stackless, uthreads, and GUIs

[Fernando Pereira]

In article <slrn8dgdru.dk5.neelk at brick.cswv.com>, Neel Krishnaswami
<neelk at brick.cswv.com> wrote:

> Johann Hibschman <johann at physics.berkeley.edu> wrote:
> > 
> > I was wondering if anyone has experimented with using stackless to
> > implement guis.  I haven't played with the package much myself (yet),
> > but it seems like you could write something nice using continuations
> > to implement dialog boxes, and so on.
> > 
> > I know most real GUIs use threads, but real OS threads are such a PITA
> > to get to compile (in my experience), that I'd like other options.
> > 
> > I'm mostly just playing with this idea for now.  Is it a reasonable
> > direction in which to be thinking?  Any accessible references?
> 
> There is a paper describing precisely this idea called "Escaping the
> Tyranny of the Event Loop: an Alternative Control Structure for
> Multi-threaded GUIs," by Matthew Fuchs. You can find it at
> 
>   http://www.cs.nyu.edu/phd_students/fuchs/
> 
> You may also be interested in reading about Fudgets, a purely
> functional GUI toolkit for Haskell. This one is a bit mind-blowing,
> though:
> 
>   http://www.cs.chalmers.se/ComputingScience/Research/Functional/Fudgets/
> 
A related, earlier, and simpler idea, based on ideas from CSP and
Esterel is Cardelli & Pike's "Squeak: a language for communicating with
mice", Proc. Twelfth ACM Annual Conference on Computer Graphics and
Interactive Techniques (SIGGRAPH), 1985,
<http://research.microsoft.com/Users/luca/Papers/Squeak.pdf>. It might
be a cool exercise to create a uthreads implementation of these ideas.

-- F