[Twisted-Python] April 20 Sprint

[James Y Knight]

On Apr 7, 2008, at 5:26 PM, Brian Granger wrote:
>> I myself having been using twisted.web2 for a while now,
>> and I also know of several other projects that are doing the same.
>> I also have done a good amount of diving into the source code,
>> trying to maybe someday understand enough to contribute back,
>> and I hope to keep up with the current status as best as possible.
>>
>> Any comments that point towards the current state of any of the parts
>> of web/web2 (what's "good", what's "bad", in the developers'  
>> opinion),
>> would be very appreciated.
>
> Wow, we have been using web2 in ipython and this it is news to us as
> well that web2 is being phased out.

I wish that people hadn't said that at this point, but since it's been  
said...

The problem, basically, is that I started the web2 and never finished  
it. At the current rate of development (really slow), there's a real  
fear that it'll never be finished. So, some (who actually are doing  
development on twisted...) feel that it would be better to take the  
more finished parts of web2 and re-integrate them into web, piece by  
piece.

I'm wary of this scheme, and I can't say I can see it being  
successful, but perhaps it's worth the shot.

My goal with web2 was to provide 99% compatibility with web1  
*resources*, but forgo compatibility at the other integration points.  
I still think this is a feasible goal to accomplish. The goal with the  
"phase-out" plan is to provide complete compatibility with all web1  
integration points *and* add the new features from web2 to web1. If  
this can be accomplished, great; I just can't say I see it being that  
feasible, and I certainly am having trouble thinking that it can get  
done *faster*.

But in any case, the take-home message of this alternate plan should  
not be "web2 is being phased out", but rather, "web2's improvements  
are being absorbed back into web." IMO, it remains to be seen whether  
or not that can actually happen, but I don't think current users have  
any serious reason for alarm at that prospect.

James