Re: outline/allout/overlay performance

[Thien-Thi Nguyen]

Ken Manheimer <address@hidden> writes:

> that looks like a nice feature.  i guess you're suggesting that
> outline-flag-region could enhance the overlays it sets with the
> presence of some customization variable.

close.  whether or not `outline-flag-region' enhances the overlays
is not my point.  rather, that the method for user enhancement of
those overlays be similar enough so that (ideally) the function
set as the value of one of these variables could be also used as
the value of the other.

> for now, however, i'm taking a minimalist approach and sidestepping
> generalization of outline-flag-region - using it as is, and employing
> defadvice so that an allout-specific isearch-open-invisible function
> in allout-mode is in allout-mode.

IMHO, congruence is a type of long-term minimalism, so we are not so far
apart aesthetically as much as temporally...

i'm not suggesting generalization of anything.  i'm suggesting the
conventionalization of `make-overlay' usage in similar (conceptually
speaking) contexts.  already, hideshow and outline share similar names
for similar operations, similar keybindings when using outline as a
minor mode, and similar algorithms for setting and tracking parent-child
relationships (this last assertion is a guess -- i suspect the set of
"tree ops" is not so large as to support wild variability).

thi