I started developing web applications in Python relatively a long time ago.
I tried almost them all in various degrees of depth: Python + CGI, Zope 2, Webware for Python, Subway (who does remember it?), Twisted Web, Nevow, Python + WSGI, Paste, Django, Pylons and maybe something else I can’t remember right now.
Nowadays most of my attention is caught by Pylons and Django. The latter has recently come to its 1.0 milestone and altough some books about it out there need update there’s a ton of documentation to start on the right track. Pylons has a different approach (but not that different in my opinion, because Django is not tight coupled as some people think it is) on how to develop applications but I’m not here to judge which one does things in the best way although I have to admit that for personal preference I’d vote for Django.
I’ve been using Pylons at work (the web part of one of our flagship product is going to be Pylons in and out) for a year and I used Django for personal and professional applications and both of them have pro and cons.
What really concerns me is the different feeling the everyday use of both give me. When I’m focused on Django I feel part of an increasing and participating community with tons of blog posts, job offerings, and niceties such as the amazing Pinax project, django people and so on. It does also feel like I’m using something rock solid despite its limitations (every piece of software has them).
Pylons, au contraire, is just a nice piece of software with which you can develop your applications but I don’t have around me the same community, the same vibe.
I hope the upcoming Pylons book is going to change that. Don’t underestimate the power of communities, there’s not only Facebook out there 
4 Comments
You might also want to take a look at web2py and Glashammer. Both are also very interesting frameworks.
Arthur: I am sure of that, despite the fact that I think fragmentation isn’t always a good thing
Yep, there’s definitely differences in the communities, I think mainly as they both hit different sweet spots. There’s going to be a lot more in common for the Django community by the nature of Django itself. Just like there’s a strong Plone community, strong Joomla community, strong Drupal community, etc. You have a large group of people making sites that generally are fairly close in core functionality.
I’d consider the Pylons community fairly strong as well, but definitely a different feel to it as the range of problems people are using it for seems a bit more varied. With Django, 98% of those using it will be using functions it provides such as its ORM, its templates, contrib modules it comes with (which all come with a set of assumptions about how the site works), etc. These things mean that most Django (certainly not all though) will be very similar in many ways.
Pylons projects can deviate heavily from one another, which means you can’t have all this re-use between members in the community, as most of them make fairly different assumptions. But some core aspects are definitely picking up steam, for example FormAlchemy has come a long ways and makes it a snap to generate forms from SQLAlchemy models.
They’ll always have different communities, as they address different ways of solving the problems at hand which generally appeals to different groups of developers.
Exactly Ben, I understand the reason why Pylons exist obviously, the problem that I see in the long run is about the fragmentation. Turbogears had the same problem too. Stale documentation or 10 ways to do a tutorial because everybody used a different set of packages. Pylons is more linear in that way, fortunately.
If the potential contributors are divided amongst different subprojects there’s less concentration of people around the same problem. Just my consideration