[Stackless] Re: [Zope-dev] Zope 2.7 running with Stackless 3.0

Bob Ippolito bob at redivi.com
Thu Sep 4 03:52:01 CEST 2003


Well, it helps to start somewhere.  I'd be happy if the 
promises/futures (or some other official way to do things async) made 
it into Python, like the way generators made it in instead of all of 
stackless.  I'm perfectly fine with the reactor pattern as-is, but I'd 
very much like nicer syntax than what Twisted has to offer (which is 
about as simple as it can be, as far as I can tell, without changing 
Python).  Perhaps it's possible to build something that can allow 
"indeterminate direction of information" in an efficient way built on 
top of promises/futures.

Personally, it would make my day if Python could figure out what code 
is "functional" enough such that it doesn't have to bother executing it 
right away.  For example, IMHO, most list comprehensions should be 
generators, but maybe converted to a list (or cached in one) on-demand 
if it's used like a list.  I also think it would be great if generators 
could implement __len__ if you can calculate exactly how many results 
it will have in advance.

I haven't been following this list for very long, so I missed the 
"logic variables" discussion.  I read some archives about it, but I 
can't find the notorious PEP.

-bob

On Wednesday, Sep 3, 2003, at 19:08 America/New_York, Jim Bowery wrote:

> Well, actually, promises/futures are exactly what I think inadequate.  
> I should have said "indeterminate direction of information" probably.  
> The question of whether something is set or read by a process is 
> impossible to establish a priori in a lot of situations where demand 
> driven computation is emergent.  This was covered a while ago in the 
> "logic variables" discussion.  The Mozart-Oz guys have given up on the 
> hard problem and dismissed it with promises/futures apparently.
>  
> Oh well...
>
> ----- Original Message -----
> From: Bob Ippolito
> To: Jim Bowery
> Cc: stackless at tismer.com ; Shane Hathaway ; Aaron Watters
> Sent: Wednesday, September 03, 2003 2:19 PM
> Subject: Re: [Stackless] Re: [Zope-dev] Zope 2.7 running with 
> Stackless 3.0
>
>
> On Wednesday, Sep 3, 2003, at 16:12 America/New_York, Jim Bowery wrote:
>
> Someone who thinks there should be more allowance for indeterminate 
> order of evaluation in highly threaded applications.
>
>
> Twisted ( http://www.twistedmatrix.com/ - website might still be 
> moving though ) does this to an extent with Deferreds (aka 
> promises/futures, I think)...
>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: text/enriched
Size: 2887 bytes
Desc: not available
URL: <http://www.stackless.com/pipermail/stackless/attachments/20030903/1506d9a8/attachment.bin>


More information about the Stackless mailing list