Nim programming language finally hit 1.0
Ecstatic Coder
ecstatic.coder at gmail.com
Tue Oct 1 05:35:22 UTC 2019
> I think no matter what you do, C++ folks will complain. They're
> just triggered by the word GC. That's why every single
> discussion thread about D outside of this forums starts at the
> GC. Then someone will mention @nogc or refcounting. Then
> someone will chime in about how you lose most of the packages
> and standard library because it assumes GC is present. And then
> people will just go "oh man, that is so complicated".
Exactly. ILHO the "BetterC' approach won't help D get new
adopters.
One the contrary, all the effort spent in removing the GC is
useless, as the GC is precisely what makes D a better language
over C++.
I'm convinced his huge amount of effort should have been spent on
making D a better Go/Crystal/etc.
As I said earlier, I *REALLY* wanted to implement Cyclone, my
CQL/SQL script runner in D.
The problem is it was SO MUCH easier to do it in Go, despite is
EXTREMELY limited compared to an object oriented language like D
: no genericity, no parametric polymorphism, no virtual
inheritance, etc.
But the language and its standard library provides ALL the
building blocks you need to implement web servers (coroutines,
HTTP, etc), and the official CQL/SQL database drivers are
complete, optimized, well maintained and used extensively.
As I already said, it takes you just a few very simple lines of
code to implement this script runner in Go, because you can
clearly see that the language itself was designed especially for
that : manage efficiently HTTP and database connections.
Which is nice, because it's the world most developers live in now.
Putting so much effort in trying to convince the few developers
developing C++ embedded applications to switch to D, while D is a
very nice garbage collected language, and so many of us use other
ones (Java, C#, JavaScript, Python, PHP, etc) to develop
connected applications a completely ineffective strategy, if D
wants more contributors.
Because as one has said above, with so few contributors working
on their spare time or funded by the D sponsors, I think the
language should strive to remain as SIMPLE as possible, and first
be enhanced to provide what most STANDARD developers need, not to
add complicated micro-features which are useless to most of us.
And unfortunately, despite their imperfections, it seems that
several new languages like Go and Crystal have at least perfectly
understood those principles.
More information about the Digitalmars-d
mailing list