One of the most interesting points he makes in this talk is about separating out what Eric Evans might refer to as supporting generics (AFAIK) from the applicat

One of the most interesting points he makes in this talk is about separating out what Eric Evans might refer to as supporting generics (AFAIK) from the applications to write into separate libraries and open-sourcing them. To me that's almost certainly the way for truth and good but fails to acknowledge that good libraries take a _really_ long time to write; ages. And then you have all this issue over being unable to change published interfaces because people rely on them now and dependency management etc. We need better programming languages before I'd be comfortable advocating that advice.