Re: in defense of types

[Matthias Felleisen <address@hidden>]

Yes, Jeff, I have heard this before. Program understanding 
does not come from types alone. As for types, the story is
wide open for powerful type systems (see ML). Of course, 
you can always make sure that your type system is so strong 
that only programs such as "print 5" pass the type checker, 
and it is easy to understand this program and to compile
it. Don't fall for such pubs. Question them. -- Matthias

P.S. The SEI link doesn't work. 

At Wed, 29 May 2002 06:31:03 -0600 (MDT), Jeffrey Palm wrote:
> Quoting Matthias Felleisen <matthias@ccs.neu.edu>:
> 
> > 
> > In a recent posting, which I promptly lost, Paul Graham, challenged
> > the
> > usefulness of static typing in programming languages. I'd like to take
> > him up on this aspect of PLs.
> > 
> > [Note: For those of you who know me from academic venues, 
> >  you will note that I "play Paul" there, but for very different 
> >  reasons. We need to challenge each other on such statements 
> >  because only strong challenges to scientific claims will ensure
> >  that these claims are 'hardened'. So, if you want to know what's
> > wrong
> >  with types and the influence of the type research community, go to
> > academic
> >  conferences.]
> > 
> > So here we go. Static types are necessary 
> >  (1) to establish basic properties of software
> > and
> >  (2) to establish absolute yet cheap abstraction boundaries between
> > modules. 
> > LL languages suffer from the lack of types, for this very purpose. 
> > 
> > ...
> >
> > -- Matthias
> 
> Static types can also lead to better program understanding:
> 
>  http://www-2.cs.cmu.edu/afs/cs.cmu.edu/user/roc/public/ICSE97-Final.ps
> 
> and optimization:
> 
>  ftp://ftp.diku.dk/diku/users/henglein/tagging-optimization.dvi.gz
>  http://www.cs.colorado.edu/~diwan/toplas.ps
> 
> Jeff
> 
> ---
> Jeffrey Palm --> http://www.cs.colorado.edu/~jdp