Nim programming language finally hit 1.0

Ola Fosheim Grøstad ola.fosheim.grostad at gmail.com
Mon Sep 30 11:40:01 UTC 2019


On Monday, 30 September 2019 at 11:05:58 UTC, IGotD- wrote:
> A lot of negativism while C++ is committing harakiri. One 
> obvious future of D is services inside Linux embedded systems.

That is certainly an option, you might be able build a community 
around the Linux embedded niche.

> Vibe. The code became much more readable and consistent. Memory 
> leakages are gone because of GC (don't hate GC, it is great for 
> most use cases).

But you can not rely on GC on small embedded systems. Even 
allocators that doesn't combat fragmentation is a relevant issue.

So if embedded is a target niche then such runtime features must 
be addressed.

If not, why would a project manager pick D over Go (beyond syntax 
preferences)?


> libraries/interfaces are more immature. Rust is the most 
> competitive contender but as complicated Rust can be D can 
> really be an option here.

But D does not provide the same memory management capabilities of 
Rust. Rust would be the stronger contender if GC isn't an option, 
and in that case D does not have a comparable memory management 
solution.


> What I think is one high priority is to get reference counted 
> GC in D because that opens up D for the performance crowd who 
> do not want stop the world GC.

Well, ref counting is too slow for the performance crowd, but it 
is better for predictable execution and limited memory usage than 
a GC. Which in some embedded contexts might be sufficient.

If embedded is the main target niche then you would want clean 
and flexible features related to stack allocation and automatic 
proving upper bounds on stack size. Such features could make any 
reasonable language a strong contender in this niche.

Being generic isn't sufficient to be a strong contender in 
specialized niches.


More information about the Digitalmars-d mailing list