I agree with your conclusion, but possibly not your reasoning.

C++ should try to be more beginner friendly. We should be willing to pay
_some_ price ("work for ourselves") for that. The question is how much.
And note that the work can also help C++ "expert(s)".

But fundamentally, C (not just C++) is a language built by combining small
expressions. a = b is an expression, that does the same thing, regardless
of whether it is in an if or not. Making a special-case for "inside an if"
is more complicated, not simpler.
Special-casing is the antithesis of language design. Special-casing is NOT
beginner friendly. It appears so on the surface, but just leads to too
many cases to memorize, so long term (more importantly - medium term) it is
unfriendly.

Once the beginner learns this fundamental rule of expressions and rule of
"combine simple concepts/blocks into bigger blocks", much of the language
becomes _easier_.

So it is not a question of beginner vs expert it is beginner vs
just-past-beginner. Simplicity of "fundamental blocks combined together"
helps the just-past-beginner. The expert can do either simplicity or tons
of special cases, they don't care, they are experts.



On Thu, Oct 24, 2019 at 3:40 PM Nicholl, Ryan via Std-Proposals <
std-proposals_at_[hidden]> wrote:

> I have to strongly oppose the suggested change. It breaks old working code
> just to make it harder to make beginner mistakes. C++ is expert oriented,
> we shouldn’t be making work for ourselves just to make the language
> slightly cushier.
>
>
>
> *From:* Std-Proposals <std-proposals-bounces_at_[hidden]> *On Behalf
> Of *Brian Bi via Std-Proposals
> *Sent:* Wednesday, October 23, 2019 12:44 PM
> *To:* std-proposals_at_[hidden]
> *Cc:* Brian Bi <bbi5291_at_[hidden]>
> *Subject:* Re: [std-proposals] Prevent Assignments from contextually
> converting to bool
>
>
>
> **External Message* - Use caution before opening links or attachments*
>
>
>
> On Wed, Oct 23, 2019 at 11:35 AM connor horman via Std-Proposals <
> std-proposals_at_[hidden]> wrote:
>
> This is my proposed to solution to one of the worst logic errors in any
> number of languages, being:
>
> if(i=0)
>
> vs.
>
> if(i==0)
>
>
>
> The proposal is to tweak the wording of contextual conversions to bool,
> prohibiting unparenthesised assignments from being eligible for contextual
> conversion. Code that intends to do this can simply put parenthesis around
> the assignment or use an init-statement, and code that does this without
> intention (i.e. intends the comparison) would be elevated from a logic
> error (the worst kind honestly) to a compile-time error.
>
> --
> Std-Proposals mailing list
> Std-Proposals_at_[hidden]
> https://lists.isocpp.org/mailman/listinfo.cgi/std-proposals
> <https://urldefense.com/v3/__https:/lists.isocpp.org/mailman/listinfo.cgi/std-proposals__;!9cTk8lEYrHI!9qEH-MtCUe6TcGGK6ll0wRqzY4HrdNbgYNjlM15ICBh_7WSj9Jj26V7ecf7H10A$>
>
>
>
> You can achieve this effect using compiler warnings. For example, with GCC
> and Clang, you can use the flag -Werror=parentheses to treat this as an
> error.
>
>
>
> I do not think we should add more rules to the language that discriminate
> among expressions on the basis of properties other than their type and
> value category.
>
>
>
> --
>
> *Brian Bi*
> --
> Std-Proposals mailing list
> Std-Proposals_at_[hidden]
> https://lists.isocpp.org/mailman/listinfo.cgi/std-proposals
>

-- 
Be seeing you,
Tony