The BDFL's boundary

I think my response to the recent news that Guido had "pronounced" Django as the Python Web framework was "so wot's 'e think 'e's doing, anyway?"

First of all, I don't think it's a big deal. Every few months something happens to get the Python world all abuzz about the number of Python Web frameworks. It might be the announcement of a new one (Django, TurboGears, web.py, Clever Harold, etc.). It may be some bit of Ruby on Rails news, which for some reason seems to strike fear into the depths of Pythoneers' consciousness. (I think O'Reilly amuses themselves by publishing book sales figures just to see how high they can get some in the Python community to jump). Whatever the stimulus, people blog back and forth about it, a handful of folks switch frameworks, just to be safe, but in the end it all dies down with the status quo still pretty much as it is. When Python Web frameworks die (which seems to be a rare event), they die with a whimper, and not a bang.

Some might think that a BDFL "pronouncement" is no ordinary stimulus, but I'd bet good money that this case is indeed quite ordinary. You see, that's where we have these lovely things called boundaries and it seems one has unmistakably been exceeded in this case. When the BDFL pronounces on a matter, it has always been with the intention, and effect, of settling a troubling argument once and for all. That's not possible in this case. It's not as with Python decorators where Guido's decision to include them, and his choice of syntax, became part of the language proper, and the only thing you could do if you disagreed was take the enormous step of forking the language. In this case, if you disagree with Guido, what do you do? You ignore him and keep using (or developing) your favorite framework. Nothing in the evolution of Python impedes you the slightest bit. In the end, I think this is what each individual developer will end up doing. A few will indeed give Django a new look, and some will convert, but in the main, all will be back to normal in a few weeks.

For my part every sniff I've had of Django makes me think it's way too large and monolithic for my taste. In other words, it's far too much like Zope. (Speaking of Zope, if it were ever appropriate to declare one Python framework to rule them all, Zope is the one with the popularity and maturity for such a role—which is why I'm glad such a declaration is not appropriate). I understand they're getting rid of a lot of the magic, which is another thing that gave me flashbacks of Zope, but I doubt that would be nearly enough for me. For now, I'll continue to work with my favorite three frameworks: CherryPy, RhubarbTart and Pylons, although I do carry some hope that the latter two will merge, and leave me with just two favorites. I'll also focus as much Web development work as I can in creating WSGI middleware, which provides the greatest flexibility.

See also:

[Uche Ogbuji]

via Copia