So, a basic philosophy that Catalyst adopted from Perl itself is the famous There is more than one way to do it. Dan points out the fallacy of having to make too many choices when adopting a new framework. I disagree that that is a deciding factor in Rails or django having more web applications listed on appliedstack than Catalyst.
I believe the extreme Rails hyping and number of books coming out to be a critical Rails success factor, and Google choosing Django as it's only language for App Engine to be a critical Django one. Having said that, I do believe that having to make all of these choices up front, while learning a new framework, has been harmful to Catalyst adoption. I don't think we need to give up flexibility to fix that problem tho.
We just need to provide good defaults. I shouldn't have to deal with choosing a template system until I've used TT for a while and decided that it sucks. I shouldn't have to decide what the best way for me to talk the database unless I have some legacy model I need to integrate with. I specially don't need to configure a wealth of Authentication modules unless I'm doing something zany.
Ranting about this is easy, but I'd like to do something more, so I've created a IRC channel, #catalyst-stacked on irc.perl.org, if you are interested in helping to build a full stack framework on top of Catalyst, come by there and join in on the discussion.