[PRL] [plt-internal] Video of Clojure talk

[Shriram Krishnamurthi]

Yadda yadda.  (That's to him, not to you.)  Yes, I did hear him say
that 2-3 times.  And I don't believe it'll make any difference to the
JVM at this point.  Unless someone has internal information that
suggests the JVM does plan to fix this problem.

Shriram

On Thu, Oct 2, 2008 at 3:00 PM, Dave Herman <dherman at ccs.neu.edu> wrote:
>> No, he gets an audience question about tail calls, recognizes that
>> they are useful for setting up networks of calls between functions,
>> and then says something that I don't understand -- it sounds both like
>> you can do that with his constructs, AND that it will blow up out of
>> memory.  (You don't need special constructs to achieve the latter!)
>
> I think you missed an earlier part of the talk that he quickly refers to in
> that comment. To expand out his point: the facility exists in Clojure now,
> but it uses up unbounded stack space and therefore isn't usable -- *yet*.
> But he mentioned earlier that at the JVM language summit he just went to,
> one language designer after another hammered on the desire for proper tail
> calls in the JVM. Based on the mood at the summit, he's optimistic that the
> JVM will eventually be properly tail-calling, at which point mutually
> recursive, tail-calling functions in Clojure will use constant stack space.
>
> Dave
>