[Twisted-Python] win32 reactors

Phil Christensen phil at bubblehouse.org
Wed Apr 2 19:15:37 EDT 2008

On Apr 2, 2008, at 6:24 PM, James Mansion wrote:
> Cory Dodt wrote:
>> Resolving a bug includes gathering requirements and building
>> consensus, but building
>> consensus goes much faster if there's an implementation handy to
>> discuss.  Even a
>> quick hack is useful as a discussion point.  A very common  
>> scenario is
>> that a quick
>> hack is eventually refined into a unit tested, UQDS-vetted
>> implementation.  However,
>> a hand-waving discussion never is.
> Call me old-fashioned, but what you are describing is the  
> difference between design-free-hacking
> followed by iteration, and actually designing something.  You know,  
> all that waterfall stuff. I know,
> its not fashionable right now.
> Seriously, though, its too late at the review-of-nearly-working- 
> code. There's too much
> pressure to incrementally fix it, and at least one participant will  
> have a sense of ownership
> in something that might be a country mile from the best solution.
> I know its easier to design when you can meet each other and use a  
> white board or
> just scribble on paper, but its still entirely possible to use words.
> So - I disagree.  I'm quite happy to hand-wave, and to listen to my  
> colleagues'
> hand-waving. If it communicates design ideas - and requirements -  
> before any wasteful
> coding, that's good. Honest.
>> Still, things get fixed when someone fixes them.  It falls on the
>> person who needs them
>> fixed to do so, no matter whether you're talking about software or  
>> rain gutters.
> Hmm.  You sure it doesn't happen after the prioritisation meeting  
> and we all get our
> steer?

Perhaps I'm misinterpreting your meaning, but I think your attitude  
is way out of line.

Free software is possibly the only scenario where you truly get more  
than you pay for, but that doesn't mean it comes without a cost.

One cost is that the project and its community dictate their own  
development process, and one person has little or no ability to  
change that inertia. It is the right, possibly even the purpose of  
the community to grow however it will.

I get just as frustrated as the next guy when I need something fixed  
RFN, but that's the price you pay in exchange for a huge library of  
high-quality code. For every time I've run into an issue with Twisted  
that the core devs were too busy to fix, there's 50 other times when  
Twisted has saved my ass.

Wasteful coding? Who do you think you're talking to here?


More information about the Twisted-Python mailing list