unbeamed 32nd stem is shortened by 0.25 ss to fit beamed stems better. (issue4243071)

[lemniskata . bernoullego]

Graham wants us to discuss this, so let's do it :)

I think that this is the only reasonable amount by which we can change
32nd stem length (4.25 ss instead of 4.5 ss); any remaining
inconsistencies in proof-sheet posted by me
(http://www.sendspace.com/file/2mzt4a) should be fixed by tweaking beams
(if they are possible to fix at all). But fixing beams is a whole
another topic, and a huge one. I'm already in the process of gathering
examples for this, but it will take a lot more time.

cheers,
Janek

http://codereview.appspot.com/4243071/