Manu:
> but I don't believe I'm alone.. the rest
> of the gamedev community will find D soon enough if the language gets it
> right...
I think games are one of the most important short-term purposes of D, despite I think D was not explicitly designed to write games.
I agree, it certainly didn't seem to be a major consideration early on, but I don't think any decisions yet made prohibit it from being well suited.
If as you say, there is some focus to generate interest from the game community, it would be really nice to have a few binary GDC cross compiler distributions available (for windows, linux users never have problems building the toolchain themselves). ARM/PPC, maybe MIPS toolchains hosted somewhere on the website might really help encourage some people get started, and I'd love to spend some time on the standard libraries for these platforms.
> If you're suggesting the reason for trapping overflow's is
> specifically to CATCH bugs like this, then maybe make is a compiler
> flag when building a debug binary? (ie. assert on integer overflow).
Right.
I think D2/D3 has also a bit of hope to replace some of the purposes of Ada language. Walter maybe didn't think of it when he designed D, but D shares some design purposes with Ada. Walter past work in aeronautical engineering leads naturally to a language that shares some of the purposes of Ada. For such purposes correctness and reliability are of the highest importance, this also means full type safety (implicit type conversions = bad) and number safety (integral overflows = bad). Defining something like a "MISRA-D" (a strict and safe subset of D similar to MISRA-C) is an option, and maybe it will produce a less butchered language.
I appreciate the attention to floating point detail made in D, and this isn't incompatible with gamedev, but making standard ints compromise basic hardware implementation just won't fly.
Compile time flag maybe to enable integer exceptions perhaps, or a special mode like you say...