mark@umcp-cs.UUCP (08/19/83)
I was on a human factors advisory panel today for a major
government (U.S.) agency, and the issue of how to use human
factors in the design process came up. There seem to be two
very different possibilities, one science (or close to it) and the other black
art (or close to it).
The first possibility is that at various stages in the design and
implementation of a project, a human factors team does an
analysis, critique, and comes up with recommendations. These
recommendations are then factored into the next iteration of the
design. For instance, you are working on a second generation
reactor control system. The human factors team analyzes what was
good and bad about the first generation, then makes human factors
recommendations for the second generation. Then the design team
comes up with a proposed system, factoring in the human factors
recommendations along with all the other constraints. This
design is human factors reviewed, the results are incorporated
in a revised design, and then detailed design is done, another
review, etc. etc.
The point of this scenerio is that the human factors people are
not on the design team, but are more akin to auditors. They can
thus do a rigorous job, performing little experiments on aspects
of the proposed system if necessary (and with sufficient
funding), etc.
The black art of human factors is actually incorporating human
factors into the system design. The system designers must know
a lot more than just human factors so cannot be dedicated human
factors engineers alone. Maybe one or more specializes in that,
but they must be more or less generalists. Why they choose one
design over another is the black art. It is one thing to
shoot down one or another aspect of a design after it is done,
but it is quite another to come up with the integrated design
which meets all the system requirements in the first place.
I don't think there is any resolution to this distinction.
The engineers who are on those design teams need to themselves
know enough human factors to come up with designs that are in the
ballpark, rather than relying on a human factors person to fix
them up. The human factors engineers need to become truly
knowledgable about the other aspects of the project engineering,
or be relegated to an audit function only.
--
spoken: mark weiser
UUCP: {seismo,allegra,brl-bmd}!umcp-cs!mark
CSNet: mark@umcp-cs
ARPA: mark.umcp-cs@UDel-Relay