i think open-sourcing widgets was a good idea, for those who can fix their own bugs, but as a case-study of community support i’d agree that it falls a bit short.
my impression is that the “o-s community” expects a certain minimum “structure” before they’ll take it upon themselves to improve/upgrade.  so, while widgets has some basic organization in place, it also has lots of weird “gotchas” that (probably?) prevent wider adoption for development.
things like scrollview having “special-case” logic built into the base class for use with date picker, instead of sub-classing it for just the functionality dp requires.  that’s the sort of design that screams “please give me unintended consequences and obscure regression bugs”.  (fe, you fix your scrollview, but therein break your datepicker)
or docs like “all widgets are groups… except scrollview, tableview”, without bothering to say why.  (would have rather seen “MOST widgets are groups, notable exceptions are … because …”)  that sort of half-doc’d stuff litters the code.
it’s “just complicated enough” that newbies probably wouldn’t touch it at all, medium-level devs might get something half-fixed and possibly suffer the above “consequences”; advanced devs might just pass it by as unmaintainable and just roll their own from scratch.
that’s my impression, $0.02, fwiw.  but i still think it’s a great that the source is available.