Re: Three Years of Computer Time

[Eric Lyons <acad!throop!acad!earth!eric>]

It is interesting to pose the question, "are there useful and/or interesting
task which can be done in the background...".  One could assert that ALL tasks
whose results are not instantaneous could and _should_ be done in the 
background, with the user in control of both priority and level of completion.

One of my favorite papers on computing is by none other than Fred Brooks, who
has coined the term (maybe he didn't coin it, but he sure likes it) Adaptive
Refinement as an approach to all computing.  His idea (not new, nor unknown
I suppose) is that a computer user should see progressively better (more
accurate, more detailed, more realistic) data the longer a process runs.  We
use this technique, for example, in AutoCAD "drag" code.

The benefit is not only an improvement in usability, but potentially HUGE
productivity increases as well.  Often one can understand intuitively whether
a process is proceding in the right direction with relatively poor cues,
at which point the direction can be changed rather than waiting for it to 
finish only to find out that it's wrong.  I believe this is useful to keep
in mind when designing software.

But I don't know what I'd run in the background on my Sun, unfortunately.
Please excuse the diversion.

Eric.