The Elastic Documentation the "tokenizer" parameter is officially deprecated, but from a functionality perspective, it seems to have been completely removed as the parameter is ignored in new indices.
This is a pretty drastic change. For the normal case, that saves a lot of pre-processing on the synonyms file that most people probably weren't doing. However, manipulating punctuation has been a historical work-around to assorted problems that pop up with phrasing, and there are plugins that depend on this behavior.
I'm not asking about new workarounds to a specific problem, just curious about the reasoning behind disabling the functionality entirely, rather than changing the default behavior.