Subject: Re: [HM] History of mathematics BY whom
From: Dinesh Maheshwari (dsm@cypress.com)
Date: Tue Feb 29 2000 - 17:57:44 EST
Dear HM listmembers,
I have not been a referee for any history of mathematics publications,
but I can relate my experience as a referee for several publications in
the field of my professional responsibilities, which is "Integrated circuits
computer aided design (ICCAD), also referred to as "Electronic design
automation (EDA)".
The ICCAD field has some important parallels with the field of "history
of mathematics"- it is very broad and diverse (covers basic electrical
engineering, semiconductor processes and physics, IC design
methodologies, statistics, efficient techniques for solving second/third
order differential equations in 100,000 or more variables, multi-valued
logic optimization and factorization, integer & real linear programming,
dynamic programming, graph theory, event queueing, compiler theory,
efficient memory-image database techniques, shape recognition &
manipulation techniques, real-time graphical user interface systems,
distributed computing etc. and several NP complete problems) and
requires in-depth abilities in each one of these sub-fields for the
complete IC CAD flow.
Although an ICCAD professional is expected to have some expertise
in each of these sub-fields, the person is usually capable of critiquing
only a handful of the above-mentioned sub-fields.
The field is very lucrative (just one copy of one piece of software for
say "routing" costs as much as a quarter of million US dollars and just
one copy of the complete ICCAD flow in software alone can cost as
much as 5 million US dollars)and attracts the best talents who try to
prove their worth in the two most prestigious conferences in the field
-ICCADC and DAC. In essence, being accepeted/rejected for these
two conferences makes or breaks carreers in the industry as well as
in the academia.
The submissions to the above two conferences are "double-blinded"
by the editor - anonymous authors and anonymous referees. The
technical proceedings for the conferences however list all the referees
for the conference. I have been a refree for both of these conferences
for some years and was quite disappointed with the "double-blinded"
approach. I noticed that even though the referees themselves were
some of the most qualified people in the industry and academia, there
were invariably about 15% of papers in the 10th or lower percentile
category (of all the submitted papers), in terms of originality & quality,
that were still accepted.
So I would be rejecting a paper from, say person A, because it would
not meet the originality and quality requirements while another paper
from person B would be accepted by someone else even though it was
much poorer than A's submission. And this, I thought, was unfair to
person A.
Talking to the other referees I found that the reasons for such a
discrepancy in the quality were -
1. Referees getting papers which were not exactly in the area of their
core competency but were rather in their peripheral area of competency.
2. Referees not taking enough time to evaluate the work.
My proposal for rectifying such a scenario was (and is) to publish the
names and the evaluations of each of the referees for all the papers
- rejected or accepted; with the names of the referees and the summary
of their evaluations published in the technical proceedings, alongside the
paper, for the ones that were accepted.
And then a rating mechanism from the conference audience that would
rate each of the papers presented so that the editor could prune out the
referees for the worst rated papers.
Also knowing the referees of each of the papers would enable the
prospective readers, employers etc. to base their decisions on the basis
of the credibility of the referees for the paper rather than the prestige
of the conference.
Perhaps, a similar mechanism might work for the purpose of refereeing
for history of mathematics also; but then not being a referee in this area,
I cannot be sure.
Best Regards,
Dinesh
-- Dinesh Maheshwari Advanced Design Methods Cypress Semiconductor San Jose, CA, USA
This archive was generated by hypermail 2b28 : Tue Feb 29 2000 - 18:27:01 EST