Mon, 03 Feb 2014 10:14:18 +0100 blanchet added a new version of 'metis' to the mix
Mon, 03 Feb 2014 10:14:18 +0100 blanchet implemented new 'try0_isar' semantics
Mon, 03 Feb 2014 10:14:18 +0100 blanchet tuning
Mon, 03 Feb 2014 10:14:18 +0100 blanchet better time slack, to account for ultra-quick proof methods
Mon, 03 Feb 2014 10:14:18 +0100 blanchet crucial fix: use right version of the step
Mon, 03 Feb 2014 10:14:18 +0100 blanchet more thorough, hybrid compression
Mon, 03 Feb 2014 10:14:18 +0100 blanchet tuning
Mon, 03 Feb 2014 10:14:18 +0100 blanchet got rid of 'try0' step that is now redundant
Mon, 03 Feb 2014 10:14:18 +0100 blanchet centralize more preplaying
Mon, 03 Feb 2014 10:14:18 +0100 blanchet tuning
Mon, 03 Feb 2014 10:14:18 +0100 blanchet centralize preplaying
Mon, 03 Feb 2014 10:14:18 +0100 blanchet tuned
Mon, 03 Feb 2014 00:22:48 +0000 paulson fixed indentation
Sun, 02 Feb 2014 21:48:28 +0000 paulson new lemmas involving phi from Lehmer AFP entry
Sun, 02 Feb 2014 20:53:51 +0100 blanchet more data structure rationalization
Sun, 02 Feb 2014 20:53:51 +0100 blanchet tuning
Sun, 02 Feb 2014 20:53:51 +0100 blanchet more data structure rationalization
Sun, 02 Feb 2014 20:53:51 +0100 blanchet rationalized threading of 'metis' arguments
Sun, 02 Feb 2014 20:53:51 +0100 blanchet refactored data structure (step 3)
Sun, 02 Feb 2014 20:53:51 +0100 blanchet refactoring of data structure (step 2)
Sun, 02 Feb 2014 20:53:51 +0100 blanchet 'primitive' is not an adverb
Sun, 02 Feb 2014 20:53:51 +0100 blanchet unform treatment of preplay_timeout = 0 and > 0
Sun, 02 Feb 2014 20:53:51 +0100 blanchet refactor data structure (step 1)
Sun, 02 Feb 2014 20:53:51 +0100 blanchet tuned code
Sun, 02 Feb 2014 20:53:51 +0100 blanchet tuned factor
Sun, 02 Feb 2014 20:53:51 +0100 blanchet tuning
Sun, 02 Feb 2014 20:53:51 +0100 blanchet made SML/NJ happier
Sun, 02 Feb 2014 20:53:51 +0100 blanchet take intersection rather than union of methods when merging steps -- more efficient and natural
Sun, 02 Feb 2014 20:53:51 +0100 blanchet merge proof methods
Sun, 02 Feb 2014 20:53:51 +0100 blanchet use Skolem proof methods appropriately
(0) -30000 -10000 -3000 -1000 -300 -100 -50 -30 +30 +50 +100 +300 +1000 +3000 +10000 tip