Throwing subpar ideas out of the window is part of the process, so no worries about my feelings. I'll just outline the thoughts that went into my other post if

Throwing subpar ideas out of the window is part of the process, so no worries about my feelings. I'll just outline the thoughts that went into my other post if you are interested in critizing them. I had also briefly considered a relay at the closest Lagrange point but came to the same conclusion as you did. Namely that off-the-shelf Starlink satellites might not be able to communicate with it which just means: it's not the solution. So let us consider a (few) separate relay satellite(s) [*] after quickly dismissing lunar ground stations: a lunar relay ground station might have the advantage of always being in sight of Earth but has the considerable disadvantage of having to live through lunar nights. This is highly impractical and would incur considerable complexity by requiring e.g. a RTG power source instead of using solar power as sats would do. So on-shell non-Starlink satellite relays it is for now. If there is more than one relay, any orbit might be fine [**]. If it really had to be a single relay satellite for some reason, a polar lunar orbit might be a possible solution that does not suffer from regular communications blackouts. Another solution might be to relax the requirement of having a continuous connection between the single relay satellite and Earth and to accept to be offline for a fraction of the orbit (at an orbital height of ~500km above Moon('s equator) a satellite would disappear behind Moon for a bit over a quarter of its orbital period), so if approximately 73% availability is fine, this would also work. Higher inclinations improve this value. One more thing because this also seemed unclear: at no point are Earth's Starlink satellites involved. All moving of data between Earth and Moon happens between the dedicated relay satellite(s) and a ground station on Earth, as this is the setup that NASA used for its laser optical communications test that I linked to from my other posting. All in all, that means that (under the assumption that Starlink satellites can operate under the conditions around Moon (it lacks Earth's protective magnetic field!)), that a few dedicated and custom-made relay satellites would all that is needed to use (my opinion:) dirt-cheap run-of-the-mill Starlink satellites for a constellation around Moon. Any thoughts on this? [*] Maybe even a few of them as material is cheap but engineers are expensive. If you can build one thing, you can usually build a second/third/n-th of it for less per-unit costs than the previous (with diminishing savings between each). [**] As stated in another post, I would rather expect the constellation to be in higher than in lower orbits so that fewer satellites would be needed to cover the same area on ground. This means that lunar mascons should not be of concern and even if they were, the satellites would still have ion thrusters to compensate for that.