[Web-SIG] PEP 444 != WSGI 2.0

Guido van Rossum guido at python.org
Mon Jan 3 04:21:53 CET 2011


On Sun, Jan 2, 2011 at 7:08 PM, Tres Seaver <tseaver at palladion.com> wrote:
> On 01/02/2011 04:31 PM, Guido van Rossum wrote:
>> On Sun, Jan 2, 2011 at 12:55 PM, Masklinn <masklinn at masklinn.net> wrote:
>>
>>> On 2011-01-02, at 21:38 , Alice Bevan–McGregor wrote:
>>>> On 2011-01-02 11:14:00 -0800, Chris McDonough said:
>>>>>> I'd suggest we just embrace it, adding minor tweaks as necessary, until
>>> we reach some sort of technical impasse it doesn't address.
>>>> Async is one area that 3333 does not cover, and that by not having a
>>> standard which incorporates async means competing, incompatible solutions
>>> have been created.
>>>>
>>> If I remember the previous Web3 discussion correctly, the result was
>>> basically that async has no business being shoehorned in WSGI, that WSGI's
>>> model is fundamentally unfit for async and you can't correctly support sync
>>> and async with the same spec, and therefore an asynchronous equivalent to
>>> WSGI should be developed separately, in order to correctly match the needs
>>> of asynchronous servers and interfaces, without the cruft and inadequacies
>>> of being forked from a synchronous gateway model.

>> Masklinn, those are pretty strong words (bordering on offensive). I'm sure
>> Alice has a different opinion. Alice, hopefully you can write down your
>> ideas for all to see? Perhaps you have an implementation too? Maybe seeing a
>> concrete proposal will help us all see how big or small of a shoehorn will
>> be needed.
>
> I didn't read that as offensive, but as an accurate description of the
> de-facto consensus of the SIG:  there is a *very* small minority who
> want next-WSGI to address async issues (AFAIK, Alice and Manlio are the
> only active proponents).

That's not what Masklinn appeared to be saying though. He implied that
it was a bad idea, not that nobody was interested. Also he used an
excessive number of negative terms ("has no business", "shoehorn",
"fundamentally unfit", "cruft"). If he had said it the way you said it
I would have remained silent. But I accept that I overreacted, and no
offense was taken by Alice, so this is all academic. The best way
forward for async is likely a separate PEP.

-- 
--Guido van Rossum (python.org/~guido)


More information about the Web-SIG mailing list