Having followed this discussion for a while, I'm trying to put my finger on why I feal uneasy about it.
The major features I see in this proposal are: * New syntax for trying multiple assignments until one matches. * Extending destructuring assignment to match constants in the LHS. * Adding syntax to tell constants from targets. * Extending destructuring assignment to match isinstance and properties. * Making _ a special variable that doesn't always gets assigned a value.
I see the other features are only discussed as part of the first one. I think each feature has (some) value on its own, could be implemented without the others, so really should stand or fall on it's own merit.
And there I found the root of my uneasiness - the supporting cast is stronger than the main actor. To me, that suggest we should change focus, to look at the usefulness of these extensions in themselves, relegating the match/case statement to syntactic sugar.