[PRL] Computer science `education'

[Eli Barzilay]

On Nov  9, Carl Eastlund wrote:
> On Nov 9, 2007 3:00 PM, Joe Marshall <jmarshall at alum.mit.edu> wrote:
> > Was I expecting too much?  We're all my questions way too hard?
> > Or did the college completely fail in its mission?  Or is the
> > general computer science curriculum this bad everywhere?
> 
> My brother works for Google and has told me his own horror stories
> from interviewing job candidates.  Your guy is way above the
> "horror" bar compared to what I've heard.  And that's compared to
> people interviewing for fully technical positions.  If this guy was
> only going to be semi-technical, he sounds like a good candidate to
> me.  A technical manager who understands that there are benefits to
> functional languages?  Even if he couldn't list a single one, that's
> a step in the right direction compared to the industry standard.

A generic "easier to parallelize and scale up" is IMO not much of a
step in any direction.  It sounds like trying to come up with whatever
template reason he's been fed, just to satisfy Joe.  In other words, I
won't be surprised to find this person arguing against using a FPL
later on based on his "knowledge".

I see this all the time in my class -- students find little hooks that
they can hold on to, and "easy to parallelize" sounds just like that.
Getting the impression that you know something based on such a hook is
more dangerous than not knowing anything.

As for the original list of questions, I think that his major flop is
with the entropy question.  I know that I'd go on forever about how
cool it is to have a measure of "information density" -- and that's
without the knowledge to tie it to chemistry/physics, an without being
interested in information theory...

-- 
          ((lambda (x) (x x)) (lambda (x) (x x)))          Eli Barzilay:
                  http://www.barzilay.org/                 Maze is Life!