Microsoft to contribute to Clang and LLVM project

[Ola Fosheim Grøstad via Digitalmars-d]

On Thursday, 10 December 2015 at 05:20:26 UTC, Joakim wrote:
> I don't see why others are so concerned about it.  A better use 
> of their time would be to chip in themselves, on documentation 
> or whatever else they're capable of contributing.
>

I think the primary concern is "what is the plan?". Without a 
clear plan it really doesn't matter what you do or not do as an 
individual with just a few hours per week. It's like voting or 
volunteering for a party with the right ideas, but no clear 
strategy for getting into position. The second concern is that 
people evaluate performance based on the official compiler. They 
evaluate Go, not gccgo, and they evaluate dmd, not ldc with an 
older frontend. This happens repeatedly when people write about 
these languages.

> sweet spot in every market.  Perhaps those are better tools for 
> those markets, while D will hit different segments of those 
> markets and new markets altogether.

That required a strategy. Like, I am now likely to pick up C 
again, just to be able to build tight asm.js. WebGL is now 
becoming mature and asm.js is becoming a massive target, but it 
takes a focused group to do better than emscripten... So you need 
a central strategy in order to organize something like that.

> I agree that Swift is a strong competitor, as I've been saying, 
> but it is currently way behind D in platform support, ie 
> currently just iOS, OS X, and largely done linux/Glibc.  Each 
> has their pros and cons and will garner their own adherents.

Swift may need 1-2 more years, but if people can replace two 
languages with one, then they have a strong adoption card. But I 
am not sure if Swift will be able to gain C speeds consitently, 
though I would not bet on it.

But it is rather obvious that being similar to Swift is not a 
good strategy. If languages get too close, then the better 
ecosystem wins.