[plt-dev] IDE and completion

If I could get it for a (typed) BSL/ISL language, I'd be happy. It's not that bad there and yes, I am willing to overlook the HO character of the language. 



On May 4, 2010, at 4:35 PM, Joe Marshall wrote:

> On Tue, May 4, 2010 at 11:49 AM, Matthias Felleisen
> <matthias at ccs.neu.edu> wrote:
>> 
>> 
>> 1. syntax: OOP guys write down the first argument first (this) and then the method call and that is the way syntax works.
> 
> Yes, but they give up the ability to have first-class methods.  The
> method name is selected
> from the scope of the object, not from the lexical environment.  There
> is no way to write
> 
> applyToAutomobile (Method action) {
>    return myAutomobile.action();
> }
> 
> (of course you could use reflection, heh heh)
> 
> In addition, there's a level of kludginess that they put up with in
> order to bootstrap
> themselves into the object world. (No first-class constructor methods,
> so `factories'
> have to be introduced.)  In FP, given an arbitrary variable like
> `aList', there are a million
> constructor functions that could operate on it:  cons, list, make-foo,
> alist->hash-table,
> etc.  If first-class functions were eliminated from functional
> programming, it would be
> pretty easy to auto-complete, too.
> 
>> 
>> Is it really hopeless for us? -- Matthias
> 
> Yeah.  We're saddled with a language that permits us to name nearly anything.
> Perhaps we need a less powerful paradigm.
> 
> -- 
> ~jrm