bug#13592: 24.2.92; C-s prompts with "Multi I-search:" today

[Eli Zaretskii]

> From: Juri Linkov <juri@jurta.org>
> Cc: 13592@debbugs.gnu.org
> Date: Fri, 01 Feb 2013 01:58:56 +0200
> 
> >> This is like the prefix "Regexp" informing the user that the search
> >> is regexp-based, and other informative prefixes as well.
> >
> > It is OK to show "Multi" when the user explicitly invokes
> > multi-isearch.
> 
> There is less need to show "Multi" when the user explicitly invokes
> multi-isearch, because the user already knows that multi-isearch
> is active.

By that token, we should eliminate all the other prefixes as well,
like Regexp, Word, etc.  Or maybe even the I-search prompt as well.
After all, "the user already knows that isearch is active", right?

I think this is exactly the opposite of a good UI: these prompts give
the user a confirmation that her keystrokes (M-c, M-r, etc.) were
received and acted upon.  Did it never happen to you that you were
unsure whether you pressed a key firmly enough for Emacs to get it, or
pressed the right key?  These prompts are very valuable in these
situations.  Likewise, if you get distracted to something else (e.g.,
outside of Emacs) and come back later to continue the search.  Or when
you are working over a slow line.  There's any number of situations
where the prompts are useful, provided that the user actually
requested those actions, and will _expect_ the prompts in response.

> OTOH, when multi-isearch is activated implicitly, it is more important
> to inform the user about this case.  This is like saying to users:
> "Beware, this is not a simple I-search".

The "Beware" thing is inappropriate in this case, because this
"non-simple" search is exactly identical to a "simple" search,
_except_ when it's about to cross to the next file.  Until that point,
the extra prompt is a distraction at best, because it is _unexpected_.

Anyway, I see that I cannot convince you, no matter what I say.  So
let's leave this at a disagreement.