pecan create test_project
Unless this command is vastly more opinionated (and less bare-bones) than Django's startapp/startproject, and with only what you've written to go on, I get the impression that what you're doing here is killing part of the benefit a framework gives you: The ability for others with framework experience to (relatively) quickly pick up your projects and work on them.
I've worked on several projects that did similar things in addition to many that more strictly followed their framework's conventions, and choice of framework aside, when you're using something lightweight already but ignoring its conventions, you might as well be rolling your own.
If that's not something that's important to you when choosing a framework, for whatever reason, then I can see this being OK (if not useful). But I don't think it's a good habit to get into if you are likely to pass off your work to others or need to bring in outside help in the future.
Padrino uses generators[1] for this and while they might have too many options, something like this for Flask would be fantastic, IMO.
Maybe I'm missing something, but I don't see a substantial improvement over Flask's (or even Werkzeug's) approach.
I personally prefer Flask's simplicity of "do as you will" over "here's an entire project for you" that lots of frameworks use.
It diesn't seem as clean on a first glance?
I know that everything is a tradeoff, bottle.py has A LOT of limitation, but to me it is the lean'est python framework. Heavy framework (WAY more features) would be django and pyramid, and midle of the road cherrypy. I'm not sure where Pecan fits in that picture.