I have not reached that point yet but as your application grows it might be better to move some code and with that some route definition to separate files. For example if I have a Dancer application called Foo I might want to move all the user related methods get '/user/this' => sub { ... }; get '/user/that' => sub { ... }; get '/user/other' => sub { ... }; to a package called Foo::User; I can then include use Foo::User; in the main Foo package and the above routes are added. Within Foo::User the /user prefix seems to be too repetitive so it might be better to have just get '/this' => sub { ... }; get '/that' => sub { ... }; get '/other' => sub { ... }; and mount that package to '/user' within the Foo application. (which by itself might have been mounted to some path. If I understood it correctly from http://advent.perldancer.org/2010/2 and from the recent presentation of Sawyer, this would be getting similar to what MVC frameworks do and what you referred to as "Namespace matching". As I was playing with this idea I implemented a small example I call Dancer::Plugin::MVC for now that would do this automatically. You just include use Dancer::Plugin::MVC; in the main Foo package and during its import() call it will load all the modules in the subdirectories of Foo. So far it simply loads the modules it finds which is quite simple. I'll try to check if I can mount the various modules in other places as well though in that case I might not want to do it for all the modules and the mount point should be the same as the package name. What do you think? Is this totally contradicting the philosophy of Dancer? Gabor