Re: Rust: kernel vs user-space

From: Mark Delany <x9k_at_charlie.emu.st>
Date: Wed, 04 Sep 2024 19:52:43 UTC
> And this is better than C because ...?

I don't think any sane person suggests rewrites for rewrites sake, but it's the new code
that needs to be created in the coming decades that is in question.

The argument goes that C is becoming niche as the diminishing pool of developers approach
retirement age. Wasn't that the most obvious trend of the last community survey conducted
a while back?

At some point finding C programmers may be as hard as finding COBOL programmers.

The second argument goes that other languages are simply much more productive and that
productivity is crucial for very large projects which need resourcing over decades.

> Lua
> Python
> Ruby
> Rust
> Zig
> dlang
> go

Above and beyond even deciding whether another language besides C is a good choice and
above and beyond picking which one(s) are most suited, the underlying issue is that
programming languages may continue to evolve and emerge and thus any choice today risks
rapid obsolescence.

It just may be that the programming industry is addicted to the siren call of the next
shiny new thing and that there are no good language choices to be made for a project that
spans decades.

If you look at each of the decades that BSD code and its forebears have lived thru, each
one produced a shiny new language with almost all of them now in rapid decline, extremely
niche or defunct.

So, any discussion like this is mostly about predicting the future, which I'm told is
rather hard:

 - Is there a significant risk in not being able to source future C programmers?
 - Will a more popular/current language attract more contributors?
 - Is it possible to chose an appropriate language which will remain popular for decades?
 - Is it possible to define a set of criteria for when such decisions need to be made?
 - Is indecision a good option to see if next-gen programmers lose their addiction to shiny?


Mark.