Bruce wrote:
[details deleted]
> However, in the totally-overlapped case, half the users are decidedly less
> happy than the other half, and the person who had the bad luck to get
> stuck at the end of the queue had to wait 11 seconds for the last record
> to come back. This is probably a very unhappy person (that's her on the
> phone right now). In the nonoverlapped case, every user got some part of
> their answer back within the first second even though they had to wait an
> average of 10.5 seconds for their entire result. That means they're less
> productive but a lot happier.

But, if the user can make use of the first part of the answer before the
entire answer is delivered, then the the entire group of users becomes *more*
productive.  Well-designed web sites are an example of this.  You can begin
reading them as soon as data begins to be displayed, so you are more
interested in "time-to-initial-response" than you are in "time-to-completion-
of-response".

Data entry operators tend to be more efficient with smoother response (they
settle into a rythim [gee, I hope I remember to spell-check this]) than they
are with jerky-but-overall-faster response.

---
Michael D. Hensley       | "Support Bill of Rights Enforcement"
Allegro Consultants Inc. | mailto:[log in to unmask]
408/252-2330             | Visit scenic http://www.allegro.com