RE: [External] : Re: can vertico approximate this ido setup?

[Drew Adams <drew.adams@oracle.com>]

>> One thing to note is that with completion-styles
>> (a b c) is that b won't kick in as long as a
>> delivers matches and c won't kick in as long
>> as b finds matches.
> 
> aha.

So it doesn't let you see matches for
one or the other.  And you don't know
which matched.

The set of matches is a function of
(1) the input domain, (2) your match
patterns, and (3) the matchers/styles.
You can't control #3 interactively,
and #1 is typically determined by the
command/code invoking matching. 

That #3 isn't accessible interactively
is OK for some use cases, but overall
it's an inherent design weakness, IMO.

Better to be able to control matching
interactively (and consequently know
what does what) - including #3 (even
#1 to some extent).

Be able to switch among style lists,
including singletons: (c), (a b c),
(b a), (a), (c f),... anytime during
matching/completion.

(Style categories were added, so #3
can at least be predefined now, in
effect allowing more than a single
`completion-styles' value.  That's a
start.)