D vs Rust

[Eljay via Digitalmars-d]

On Sunday, 3 September 2017 at 19:40:27 UTC, Ali Çehreli wrote:
> Been there, done that. Decision makers rationalize any other 
> choice pretty easily:

In my fantasy, I'd be the decision maker.  ;-)

In reality, along all my career, usually the decision as to the 
language used was made usually by (non-engineer) management, or 
sometimes by the principle developer by fiat.

And merits of the language was but one of many factors, but I 
think the foremost factor was whichever language the decision 
maker was familiar with, or enamored with.  (In your example, 
hence... PYTHON!  I actually like Python, but I've only used it 
for scripting purposes.  Less than 1 kLOC.)

Often times in the projects I have been on, that decision maker 
has long since moved on to different projects or left the company.

And the language decision is nigh impossible to change.  Except...

On my current project, the original language was Object Pascal.  
At some point in the early 90s, the Object Pascal code was 
mechanically transformed into C++89 code.

Even though that was a very long time ago, there are still Object 
Pascal idioms present in the code.  Which are not idiomatic C++.

Semi-ironically, Object Pascal is still around, and has it's own 
following.  Although now referred to as Delphi or Oxygene or 
Freescal.  They're all Object Pascal.

Regarding the project language choice, the one thing I have found 
consistent from all my projects is:  picking a locked-down 
proprietary language is invariably a bad decision.  Languages 
like PowerBuilder Powerscript or Forte Tool or Meditech Magic or 
PickOS PickBASIC.

My apologies to devs that enjoy Powerscript or Tool or Magic or 
PickBASIC.  I'm not saying that those languages cannot be 
enjoyable.  The nature of proprietary languages make the projects 
tied to them at their mercy.

Whereas with D, if I don't like dmd, I can use gcd, or ldc.  Or 
if D doesn't support my favorite platform -- obviously Amiga OS, 
the one best OS ever made -- I could in theory do the work to 
leverage D as a front end to my toolchain.

Hmmm, maybe that doesn't make sense.  Sorry if I didn't express 
myself clearly.