|
Address to my
students
Reading the
literature almost always tempts me to conclude that the quality of the
standard paper is below average. Let me list a few of the standard
shortcomings of the standard paper:
- the nontechnical
prose is salestalk
- the technical
prose is sloppy and verbose
- the notations are
ill-considered and pompous
- the formalisms
adopted are clumsy
- the definitions
are unclear and incomplete
- the proofs are
omitted or a caricature of what they should be
- the problem
tackled is not worth solving (e.g. because it is easily avoided)
- the author fails
to challenge his tacit assumptions and to separate his concerns
- the problem can be
presented and dealt with better and in only a tenth or a fifth of
the space.
So far, so bad. Worse
often surfaces when one discusses these shortcomings with the author.
While agreeing in an abstract way, he yet defends his articles with
arguments in the following vein:
- the paper is
addressed to the XYZ community, which habitually expresses itself in
that style, deviation from which would result in failure to reach
his intended audience
- the shortcomings
don't matter because the intended audience knows how to resolve the
ambiguities (if it notices them) and how to supply what has been
omitted
- had he followed
the suggestions for improvement,
the paper would have
become so short that no one would notice that he had achieved
something
- in improved form,
the paper would be rejected by the editorial boards and program
committees of the XYZ community and, being untenured, the author
cannot afford the luxury of that rejection.
Having a solid germ
of truth, his defense sounds almost plausible. But we should not forget
the following:
- scientific
research as planned by managers is planned as if the available
budget is the limiting factor, whereas the problems worth tackling
are our scarcest resource
- the
publish-or-perish syndrome has opened the flood gates for the
write-only journals and the analogous speak-only conferences; there
are more of those than we can form competent editorial boards and
program committees
- education has
always been an up-hill struggle, and it is naive to expect from the
XYZ community gratitude for showing that its pet problems are
disposable
- life of a
scientist was never meant to be comfortable; exciting, yes, but
comfortable, no.
We should never
forget the crucial distinction between the salesman and the scientist.
It is the salesman's duty to please his customers or, if that is too
difficult, to fool them. It is the scientist's duty to raise our
abilities by increasing our standards for quality and effectiveness, to
show what can be done well by pursuing the just possible, and to
clarify ruthlessly,
independently of the fact that complexity sells better.
In these respects,
there is no room for compromise: the alternative is no less than
scientific corruption, and remember that the commonness of the latter
phenomenon does not make it respectable.
Of course, you are
free to join "for your protection" an XYZ community by
accepting its standards as law; but if you do so, know that the
scientist in you has been replaced by the party member.
Austin,
14 April 1986
prof.dr. Edsger W.
Dijkstra
Department of Computer Sciences
The University of Texas at Austin
Austin, Texas 78712-1111
United States of America |