Wording in sections 2 and 3, with attention on a couple of points:
authorNorman Ramsey <nr@cs.tufts.edu>
Wed, 28 Jul 2010 20:59:07 +0000 (16:59 -0400)
committerNorman Ramsey <nr@cs.tufts.edu>
Wed, 28 Jul 2010 20:59:07 +0000 (16:59 -0400)
commit7974cfeb38ea80dbe9010f32b4935e6e39779765
tree7b74ad2341e6e1843c3552a1f4a861debfbca5e4
parent7393f7d50d2d16a72c3fb5d66cb9af52ae30cd68
Wording in sections 2 and 3, with attention on a couple of points:

  - don't even suggest that the empty fact is an appropriate
    start fact for constant propagation

  - the example that mentioned loops wasn't a loop

  - point out in section 3.1 that shape is a new thing
paper/dfopt.tex