Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!seismo!rochester!kodak!ornitz
From: ornitz@kodak.UUCP (barry ornitz)
Newsgroups: sci.philosophy.tech
Subject: Re: Just a test
Message-ID: <854@kodak.UUCP>
Date: Tue, 19-May-87 12:03:32 EDT
Article-I.D.: kodak.854
Posted: Tue May 19 12:03:32 1987
Date-Received: Wed, 20-May-87 05:13:28 EDT
References: <7585@orchid.UUCP> <3439@sunybcs.UUCP> <1767@kitty.UUCP>
Reply-To: ornitz@kodak.UUCP (barry ornitz)
Distribution: sci
Organization: Eastman Kodak Co, Rochester, NY
Lines: 62
Keywords: Honeywell, Taylor, pilot plants, human factors engineering, qwerty
Summary: Human factors for process control computer systems

This posting is in response to Larry Lippman's posting about human factors
engineering for process control computer systems.  I am sure Larry (a kindred
spirit if I ever found one) already is aware of the following information but
it might be useful to start discussions.

I work in the research labs of the Eastman Chemicals Division of Eastman
Kodak in a group that designs and builds custom instrumentation.  We are also
involved in making our instruments easy to use by un-trained operators.  One
technique we have found to work well is to imbed a computer in an instrument
in such a way as to not let the operator even know a computer is present -
no menus, no icons, no CRT screen, no keyboard; just a few buttons or switches
and perhaps some lights, and generally a digital readout like a digital volt-
meter.  We often include an RS-232 connector in back where we can plug in a
terminal for maintenance or software changes but the terminal is rarely placed
in the operating environment.  Even though the software may be doing very
complex calculations and handling self calibration, etc., the operator is not
generally aware of what is going on inside the instrument.  We have found that
this technique works well for instruments that can have all of their software
ROM resident - in case of a glitch, the operator cycles power to the instrument
to reset.

As for large process control systems, most vendors offer exceptional graphics
for the process.  Honeywell and Taylor come to mind but there are scores of
other companies around.  Most of these systems have a "panelboard" simulation
mode where graphics are used to simulate the layout of conventional PID
controllers.  Trend displays are also simulated to look like strip-charts.
We have found, however, that after using the system for several weeks, our
operators generally move to more modern information presentation methods.
Tabular displays or process graphics with process variables displayed next to
the displayed process unit are common.

We found that several years ago, with a large Honeywell system, the operator
keyboards were sequential alphabetically, i.e. A B C D E F G
                                               H I J K L M N, etc.  Now we find
that conventional QWERTY keyboards are usually supplied.  With the advent of
cheap home computer systems, everyone seems to have been exposed to the QWERTY
layout.

When we first placed a computer control system in our research pilot plants,
the operators were resistant to its use saying that it could never control
the processes as well as they could.  After using it for several months, they
grew to like the computer since it relieved them of the drugery of manually
recording data and turning valves, etc.  Now they gripe if the computer is
down for even a short while, claiming that they can't run the complex processes
without the computer.  My personal opinion is that they were afraid that the
computer would replace them.  In actuality, it has freed the operators from
much of their daily drugery and allowed them to understand the processes
better that they were trying to control.  They now use the computer as it 
should be: as a tool.

I hope this can help stir up some interest in this group.  My apologies to
Larry if it was boring.

 -----------------
|  ___  ________  |
| |  / /        | |  Barry L. Ornitz       UUCP:...!rochester!kodak!ornitz
| | / /         | |  Eastman Kodak Company
| |< < K O D A K| |  Eastman Chemicals Division Research Laboratories
| | \ \         | |  P. O. Box 1972
| |__\ \________| |  Kingsport, TN  37662       615/229-4904
|                 |
 -----------------