[Edu-sig] Lowering the Barriers to Programming

[Scott David Daniels]

Kelleher and Pausch's ACM Computing Surveys 37(2) paper.

Well, I took my time getting through the paper, and like the rest of you
here was left dissatisfied.  The first problem I have with the paper is
is that, while there is a large body of work that it covers, I am
unable to discern a desiderata for whether any particular work will be
included.  I certainly see Python as one (very successful) attempt to
address the "Lowering the Barriers."  Not seeing any criteria for
inclusion or exclusion leaves me feeling that this paper is about "a
bunch of stuff I read."

Clearly there has been a lot of effort here in analyzing the subject
systems; simply thoroughly reading the system descriptions would be
exhausting.  The paper though, leaves me with the impression, "I read
a bunch of stuff, and this is kinda-sorta how I see the stuff I read
can be classified."

A survey like this should either start with a taxonomy and show how
efforts fall into this taxonomy.  Such a paper is about the taxonomy,
and should concentrate on how well the taxonomy works.  On the other
hand, the paper could describe a way of accumulating research, and
then produce a taxonomy from observation on the accumulation.  Neither
seems to be the case here.

A kvetch: the SP/k claims PL/1 evaluates 25 + 1/3 as 5.33333.  How could
this be true?  When stating a possibly surprising fact, proofreading is
indicated.

Why is COBOL in there? If it is, FORTRAN and ALGOL certainly belong, and
wherever those three belong is where Python belongs.  I suspect that
Turing is in this group, and I don't know that Turing was a "stripped
down for teaching" language.  Claiming that BASIC's "LET" statement
is somehow simplifying the language for the student does not convince
me; I think LET simplifies the interpreter, not the student's task,


As to the chosen hierarchy, the top-level distinction confuses me:
       Systems to teach programming for its own sake
   vs. Systems to teach programming in pursuit of another goal
To which class do systems to teach programming in order to teach
Computer Science belong?  Determining the "primary aspect of
programming that the system attempts to simplify" seems equally
troublesome, requiring a crystal ball -- I have no confidence in
reading this paper that another person would cut the boundaries
the same way.

--Scott David Daniels
Scott.Daniels at Acm.Org