Maybe the best approach is to do as Ruby on Rails did -- do something real, and then create your tool/infrastructure around/in reaction to that (in Rails' case,

Maybe the best approach is to do as Ruby on Rails did -- do something real, and then create your tool/infrastructure around/in reaction to that (in Rails' case, it was Basecamp and subsequent applications, as well as contributions by other working practitioners -- not stuffy people sitting on committees writing specs instead of applications that people use). So, in order to make a CMS, you ought to have some content to manage. Make your CMS really, really good at managing that particular collection of bits you care about so dearly without making things complicated. That seems like a good strategy to me. Yes? No? To answer your question directly, if you think you can create some useful software that scratches your itch in a novel way (slightly novel is more than sufficient), by all means work on it, and post a follow up here so we can try it out! :)