A number of you have seen the rather misleading article published in
yesterday's Washington Post. The best I can say about the article is
that it depicted MSR as a pretty irresistible place to work - which is,
of course, true! I guess on the plus side this may mean a lot more
great people beating a path to our door.
On the negative side, though, the reporter bought into the clich=E9d
notion that whatever Microsoft does must somehow be bad for somebody and
in this case decided that MSR was "picking clean" the academic world and
"eating the seed corn" of tomorrows PhDs. In order to make this point,
quotes were taken out of context or simply "rearranged" and facts were
bent. Both Jim Morris and Raj Reddy, for example, have disavowed the
article and their quotes in it.
Because I prize our relationships with our academic colleagues I decided
that I didn't want to see this distorted view of reality stand without
comment. In particular I wanted to correct the many factual errors and
confront the unfounded innuendo in the article. I have for that reason
submitted this letter to the Editor of the Washington Post. I'm certain
it will be heavily "edited" by the Post's editorial staff before
publication, but here is the original version. It contains a number of
facts and statistics that might come in handy for those who may be asked
questions about the article or about MSR's role in the community.
Dear Editor,
The article "Microsoft Skims Off Academia's Best for Research Center"
painted a very misleading picture. By selectively including comments
and quotes that fit it's "brain drain" premise and by including or
implying incorrect information it left a very false impression of
Microsoft Research (MSR) and it's relationship to academia.
The underlying premise that the academic world is being drained of its
faculty by MSR in particular or the industrial research world in general
is demonstrably false. According to the Computing Research Association's
1998-1999 Taulbee study, less than 2% of all tenure track faculty leave
each year for non-academic positions (roughly 50 out of 3000) and that
number has remained largely the same for much of this decade. Moreover,
the number of total faculty counted in the Taulbee study is currently
increasing, not decreasing. The total number of tenure track faculty is
projected to increase by more than 6% in 1999-2000. In fact, the total
number of tenure track faculty is projected to increase by more than 23%
over the next 5 years. Undergraduate CS enrollments have mushroomed
since 1995, the number of students receiving PhDs in Computer Science is
increasing and, most importantly, new PhD enrollments are up more than
23% in just the last year (or by more than 300 students). These are
hardly statistics that support a "brain drain" hypothesis.
Looking beyond the general point about "academic brain drain", the
article incorrectly implied that Microsoft Research's growth has largely
been at the expense of academia. For example, in the third paragraph it
states that MSR is "aiming for 600 'faculty'" by next year. While it is
accurate that Microsoft Research plans to grow to 600 researchers, the
use of the word "faculty" in quotes is very misleading. It implies that
these 600 people are "faculty" in the sense of "teaching faculty" at
Universities which supports the notion that somehow Universities will be
"picked clean" by MSR hiring. The reality is very different. 90% of
Microsoft researchers have come from Industry, were hired by MSR as new
PhDs or have come from non-teaching research positions.
As a case in point, when Raj Reddy is quoted as saying that "15 or 20
top people" have left to go to Microsoft, the article implies that these
people were teaching faculty. There were not. Most of the people Raj
is talking about were PhD students finishing their degrees, post-docs or
staff. Microsoft has hired only 4 teaching faculty from CMU in 7.5
years. This obviously cannot have been a major drain on an organization
the size of CMU's School of Computer Science - which has more than 750
faculty, staff and graduate students. CMU actually has more faculty and
a larger budget today than it did when Microsoft Research was founded in
1991.
In fact, the entire discussion of how many teaching faculty Microsoft
Research has hired was based on bad data. Only around 10% or roughly 40
of Microsoft's researchers held academic teaching positions of any kind
-- not just tenured full professors as the article asserts -- when they
were hired by MSR. This is a negligible number when compared with
thousands of professors in the Computer Science academic community -
especially when one realizes that these people were hired over a period
of more than 7.5 years. On a yearly basis, MSR hires represent less than
0.2% of total faculty or roughly 6 per year out of a total of 3000. Far
more faculty have been leaving Universities to go to startup companies
(and thus leaving the research community altogether) than have been
coming to MSR.
Even the article's quotes may not have been accurate. Raj Reddy has
written to the Editor of the Washington Post to express his dismay at
the tone of the article. Jim Morris has said that he was "badly
misquoted".
Realistically, people move around a lot in the research world. CMU
hires it's faculty from other institutions just as they hire from CMU.
Other Universities also hire senior faculty from each other or bring
them in from established Industrial research labs. Jim Morris (who was
quoted in the article) taught at Berkeley and worked at Xerox PARC
before going to CMU. A number of other CMU faculty came to CMU as senior
hires from other institutions (e.g. Dana Scott, John Reynolds). David
Dobkin (who was also quoted) taught at Yale and Arizona before being
hired at Princeton. This is normal and healthy and has always been the
case for as long as I have been involved with the field. According to
the Taulbee study, almost twice as many faculty move from one position
to another each year as leave to go into non-academic jobs.
Or course, a story which says that "this year Microsoft will only hire
roughly 6 University professors which is 0.2% of the total" isn't very
newsworthy. The fact that despite these hires the total number of
professors in Universities is projected to increase by more than 6% or
over 200 professors and that PhD enrollments are up 23% is much less
interesting than word images of Microsoft "picking clean" a helpless
academia. Although dramatic, the story as written is unsupported by the
facts.
Founded at a time when other Corporations were downsizing their research
in Computer Science, Microsoft Research has become one of the strongest
CS research organizations in the world. It was built on a University
model of openness and unfettered investigation and it has been a strong
supporter of University research in the form of millions of dollars of
unrestricted research grants per year and a growing graduate student
fellowship program. It has most assuredly not "spurred a university
brain drain". In fact, the positive example it provided may have been
at least partially responsible for a resurgence in Computer Science
research in academia and industry.
Sincerely,
Richard Rashid
Vice President
Microsoft Research
Microsoft Corporation