| |
|
June 4, 1998
Often cited, sometimes false
Bill Schweber, Technical Editor
Often, conventional wisdom is the product of a
single story repeated so often, and by so many people, that it takes on the aura of
validity.
I'm sure you've all heard the legends about the
much-maligned but omnipresent QWERTY keyboard. Because early typing mechanisms would jam
at faster speeds, Christopher Sholes and the co-inventors of the mechanical typewriter
laid out the QWERTY keyboard in 1867 in an effort to slow down the typist. Although a
1940s study by the Navy deemed a keyboard layout designed by Professor August Dvorak in
the 1930s more efficient than QWERTY, the public never adopted Dvorak's design. We'd
probably get much more work done today if we had adopted Dvorak's layout, now easily
achievable with a software-mapped keyboard. So many people have repeated and cited the
QWERTY stories that people rarely give the tales' veracity a second thought.
The problem with the QWERTY legend is that it may be
only partially correct. Researchers looking into the story as a starting point for some
unrelated economic analysis now say there is no evidence that the keyboard was designed to
impede typing and that Dvorak, who stood to gain both reputation and royalties if the
public adopted his layout, possibly conducted the Navy study himself. We'll probably never
know the real facts in this matter, because the inventors did not leave design-rationale
documentation.
The same framework applies to assumptions you make when
designing and debugging your system. You probably make many large and small
assumptions--some consciously, some subconsciously. You probably follow guidelines for IC
or board layout, for example, that should increase your chances of success. You follow
rules that seem to work for applying power-supply and EMI bypassing. And, when you debug,
you probably begin with some assumptions about what is happening, what should be
happening, and why some things are not happening.
The real challenge for a good designer is to know when
and how to step back and reassess everything. Is what you think is happening really there,
or is it a combination of instrumentation and observation interaction? Are some of the
design guidelines you followed based on hearsay rather than rigorous validation? Are you
using differing standards of judgment for what you see in the same way a critic might
review a just-opened restaurant, saying it will get better with time, but review a theater
performance on its opening night and advise against seeing it, because it had bad acting.
Often, you'll find that conventional wisdom isn't wise
but is the product of a single story repeated so often, and by so many people, that it
takes on the aura of validity. For example, most studies on shock-therapy effects of
various voltage and current levels on humans have taken on this aura; despite numerous
literature citations, most references point back to just one or two animal studies done
long ago. HL Mencken found out--the hard way--about putting the message back in the
bottle. In the 1920s, he made up a history of the bathtub and presented this story as
fact. People repeated this story so often that it eventually found its way into
encyclopedias, and Mencken spent years trying in vain to correct the mess he made.
|
