CLIM mail archive
[Prev][Next][Index][Thread]
Re: Moving objects and incremental redisplay
Date: Tue, 28 Sep 1993 14:40 EDT
From: Curt Eggemeyer <curt@eraserhead.jpl.nasa.gov>
Just to put my $.02 in, from what SWM said.
>CLIM's incremental redisplay facility operates at too low a level to be
>as efficient as the 4 steps above. CLIM doesn't know, for example, that
>you just moved one little object. It has to go figure it out by looking
>at the whole output record tree.
Maybe in CLIM version 3 ... the incremental display facility would have
an option of working with either a top-down (as it is now) or a bottoms-up
way. Many of my particular display domains would really exploit the
bottom-up capability if it were there.
Can you design a language that specifies how the program tells CLIM
about this? I'm serious.
The big problem is, quite simply, that incremental redisplay is a
*really hard* problem. It's easy to think of a bunch of special cases:
the Emacs redisplay special case, the line-at-a-time special case, the
simple graphical editor special case, the spreadsheet special case. And
it's easy to write code to implement any of these well. It's relatively
easy to specify the general case -- that's what UPDATING-OUTPUT does.
But it's hard to implement that well, and it's hard to figure out from a
general specification which of the special cases you might fall into.
Several of us have been grappling with this on and off for years, and
none of us has a solution.
Follow-Ups:
References:
Main Index |
Thread Index