M. Edward (Ed) Borasky wrote:
 > Ah, but you can *buy* or download for free really good compilers for
> functional languages. See the conference announcement I posted a couple
> of days ago. Maybe you'd even like to come visit the City of Roses? :)

I'm awfully impressed by the recent quality of the compilers for 
lambda-languages, especially Lisp. This was far from true a dozen years 
ago when I was going through the exercise. Also, I was a professional 
compiler-writer at the time. I ditched that career about five minutes 
after I read the Dr. Dobbs article announcing Java, in May 1995. (Well, 
not completely. That summer I wrote a bytecode->native x86 compiler. But 
the handwriting was on the wall.)

Interesting, your comment about Rails. (Although now we're mixing CoC up 
with my intuitions about fundamentally restricting language power.) You 
notice how people talk about Rails as the "DSL of the web"?

Part of what I'm interested in is permitting arbitrary Ruby code to run 
in IoC containers with DI. That does imply some restrictions on how you 
write your classes, not least so the framework can automatically give 
you diagnostics and management tools. The simple fact that people use 
TDD effectively in the agile world shows that there is something to 
this: you're writing code-objects that are designed to be testable. 
Well, I'd like to discover what are the best practices for writing 
code-objects that designed to be distributable and composable. (Of 
course we already know about an approach that *won't* work- 
convention-based reflection in the manner of Java Beans.)

-- 
Posted via http://www.ruby-forum.com/.