[r6rs-discuss] Thoughts on Scheme's Future

Joe Marshall jmarshall at alum.mit.edu
Fri Sep 4 16:30:24 EDT 2009


On Fri, Sep 4, 2009 at 1:21 PM, Anton van
Straaten<anton at appsolutions.com> wrote:
> Joe Marshall wrote:
>> Similarly with threads.  These are a disaster.  It is very hard to program with
>> concurrency.  It is harder still if your concurrancy mechanism is something
>> as primitive as a thread.  In nearly every implementation of Common Lisp
>> I've worked on, they've made major errors in concurrency control.  And these
>> are the *vendors*.  The users have no hope of getting it right.
>>
>> I don't know what the correct solution to concurrency is, but I'm sure we
>> can do better than a thread library.
>
> I don't know what the correct solution to control flow is, but I'm sure
> we can do better than first-class continuations.

Agreed.  Although this could be interpreted in two different ways, both of which
I agree with.

    1.  Easier to understand user constructs such as structured
exceptions, non-local
         exits, co-operative co-routines, etc. should be what users turn to
         rather than ad-hoc grabbing of continuations.

    2.  First-class continuations should NOT be used as a model or
         implementation of threads.  (You can crudely mimic time-division
         multiplexing, but not true asynchronous parallel computation.)

-- 
~jrm



More information about the r6rs-discuss mailing list