[net.works] Responsiveness of the Lisa

byard@dca-ems (02/11/83)

From: Larry Byard (WSE-EUR) <byard at dca-ems>

We have read with interest about Lisa in the WorkS (thank you),
and in articles in Time, Newsweek, BYTE, InfoWorld, Businessweek,
etc.  At first glance, the machine looks(!) very impressive for
the money.  But then one sees hints here and there that all is
not well.  I refer to comments on the machine's responsiveness.
    Digressing a bit... Last summer we investigated XEROX's Star.
I first saw the Star at the Hannover Fair and was very favorably
impressed by a DEMONSTRATION of the machine.  I managed to sit at
the keyboard/mouse for awhile and was able to learn the basics of
the machine in about 30 minutes.  Soon thereafter, I started to
really notice how very slow the system was.  Nevertheless, at
that point, I was enthusiastic about its potential for meeting
our requirements.
    Arrangements were made to demonstrate the Star system to our
staff.  XEROX said they would set-up a faster Star, one with a
later version of the software and more memory.  They brought the
Ethernet, two Stars, a small laser printer, etc.  They did an
absolutely superb job of showing their products.  I must say that
most of the staff officers were INITIALLY, very favorably
impressed with the system.  We had to "drag" some people away
from the workstations--they were having that much fun!
    Although faster than the version I saw in Hannover, the Star
system was still very slow.  As one became more familiar with the
machine (and some of the razzle-dazzle began to wear off)--that
is, as the man part of the man/machine interface became more pro-
ficient and able to operate the various tools faster--the inada-
quate responsiveness of the system became more and more
untolerable.  To make things worse, the system didn't print or
send messages in background.  One had to literally sit there for
minutes(!) while documents were printed or transmitted.
    The XEROX reps responded to our comments about Star's speed
with promises that the system would be 'speeded-up.'  Mainly,
because it was so sluggish, we concluded that Star was not a
suitable solution to our needs.  In my opinion, XEROX put the
system on the market before it was ready.
    Do I see a repeat of the same sort of thing with the Lisa?
How many of those glowing reports are based on hands-on
experience?  How many were overly influenced by the geee whiz
factor, or by demonstrations which did not make the briefly men-
tioned responsiveness problems as apparent as they would be after
several days of actual use?  Will the user be very happy during
the first 30 minutes of exposure to the machine, while learning
its basics, and then, after using it for two or three days,
become completely disgruntled because the machine can't keep up
with him?  Is the Lisa ready for market?
    Regardless of the Smalltalk philosophy of designing it in
after the fact, responsiveness is a vital part of the ergonomic
equation.  In my opinion, an "ergonomic engine" (if I may coin a
term) presents a new problem, a "two-edged sword": the bit-mapped
display, conceptual hierarchy, modeless operation, and mouse make
tasks seem easier yet take more hardware and software overhead,
and time or speed to accomplish.  A tool for the mind, operating
at levels of abstraction more easily assimilated by the mind,
will doubtlessly have to respond in terms of real-time thinking
or it will soon bore and alienate the user (particularly, the
computer-naive user), no matter how "nice" it is and how easy it
is to learn.  That is, if it appears simple, it must be fast,
very fast.  Any comments?

Larry Byard
byard@dca-ems
WWMCCS System Engineering - U.S. European Command
Stuttgart, Germany

crp (02/16/83)

Human interfaces is an area that needs a LOT of work before all of
our human brethren are going to be able to use computers to do more
than frustrate themselves and generate incorrect results at rates
never before attainable.

One person from our group here has seen and played with a Lisa
and he says that once an application gets cranked up it is
quite peppy (the startup time, however, is something else).
I suspect that they put a Meg of memory on the machine standard
for exactly that reason, but a faster disk is a WHOLE BUNCH
more expensive.

I agree that if you make something SEEM simple,
it had better happen fast -- and that isn't a bad start
on some "design principles" for the average person using
a machine to accomplish some work for which the machine is
only a tool (rather than an end, as it is for most of us).