[PREV - ALGORITHM]    [TOP]

COMP_SIGH


                                             March 31, 2023
                                             April 05, 2023
                                             April 08, 2023



                                                  I was looking up stuff
                                                  by the authors of
Mehran Sahami did some work on revising
Stanford's "Computer Science"                            SYSTEM_ERROR
curriculum, and as you might expect,
he's a booster for Computer Science
academic programs.  This talk on the
subject is less than impressive...         https://www.youtube.com/watch?v=9UuqkXK1xy0

  "The World Needs More Wizards and Witches" (2012)
  by Mehram Sahami, at TedX Gunn High School            With "TedX", the X
                                                        somehow seems to stand
                                                        for "mediocrity".

                                                        There are lessons for
                                                        us all here, though,
                                                        primarily "don't
                                                        dilute your brand".

He presents some charts labeled things like
"Understanding Supply Versus Demand", where
he's looking at numbers of graduates in a field
and numbers of jobs in the field.  He regards
it as a Bad Thing there are people doing computer
jobs that don't bother with CS degrees:

    "The number of jobs projected to
    exist is three times the number
    of graduates."                             (You have to wonder about
                                               the data, too: What's a
    "The majority of people who are            Mathematics job, what's a
    doing jobs in computing were not           Physics job?  How do you
    trained in computing.  And we              do the classification?)
    wonder why our credit card
    numbers get stolen when we make
    online transactions."


        This presumption that Computer Science geeks
        are more competent programmers and sysadmins
        than people from other fields-- and that it's
        only those *outsiders* who are creating             It would be
        security leaks-- is not actually established        interesting if you
        by any data he presents.                            *could* establish
                                                            it-- do security
           There's a style of argument that                 mistakes correlate
           *starts* with data, but does                     with academic
           without it at key points along                   backgrounds?
           the way, hoping you won't notice.



    The interesting thing to me here is
    that this data shows that few
    employers really care if you've got       Jumping to another field:
    a degree that says "Computer              anecdotally I can tell you
    Science" on it.  People in the            that physicists working
    industry do not automatically accord      as engineers had a poor
    CS geeks with respect, and there          reputation among the other
    might be reasons for this.                engineers at the nuclear
                                              facility I was working at
                                              in the early 1980s.
    Training doesn't necessarily
    improve performance, it                   While I was there, there were
    depends on whether the                    indeed some serious mistakes
    training is relevant for                  made by a guy with a physics
    what you need to do.                      background, that I'd attribute
                                              to a kind of arrogance-- physics
                                              people have the deep deep
      If one of the main things you           conviction that they're so smart
      learn is to take pride in your          they can just wing it, and do
      background, that can actually be        just as well as experienced
      counterproductive: engineering          engineers.
      requires a kind of humility-- or
      you might say "paranoia": a                E.g. I heard an electrical
      recognition of how easy it is to           engineer complaining about
      screw things up.                           them building a system with
                                                 twisted pair cabling rather
         (That's what "murphy's law"             than coaxial cables: "Noise
         tries to remind you about.)             immunity isn't something you
                                                 build in later, you know."

     Computer Science as a discipline
     is united in it's convinction               To be fair, I've also heard
     that they know what you need to             of some mistakes made by
     know, and they feel no need to              people with engineering
     prove that they're correct about            backgrounds-- e.g. getting
     this.                                       an order wrong for custom
                                                 bolts made of somewhat
                                                 exotic materials, wasting
                                                 quite a bit of money.
                                                 That kind of thing doesn't
                                                 take arrogance, just weird
                                                 mental glitches that we're
                                                 all prone to.



    Sahami suggests that the reason the CS field
    became less popular was the reputation of
    the field, where everyone is supposed to be       This sneer reminds of
    pulling all-nighters, staying up on Diet          the old hackers
    Coke and Doritos.                                 vs. designers split...


    He then seques into talking about his new CS
    curriculum at Stanford, talking about how it places
    emphasis on getting across The Power of Computing,
    because programming is just a means to an end.

       (How knowing this is supposed to
       keep you off the Diet Coke escapes me.)


    But anyway, "computing is something for everyone";
    he argues for a "big tent" of Computer Science;
    We need to understand "what computing is about"...


He then points to a doubling of the numbers of new CS majors,
and tries to act like his new curriculum deserves the credit
for this-- it's hard to see how this can really be:
the phenomena is more widespread than just at Stanford.

(As he comments: "this is now becoming a national phenomena.")

Again, the data gets sketchy later in the story...

    If you were interested in this subject, wouldn't you try
    to get your students to tell you what got them
    interested in the degree?  Give the new recruits some
    surverys, ask if they're aware of the new curriculum and
    how they heard about it...

Also, this talk was in 2012, three years after the new
curriculum... how are the trends fairing now, ten years later?




--------
[NEXT - DIVISION]