Just like people dream of rock star programmers, instead of focusing on the miracle projects that people are supposed to produce in this 20% "mess-around" time (which is not, by definition, messing around if you expect them to actually produce something reliably) maybe one should focus on the value added to the individuals.
Messing around with a new framework or some prototype in your current language of choice means that you won't be (or won't have to) trying some new stuff on a production-class project. As an individual it is a great opportunity to hone your skills and knowledge, and help find the fun again in what could otherwise be a bland job.
This is actually iteration zero of a project, dedicated not to hopeful production of worthwhile deliverables, but production of the minds that will make (other) projects happen. Because when you will start working on iteration one of a new project, a new feature, then it is too late to mess around.
It just happens that sometimes great things could come out, but they're a statistical oddity.
Messing around with a new framework or some prototype in your current language of choice means that you won't be (or won't have to) trying some new stuff on a production-class project. As an individual it is a great opportunity to hone your skills and knowledge, and help find the fun again in what could otherwise be a bland job.
This is actually iteration zero of a project, dedicated not to hopeful production of worthwhile deliverables, but production of the minds that will make (other) projects happen. Because when you will start working on iteration one of a new project, a new feature, then it is too late to mess around.
It just happens that sometimes great things could come out, but they're a statistical oddity.