A modest question: Should we abolish the PhD degree?

Jeff Bone jbone@jump.net
Tue, 28 Aug 2001 22:56:52 -0500


Russell Turpin wrote:

> Jim Whitehead asks:
> > If you know of a better vetting mechanism for determining
> > whether someone is capable of doing original research,
> > I'd like to know about it. ..
>
> Why not look at the papers they publish?

Because it's no guage of the depth of either the insight or relevance
of one's "research" in this industry.  The definition of "paper" is
too stringent as is what constitutes "publishing."  Peer review is an
imperative in our industry, but there are many alternatives other
than journals.  I've seen many an occasion when internal corporate
publications were better researched, more detailed / quantified, and
had more far reaching conclusions than the average dissertation.  But
those things never get seen by the public, even though they may be
reviewed by as- or more-qualified "peers."  Academia is a club in
which it's easiest for the leisure class to participate, and an
exclusive political club with its own bullshit politics.
"Publications" --- in the classic sense, in many journals --- are
merely a merit badge awarded by this club.

I'm all for Dick Gabriel's MFA in Software suggestion, as it appears
to realize there's more to computer science than writing an
impenetrably dense dissertation.  No disrespect for Dr. Turpin's
achievements or dissertation intended --- I've got a bound copy of
his dissertation, and it's a keeper --- if he lets me keep it. ;-)  I
penetrate it weekly. ;-) :-)

> And then:
> > Why do you separate paper writing from the PhD
> > process? For me they were inseparable.
>
> You make my point for me. If they are so inseparable,
> what is the point of anything else?

That's all good, if possible.  Often the most pertinent research ---
if being done under a corporate umbrella --- is unpublishable due to
IP issues.  What then?  How do you balance between your / your
employer's fiduciary responsibilities and your academic goals?  The
people most punished by this dichotomy are those that sit ahead of
the curve, who are not groping through a learning process in the
course of the PhD.  Often PhD committees are as inept as the patent
office at recognizing a simple repositioning of prior art.  Often the
people most punished by the current academic gauntlet are those that
are attempting to apply truly innovative insight and push forward the
knowledge curve.

> And not all of it has to do with
> learning all there is about a specific topic.

And therein lies the crux.  Most CS-oriented degree programs think
they are preparing someone for a career, when in fact all they are
doing is enacting a poorly-scripted shadow play of what happens in
the private sector.  I read *lots* of dissertations.  Just last
weekend, I threw away 6 filing cabinet drawers full of printed
academic research material in CS that I'd accumulated over the last
15 years.  I'd estimate that it included over 75 doctoral
dissertations.  (I kept a few, Kent Dybvig's for instance.)  Most of
that stuff is almost completely irrelevant.  There is an order of
magnitude more important, insightful, impactful thought embodied in
the most poorly-written 6-page hack-doc generated by the Mojo Nation
folks than there is in the average PhD.

> I'm not arguing against the University. I merely raised
> a question about how it structures the lower rungs of
> its career ladder.

There should be no binding whatsoever, as a reasonable compromise.
Myself, I try not to hire people with advanced degrees in roles where
they're expected to innovate or produce code.  The esteemed and
aforementioned Dr. Turpin stands at the head of the pack as one of
the few esteemed doctors I know who I actually trust to produce
code...  :-)  (That's a kudo, Russell, take it and run! :-)

jb