Beware of dropping “entry points” to uLisp.
I am fond of the DUE, as a “most powerful Arduino board” from the quasi “classical” ones. Don’t forget that many people buy “an Arduino” — for support and compatibility reasons. Who cares how good your gadget is if nobody helps you with it? DUR is also the target of a lot of projects out there, like RunCPM; it is also used by Altair 8800 emulators. — Personally, I have outgrown it in practice, but I think it remains a nice “entry point” to such hacking adventures, AND to uLisp. If you drop it, I would just recommend that you support at least the Line Editor with it in a sort of “preliminary finalising release”.
The BBC micro bit is unfortunately even more of an “entry point” — though a pretty weak system. Thing is: a kid is LIKELY TO OWN a BBC micro bit. Like, there are other microcontrollers out there. But do people OWN them in practice?
Whether you “sufficiently care” about these considerations is, of course, perfectly your decision. (And a matter of experience: despite wide ownership of BBC micro bits, how often did anyone come here with one to ask for help? — On the other hand, quite a few fellows seem to be going for the Adafruit Grand Central M4, myself included, though it is by no means a “typical” system. So what matters “here” and “out there” may differ.) — What you could consider is doing a “Tier 2” categorization, the way BSD OSes do: they do not say that they let “Tier 2” quasi “die”, but they make clear that Tier 2 gets less attention. So you may leave these platforms at a “secondary” status, without entirely calling them unsupported evermore.
Can one get selected “past” uLisp releases? Particularly if you drop a platform, that one may become an issue.