[woodruffw]

“There’s a standard, but it’s performative” is a different argument than the top level one.

But even if we accept that, it doesn’t seem like a good comparative argument: anybody who has written a nontrivial amount of C or C++ has dealt with compiler-defined behavior or compiler language extensions. These would suggest that the C and C++ standards are “performative” in the same sense, but repeated claims about the virtues of standardization don’t seem compatible with accepting that.

[cwzwarich]

The original purpose of the C standard was to solve the problems created by the diversity of increasingly divergent implementations of C. They studied existing behavior across systems, proposed new language constructs, and it was generally a success (look at the proliferation of C in the 90s across many different systems and architectures).

The actual informal semantics in the standard and its successors is written in an axiomatic (as opposed to operational or denotational) style, and is subject to the usual problem of axiomatic semantics: one rule you forgot to read can completely change the meaning of the other rules you did read. There are a number of areas known to be ill-specified in the standard, with the worst probably being the implications of the typed memory model. There have since been formalized semantics of C, which are generally less general than the informal version in the standard and make some additional assumptions.

C++ tried to follow the same model, but C++ is orders of magnitude more complex than C and thus the standard is overall less well specified than the C++ standard (e.g. there is still no normative list of all the undefined behavior in C++). It is likely practically impossible to write a formal specification for C++. Still, essentially all of the work on memory models for low-level programming languages originates in the context of C++ (and then ported back to C and Rust).

[tialaramex]

Well, the memory ordering model was developed for C++ and is used in C and Rust. But e.g. C++ does not have a pointer provenance model, which is arguably almost as important in this context. It turns out one of the things we really care about with multi-processing, and thus memory models is linked list hacks, and those only work if you have provenance rules, which uh, C++ just has a shrug emoji where the provenance rules would go so that's not great. C has an ISO document, although it's not part of the ISO C standard it's just another document so far, but Rust has specified provenance.

Also, the C++ ordering model is defective in the sense that while it offers the orders we actually use it also offers an order nobody knows how to implement, so it's basically just wishful thinking. For years now the C++ standard has labelled this order "temporarily discouraged" as experts tried to repair the definition and C++ 26 is slated to just deprecate it instead. Rust doesn't copy that defect.

[csande17]

They have a document that they sometimes describe using the word "specification", but its README clarifies that it's not actually a specification:

> The FLS is not intended to be used as the normative specification of the Rust language

[woodruffw]

To be clear, my argument doesn't hinge on whether FLS is a normative specification of Rust or not. The argument is that being "specified" is neither necessary nor sufficient for language maturity or quality.