several years we dropped suport for the mac.
This was becaused the mac sales did not even cover the support costs. We factored manuals and training guides into the support costs, and discovered that the anualized sales of the mac product never exceeded about 3% of the total sales and the support costs were higher than that.
We got to choose. We could enhance the products that sold or we could support the mac. The answer really was simple when it came down to that. Even though I was (among other things) the manager of the mac development group.
The support issues came down to several things.
1) The support cost per sale was higher, partially due to few multiunit sales to a single customer, partially due to corprate sales where the mis dept only had a few macs and they didn't under stand them and partially due to end users not knowing much about their systems.
2) The cost of pubs was higher on the mac platform than elsewhere. It ws lowest on unix/linux. The standards that the mac community imposed on the pubs was such that our standard pubs could not be just modified, but had to be compleatly re-written.
3) We needed to maintain a mac-lab for development and support and it was expensive, with individual machines costing about twice as much as windows boxes. Further, tat the time, conectivity to file servers and corporate support apps (various tools, code control, and problem reporting systems) was non existant, requiring involved proceedures and additional machines (running windows) to solve.
Developement was also problematic, but didn't really enter into it.
For the simple development problems Java would have reduced some of the issues, and the current BSD based mac oses would resolve the tools ones, but the others remain. In the end it was just too much of a pain for the amount of sales.
Given all of this I find the dive gear makers lack of support for macs (and the same problem exists with various other areas, including amature radio, navigation and maping system, etc.) to be not suprising.
Still releasing the device API specs -- if even to a limited group of more 'professional' developers, with NDAs etc in place -- seems to be short sighted and defeatest. This could provide a bit of a competive edge to the makers that did it, albeit a small one.