Re: peasant revolt against DrScheme!

[Michael Vanier <mvanier@bbb.caltech.edu>]

> From: Greg Pettyjohn <GregP@bsquare.com>

> 
> I doubt that the young Schemer will be put to work (re-)designing whatever
> product
> from the ground up to be written in Scheme. My doubts are motivated by the
> very
> large population of C/C++ programmers who will be the Schemers' managers and
> who
> aren't due to retire for a few decades.
> 

In a side project I'm involved with, we are doing a lot of web
programming.  We're using python, because I predicted that we'd be able to
get people up to speed in it in a few days (and I was right).  The learning
curve for scheme is much more steep, although as Shriram says, that's
because most of us (including me, sad to say) still see programming through
C-colored glasses.

On that note, there's a fascinating old article by Guy Steele and Gerald
Sussman called "Lambda, the Ultimate Imperative" which shows how to model a
lot of imperative constructs in scheme.  It's available online at:

ftp://publications.ai.mit.edu/ai-publications/0-499/AIM-353.ps

I found it very enlightening.

> 
> So let me add to Mike's hypothetical quote: 
> 
> "you're thinking about programming in the wrong way; you have 
> to totally unlearn all of your ideas about programming and start over
>  *and* you have to re-write your system in Scheme because it's broken
> by virtue of having been written in C/C++."
> 
> Even if the established body of imperitive programmers *believed* this
> sentence,
> they'd still have a hard time owning up to it.
> 

And they wouldn't believe it unless scheme could compete with C/C++
speed-wise.  That's why I think the target should be scripting, where all
of scheme's strengths are manifest.

> So, in conclusion, I think that how to educate Caltech CS freshmen is only
> a small piece of the puzzle. I think that we are faced with is the more
> monumental
> task of producing a "paradigm shift" in the industry.
> 

Agreed.

Mike