On May 17, 2020, I did a nearly complete rewrite of the Wikipedia article on 'The Minimum Description Length Principle'
. The purpose of this rewrite was to correct a very damaging misunderstanding of that phrase that pervades the machine learning and statistics community.
The misunderstanding is that most-widespread use of that phrase refers not to a "principle" but to a pragmatic degeneration of AIT.
This is very damaging because the degeneration of AIT, to non-Turing complete descriptions, leads to a pervasive impression that application of AIT is ineffective.
Because the subscribers to this list are more lettered, I would appreciate any reviews if not edits of my attempt to correct what I perceive as damage to the field.