More on bad attitudes that gain dominance in TOC [2016]
by Oded Goldreich
[outline, July 2016]
Preface: Prior related opinion statements of mine include
What's currently wrong with STOC/FOCS,
and On Struggle and Competition
in Scientific Fields.
For several years now, I have been concerned about the
growing dominance (in TOC) of some bad attitudes
(while noting that this phenomenon is not confined to TOC).
I am referring to the violation of balance over three non-orthogonal axises.
- CONTENTS VS COMPETITION: In terms of our conferences,
the question is how focused we are on the actual contents
of the program of a venue versus its competitive aspects.
The latter surely exist and are healthy up to a certain extent,
but become damaging and pathological beyond some point.
At the extreme, our conferences turn into pure competitions,
and when combined with (2) they become weight-lifting competitions.
- CONCEPTUAL VS TECHNICAL: In terms of our conferences,
the question is how focused we are on the message communicated
by a scientific work, its meaning and its value
versus the perceived difficulty of reaching the
ability to communicate this message.
Indeed, a priori (as opposed to a posteriori)
trivial messages are of little value,
but the issue is of the tendency to dismiss
valuable messages due to their perceived non-difficulty
(e.g., lack of non-elementary mathematics).
- LISTENING VS CRITICIZING AND GRADING: Our reviewing process
takes the form of exam grading with reviewers rushing to
find weaknesses and score according to predetermined concepts,
instead of first listening to what the work has to say.
I am not promoting non-critical reading, but rather promote
careful reading focused at understanding the message that
the work tries to communicate, and turning to evaluation
after such an understanding was gained.
An aspect that is most related to the 1st concern is that submissions
are written with the reviewers in mind, rather than with the actual
readers in mind. This is related to setting acceptance to the venue
as the main (or sole) goal, rather than setting communicating to
interested readers as the main goal.
When a CFP talks
of informing the PC abouts the merits of the submission,
it encourages targeting the reviewers rather than the readers.
Furthermore, if the authors can communicate valuable information,
why should they communicate it to the PC rather than to all readers?
(I.e., why deprive the interested reader from such information?)
Clarification re (2):
I am not claiming (nor do I believe) that
there is a negative bias against conceptual contributions.
It is just that conceptual contributions do not count much,
when it comes to making decisions regarding the program of conferences.
That is, everybody pays lips service to conceptual contributions,
but very few actually give these contributions any (non-negligible) credit
when reviewing submissions; the focus is solely on the (perceived)
technical difficulty (a.k.a effort invested in weight-lifting).
For sure, if the focus is on weight-lifting
(and conceptual significance is disregarded),
then you get a technical strong program,
but one that does not excel in conceptual contents
(since these were not given much weight when making choices).
In order to do better on the conceptual front,
one must assign it some (significant) weight.
There are no miracles: if you try to optimize parameter P,
then you get a high P-value (assuming that you do a good job),
and if you make your selection while disregarding parameter Q,
then you are not going to get high Q-values.
[Apr'18]
Back to Oded's page of essays and opinions
or to Oded's homepage.