Re: Morally equivalent

[Thibaut Verron <thibaut.verron@gmail.com>]

Le lun. 17 oct. 2022 à 09:50, <tomas@tuxteam.de> a écrit :

> On Mon, Oct 17, 2022 at 09:44:15AM +0300, Eli Zaretskii wrote:
> > > From: Akib Azmain Turja <akib@disroot.org>
> > > Cc: help-gnu-emacs@gnu.org
> > > Date: Mon, 17 Oct 2022 10:49:16 +0600
> > >
> > > This should be:
> > >
> > > --8<---------------cut here---------------start------------->8---
> > > This is almost equivalent to (setf PLACE (cons NEWELT PLACE)),
> > > except that PLACE is evaluated only once (after NEWELT).
> > > --8<---------------cut here---------------end--------------->8---
> >
> > Why not just "This is similar to"?
>
> Back to topic, yes. It seems that this phrase is too jargon-ish to be
> easily understandable to everyone.
>

Isn't there the technical term "functionally equivalent" for something that
will have the exact same outcome with possible differences in
implementation?
If the goal is to reduce the jargon, it could just be "the end effect is
the same as that of (...), but PLACE is evaluated only once".

It's not rigorously true, because the outcome will be different if
evaluating PLACE has side effects. But imo this is such an outlandish
scenario (if it can even happen in the first place) that it doesn't need to
be the primary focus of the docstring of push. And someone who is dealing
with such a scenario should be able to infer the potential consequences
just by being reminded that PLACE will be evaluated as many times as it
appears.

I personally find "morally equivalent" or "almost equivalent" clearer than
"similar": "similar" could refer to any other aspect of the object, besides
the functionality.

Best wishes,
Thibaut