In my opinion, a version number should not be assumed to be a float. Version numbers can also look like this: "12.3.0.3". This is where "float" fails severely.

My method is something like Python's "sys.version_info.major", where Python's version number contains several decimal points.

---