Re: [stack] disallowing recursive definitions

On Mon, Mar 3, 2008 at 10:12 AM, John Cowan wrote:

> Joe Bowbeer scripsit:
>
> > As far as I know, Dijkstra didn't have a beef with recursion.
>
> It seems to me from _A Discipline of Programming_ that he thought
> recursion a special case, relatively unimportant: the programming language
> he introduces there does not admit it.  And he is very much on record as
> condemning the reduction of iteration to (tail) recursion as a false
> reduction, _ignotum per ignotius_.
>

But he mentions recursion favorably in the letter you were using to indict
it.  I'm confused.



> > Recursion is the simple, elegant solution.
>
> Recursion can be employed elegantly, when it mirrors induction, or
> inelegantly, when it is an invitation to make a mess of your program.
> Not all syntactic sugar is oncogenetic: there are good reasons to use
> both let and lambda instead of just lambda.
>

By "mess" you mean obfuscated and unintelligible?  I agree.  Recursion is an
elegant concept that can lead to terse, confusing code.  But so, in my
opinion, can concatenative programming, so I have trouble understanding your
complaint.  If you are saying that adding recursion to concatenative
programming is beyond the pale, I think I could agree with that.

--Joe


[Non-text portions of this message have been removed]