It ought to go without saying that the goal of diversity of gender or ethnic origin does not generally conflict with excellence in research. For instance, in recent years my department has interviewed several women candidates who were uniformly superior to their male counterparts.

However, in specific faculty searches it may be that the potential research stature of a certain white male candidate is perceived as exceeding that of a certain female or minority candidate. The latter may be stellar, but the former’s intellectual light shines just a bit brighter. If the discrepancy is comparable to the rather high level of uncertainty inherent in measuring a candidate’s potential, some may invoke the additive argument.

However, this argument seems to rest on two questionable assumptions: departmental excellence (however measured) is the arithmetic sum of the individual levels of excellence of its faculty members; and the success of an individual researcher is independent of the surrounding environment.

Both are wrong. Excellence in research (individually or across a department) is a nonlinear function of interdependent factors. For instance, in a department that makes itself attractive to a broader pool of graduate students through the composition of its faculty, all researchers benefit from the resulting potentially improved quality of the department’s student body. This also holds when attracting new colleagues, including so-called superstars. When female or minority candidates are at, say, the top of the list in a particular search, they (like everybody else) also consider a department’s environment when choosing which job offer to accept. Moreover, a more welcoming, collegial, diverse faculty often leads to better and more frequent collaboration, as well as to more vibrant research.

The question is not whether to compromise between excellence and diversity but how best to foster excellence, with diversity a part of the equation.

carlo tomasi, durham, nC

said computer science is “the only
scientific community that considers
conference publications as the prima-
ry means of publishing our research
results,” asking, “Why are we the only
discipline driving on the conference
side of the ‘publication road?’”
As an old timer, I can say that in the
early days, there was a belief (conceit
might be a better word) that the field’s
pace of discovery was happening so
quickly that only conferences, with
subsequent prompt publication of pro-
ceedings, could communicate results
in a timely manner. As a corollary, the
traditional peer-reviewed published
literature review fell behind, as it was
relieved of temporal pressure through
the published proceedings.

These days, the pace of discovery in the biological sciences, including molecular biology, genomics, and proteomics, far exceeds that of computer science. Yet the gold standard of publication in archival journals continues. It is the ultimate irony that computer science, along with various disciplines in the physical sciences, employs the tools developed by computer scientists to ensure timely dissemination of research results through the online editions of their publications. Science, Nature, Cell, and other leading journals routinely present their most important articles in online form first. If, perhaps, computer science would make greater use of its own tools, the shoemaker’s children would no longer go barefoot, and published proceedings would fade into its proper historical niche.

stuart Zimmerman, houston, tX

more to celebrate in RDBms history
Gary Anthes offered good reporting but
also some serious errors concerning
pre-RDBMS history in his news article
“Happy Birthday, RDBMS!” (May 2010),
saying “In 1969, an ad hoc consortium
called CODASYL proposed a hierarchi-
cal database model built on the con-
cepts behind IMS. CODASYL claimed
that its approach was more flexible
than IMS, but it still required program-
mers to keep track of far more details
than the relational model did.”
Please compare with the following
basic facts as reported in Wikipedia:
“In 1965 CODASYL formed a List Pro-
cessing Task Force. This group was
chartered to develop COBOL language
extensions for processing collections
of records; the name arose because
Charles Bachman’s IDS system (which
was the main technical input to the
project) managed relationships be-
tween records using chains of pointers.
In 1967 the group renamed itself the
Data Base Task Group and in October
1969 published its first language speci-
fications for the network database
model, which became generally known
as the CODASYL Data Model.”
The Integrated Data Store (IDS) has
been in continuous productive use
since 1964, running first on GE 200
computers. In 1966, it began support-
ing a nationwide, 24/7, order-entry
system (OLTP). And in 1969, running
on the GE 600, it began supporting a
shared-access (OLTP) database, com-
plete with locks, deadlock detection,
and automatic recovery and restart.

 

Wrong side of the Road In his Editor’s Letter “Revisiting the Publication Culture in Computing Research” (Mar. 2010), Moshe Y. Vardi

Communications welcomes your opinion. To submit a letter to the editor, please limit your comments to 500 words or less and send to [e-mail].

© 2010 ACM 0001-0782/10/0700 $10.00