[temp.constr.concept] Avoid talking about normalization of constraints

[lprv]

No description provided.

[jensmaurer]

I think this is intentional. We do some weird normalization here.

[lprv]

In that case, how about:

Checking whether CI'' is satisfied can lead to substitution into, and normalization of, concept-ids of other concept-dependent constraints.

"Normalization of concept-dependent constraints" feels wrong to me because normalization is the process of turning expressions into constraints; a concept-dependent constraint is not itself an expression.

[tkoeppe]

@hubert-reinterpretcast, @opensdh could you speak to whether the status quo needs to be cleaned up here?

[tkoeppe]

Or @cor3ntin?

[cor3ntin]

I think normalization is what we mean here. Yes, we will also check satisfaction recursively but this is not notable. however, the fact that concept dependent constraints would be further normalized as they become non dependent is notable, as it is unusual.

status quo is therefore preferable.

[tkoeppe]

Thank you!

[lprv]

So am I wrong in saying that "normalization of [...] constraints" is a category error? Because that is the main concern of this PR, and I don't think it was addressed by the above explanation.

We could find a form that keeps the term "normalization" in the sentence without using it incorrectly; my previous comment provides one suggestion.

[hubert-reinterpretcast]

So am I wrong in saying that "normalization of [...] constraints" is a category error?

You're not wrong. We will need to see how a concept-dependent constraint is structured to see what is getting normalized.

[hubert-reinterpretcast]

@lprv, please open an issue re: the category error here: https://github.com/cplusplus/CWG/issues