[plt-scheme] Re: Change the World

On Mar 21, 2005, at 09:33, Matthias Felleisen wrote:
> On Mar 21, 2005, at 8:25 AM, Neil W. Van Dyke wrote:
>> (I blame MIT for letting undergrads use Java.)
> And it's going to get worse.

Describe the class of problems that are mostly likely best handled by 
Scheme / Lisp, and there will always be a place for Schemers and Lisp 
hackers.  Then there's the Haskell camp backed by Microsoft.  They 
probably don't fund it unless they see a pot of gold in thar' hills.

I'm trying to guess what makes it hard for Scheme to "catch on" in the 
mainstream.  I don't think it's parentheses.  I think it's that bad 
Scheme blows up in the developer's face more dramatically than bad 
Java.  I think bad Java programs limp along half working and half not 
working until someone fixes the parts that don't work, whereas bad 
Scheme programs may not work at all until someone figures out where the 
problem is.  A Java klutz will break a portion of a Java program, but 
it will still klunk along.  A Scheme klutz might easily cause more 
mysterious death.  These are my past impressions.  DrScheme, modules, 
SchemeUnit, PLaNet, &c. are paving the way to a safer future.

A well written Scheme program is a work of art.  A good Java program is 
rarely so graceful, just tidy.

Speaking of art, I wonder why Knuth has stayed away from Lisp.  I also 
wonder how much of TAOCP could be expressed in Scheme.

Geoffrey
--
Geoffrey S. Knauth | http://knauth.org/gsk