I agree with all of Noel's points, though mainly #3 and #6 because these are the only areas where BlitzMax is actually inferior to Blitz3D. A garbage collector that fails to collect circular reference is practically useless: memory leaks in a garbage collector should be an oxymoronic statement, but not with BlitzMax. Garbage collection must be transparently functional, otherwise it only adds way more complexity and problems than it's supposed to be solving.
I agree. Working with an obscure product is no fun, for the reasons you mention. If anyone doubts this, compare the example / documentation / support base of C# or Java vs. BlitzMax. I'm not saying that I expect BlitzMax to reach the status of C#/Java/C++, but it could do a whole lot better in terms of marketing.
The problem is that BlitzMax exists in a world completely cut off from mainstream programming. It's been around for a while, but there are still no books on programming with BlitzMax, popular libraries don't offer BlitzMax modules on their sites, and none of this will ever happen. So it's a developmental dead end, and will never be taken seriously or even noticed, even though BMX and its associated libraries are far better than a lot of other languages out there.
I agree. Working with an obscure product is no fun, for the reasons you mention. If anyone doubts this, compare the example / documentation / support base of C# or Java vs. BlitzMax. I'm not saying that I expect BlitzMax to reach the status of C#/Java/C++, but it could do a whole lot better in terms of marketing.