[xormapmap]

So many of the so-called "C alternatives" end up doing way too much. I don't need algebraic data types or classes or an integrated build system or a package manager.

What I would like to see is a language that is essentially just C with the major design flaws fixed. Remove the implicit casting and obscure integer promotions. Make spiral rule hold everywhere instead of being able to put const at the beginning of the declaration. Make `sizeof()` return a signed type. Don't allow mixed signed/unsigned arithmetic. Make variables/functions private by default i.e. add `public` to make public instead of `static` to make private.

Keep the preprocessor and for the love of god make it easy to invoke the compiler/linker directly so I can write my own Makefile.

[baranul]

Arguably, what you describe, is closer to what C2 was/is[1]. By the way, C2 is still alive, for those that care to look.

C3 (link[2]) is a fork of/inspired by C2, which appears to have incorporated a lot of Odin and Jai "flavoring". In the case of both C3 and Odin, it can be argued that part of their popularity is that Jai isn't publicly released. Consequently, they seem to pull in a lot of the crowd, that would be attracted to Jai. Another aspect of this, is the more C3 promotes itself (whether intentional or not), the more likely C2 will get faded out. Many will likely think C3 is the next iteration of C2 or simply know the name more, because pushed on HN and other social media.

[1] https://github.com/c2lang/c2compiler

[2] https://github.com/c3lang/c3c

[lerno]

Isn’t it a somewhat unfair characterization that ”C3 promotes itself more” and ”is pushed on HN and other social media” and that because of this C2 for some reason experiences harm?

C2 is over 11 years now. C3’s recent breakthrough this last half year is unlikely to have had much impact on its ability to grow the last 10 years.

[baranul]

This thread literally meets the definition of self-promotion. So no, don't think the characterization is unfair.

If your programming language had a different name, then I would agree, but it doesn't. Many will assume that C3 is the next iteration of C2 and that it's outdated, despite the fact that C2 is still in development.

[lerno]

I asked and got Bas's[1] permission to use C3 as a name when I first started out.

[1] Bas van den Berg, the author of C2.

[jstimpfle]

There is no spiral rule, that is a misconception. Look up how "declaration follows usage" in C. E.g. https://eigenstate.org/notes/c-decl

While this rule has become somewhat diluted when C developed and gradually took on features from C++ (like types in function parameters), it's still very helpful to understand the guiding principle. (But those inconsistencies that crept in over time are also the reason why newer languages don't do that anymore).