Re: 7 logical-xor implementations in source tree

[Andy Moreton]

On Tue 30 Jul 2019, Philippe Schnoebelen wrote:

> On 2019/07/30 11:36, Alan Mackenzie wrote:
>> Less importantly, I don't like iff being used in this way.  I'm not sure
>> why.  
>
> Pros: (1) short (like and & xor)
>
>         (2) exactly what we mean when writing assertions in code
>
>            e.g.      (cl-assert (iff (listp arg) (listp res)))
>
>             e.g.     (should (iff  (= b (max a b))   (<= a b)  ))
>
> Cons: (1) name not known outside of mathematicians;
>
>           (2) is an acronym
>
>           (3) is not purely boolean, has order of evaluation, allows
> side-effects (like and & or)
>
> --Philippe

Please do not add `iff' as a symbol under any circumstance.

Short names must be universally understood by readers of the code with a
wide range of expertise and experience. A short name must also be needed
due to frequent use in typical code. This usage meets neither test.

    AndyM