How do Monk skeletons compare to generators?
The idea of having a skeleton versus a code generator is that we wanted to evolve the structure of the skeleton separate from the binary, thus the structures can always be up to date without the need to update the monk binary. This is good also because you can have many different monk skeletons configured: say you have the default one with Ohm (our Redis layer), another one with ActiveRecord and yet another one with Sequel. You can be sure at any point that starting a project will give you the best combination, not what was best at the time you installed the monk gem. Having the binary generate all those different structures would be overkill, and would concentrate the power on the people that release the monk gem. This way, anyone can maintain their own skeleton with their stack of choice. Some benefits of having functioning skeletons (rather than a generator) are: • Integration: it’s easier to test that the base application works and that there are no conflicts in the dependencies. •
