Subject: [Fwd: Help with type advocacy]
From: Craig Chambers (chambers@cs.washington.edu)
Date: Fri Oct 05 2001 - 18:05:07 PDT
FYI. Matthias vs. Bob!
-- Craig
Matthias Felleisen wrote:
>
> [----- The Types Forum, http://www.cis.upenn.edu/~bcpierce/types -----]
>
> Ken,
>
> while I like Bob's answer a lot, let me urge you to make your presentation
> a lot more concrete. Many undergraduates who enter PhD programs just don't
> get the abstract talk that many of us speak. They have never heard it, and
> they may have been told at their undergraduate institution that C++ or Java
> is all they will ever have to know to build a system. (That is often just
> implied because N - 2 courses are taught in that language of fashion.)
>
> Assume they know C++. Show them small snipets of code that compile just
> fine. Then ask them why programs that link in these snipets crash, hang,
> memory dupm, print "the answer to 4+5 is 'foo'", etc.
>
> Assume they have heard of Pascal. In that language, you couldn't do many of
> these things. Then show a tomb stone:
>
> RIP: Pascal, restricted to death by its type system.
>
> Explain. Mention that Java's type system suffers from lots of restrictions,
> too.
>
> Then tell them that ML and Haskell exist, and that their type systems are
> sound.
>
> Now challenge them:
>
> (1) how can you learn to use an existing language in a safe manner?
>
> from pl people! perhaps some students who "know" that all languages
> are the same may sit in your courses and try to learn how to improve
> their coding
>
> (2) how can we (as a community) learn to build better type systems:
>
> build real, interesting systems in languages where we can tinker with
> the type system.
>
> Be concrete. Abstraction is what they must learn.
>
> Good luck. -- Matthias
_______________________________________________
Cecil mailing list
Cecil@cs.washington.edu
http://majordomo.cs.washington.edu/mailman/listinfo/cecil
This archive was generated by hypermail 2b25 : Fri Oct 05 2001 - 18:06:03 PDT