Could D have fit Microsoft's needs?

[JN]

On Friday, 26 July 2019 at 03:29:33 UTC, vladislavp wrote:
> The number of negativism in some posts is very high.
> To the point where it is actually hard to understand the 
> technical details behind
> particular grievances.

I think the issue is that many flaws of D are not strictly 
technical. For old languages like C/C++ the idea was that 
language is king, makefiles are the ultimate build tool and IDE 
support means vim syntax highlighting plugin. Those times have 
changed. The ecosystem around the language, in form of available 
packages, frameworks, boilerplate projects, IDE plugins is just 
as important, if not even more important as the language itself. 
Just look at Go. It's such an uninspired language with hardly any 
notable features other than goroutines, and yet it's popular 
because it's easy to get started with, easy to find ready 
packages to use for various tasks, fast to compile and easy to 
deploy. All while maintaining good performance, as most native 
languages do.

The vision document has not been posted in a while, and most of 
the progress comes in form of bugfixes, memory safety related 
DIPs or extern(C++) related changes. But overall, I feel like 
many people believe that D is wasting it's potential. Now, 
normally, it wouldn't be an issue, and D would move forward at 
its pace as it always did. However, Rust happened, and people see 
how quickly it exploded in popularity. I guess people like to 
downplay popularity of Rust here, but for many new projects Rust 
is just the new default, if you don't need compatibility with 
legacy code. It's still not a language I'd call mainstream, but I 
believe in several years the question asked will be "why are we 
doing it in C++? can't we just do it in Rust?".

>
> As for Microsoft, It would be interesting to hear how they view
>  D's BetterC and its, hopefully soon, upcoming DIP 1000
> for their Driver Verification (SLAM)
> https://www.microsoft.com/en-us/research/project/slam/
> And building devices drivers in BetterC (since it specifically 
> allows to use C's or C++ runtime).

Oh. That one is simple to answer. I'm not Microsoft, but the 
answer will be along the lines of: "Our engineers are not 
familiar with the D language. While it has some interesting 
ideas, we really enjoy the opt-out safety features of Rust and 
the compile time checks provided by the borrow checker. Combined 
with a powerful toolchain and a very active community, it seems 
to be the future of native GC-less programming. Also, -betterC 
mode seems to be barely documented and it's hard to say what the 
future of it will be".