Author rhettinger
Recipients dalke, jcea, rhettinger, terry.reedy
Date 2019-08-26.04:50:12
SpamBayes Score -1.0
Marked as misclassified Yes
Message-id <>
The anti-correlation algorithm seems like a plausible heuristic but it can't really know more than the user does about the semantic content of the sets.  Also as Terry pointed out, this will have user visible effects.

This likely should be published as a blog post or recipe.  A user can already control the pairing order and do this or something like it themselves.  It is more of a technique for using sets that it is a core set algorithm (it reminds me of using associativity to optimize chained matrix multiplications, though than can be done precisely rather than heuristically).
Date User Action Args
2019-08-26 04:50:12rhettingersetrecipients: + rhettinger, terry.reedy, dalke, jcea
2019-08-26 04:50:12rhettingersetmessageid: <>
2019-08-26 04:50:12rhettingerlinkissue13653 messages
2019-08-26 04:50:12rhettingercreate