Re: emacs rendering comparisson between emacs23 and emacs26.3

[Eli Zaretskii]

> From: Richard Stallman <address@hidden>
> Cc: address@hidden, address@hidden, address@hidden
> Date: Tue, 31 Mar 2020 22:07:38 -0400
> 
>   > When we know a new feature causes significant slowdown, we either fix
>   > it, or provide a way to disable or work around it.  The problem is, we
>   > don't always know there's slowdown, as it frequently happens only in
>   > specific rare use cases.
> 
> 1. If the slowdown cases are truly rare, I don't think they will bother
> users very often.

A frequent case is that the use case is rare on the average, but for
some user(s) it is not so rare.

> 2. I think we should be able to anticipate which changes _might_ cause
> noticeable slowdowns.  They will be changes that will come into play
> frequently in basic activities such as cursor motion, basic editing
> operations, or redisplay.  We may not anticipate 100%, but let's start
> trying.

I think we do try: patch review frequently has these aspects
discussed.  But it is a well known fact that humans are lousy
predictors of code speed, which is why the advice is always to profile
before deciding where to optimize.  I think we miss the slowdowns for
the same reason.