I have twice written here that there was a bug in the routines I used to identify complete anticipations, which are those that share an exact position, possibly with reflection and/or rotation. I am glad to say that I have now found that bug and, because is was a bug in the design, completely rewritten those routines. That reworking is not quite finished (I still have to look for positions that are both reflected and rotated), but I have now identified several more complete anticipations, and lots of reflected duplicates. The anticipations have been tagged and the duplicates have been deleted. This explains why, even given the problems that I have had time to enter in the last few days, there are very few more problems in Meson than there were at the time of my last post.
All of the above has interrupted my investigation of 'AntCheck' output and I still have about 200 anticipation/posticipation pairs to check out. There was so much output this time that I decided to delete any output with a comparison ratio of less that 85%, which amounted to about 700 pairs. Too many of those at less than 85% were false positives and I was wasting a lot of time.
WCSC preparation work is beginning to take over, so it is likely that posts here will become even more infrequent until I have everything ready for Rhodes.
Subscribe to:
Post Comments (Atom)
0 comments:
Post a Comment