On Sat, Sep 17, 2016 at 3:06 AM, Imants Cekusins <imantc@gmail.com> wrote:
that's the while point. If we could agree on a standard serializeable model,

That seems like a big "if". Especially since many dev tools exist to extend the model, and quite aside from "so where's the 'standard' now", conflicts you can currently control (mostly) suddenly become problematic. (I'm tempted to point to how gtk2hs's configuration phase works. pTk may be an even more severe example, although non-Haskell.)

--
brandon s allbery kf8nh                               sine nomine associates
allbery.b@gmail.com                                  ballbery@sinenomine.net
unix, openafs, kerberos, infrastructure, xmonad        http://sinenomine.net