It's been a bit of an open secret that the current path we're on with the SP is a dead end for technical and "people" reasons. We entered into a dialog with members to be above board about this and to table the question of whether there's a clear need to replace the SP with something ourselves or treat this as a problem with adequate existing solutions (other SAML modules, OIDC, "who cares about Apache in 2020?", etc.)
We've gotten plenty of feedback that doing something would be preferable to nothing. While we have not yet committed to a course, I have started to sketch out some design constraints for that on this page as well as (in much less detail) a high-level strawman of what we might do.
We would like to gather feedback to those ideas or get new ideas, as well as solicit any interest in working with us on this if you're so inclined. Right now, we haven't really assigned anybody specific to work on this or slotted it into a schedule of work other than to assume that it needs to happen within the next few years.
Anybody is welcome to add comments to the page or (preferably) just express themselves on the dev list about any aspect of this.
I don't know exactly where this is going to go or when but I would guess that once 4.1 is done and stablized, that I'll be retasking myself toward this work in either a direct or an oversight capacity depending on who ends up working on it, so I expect some level of activity around this during the second half of 2021 and in earnest in 2022-2023 if it happens.