Conversation

ᵉᵈᵈʸᵇ
@eddyb_r
me: I think I get it (Dunning-Kruger anyone?) but I've never tried implementing any of it and have trouble explaining it to other ppl so delegating is hard also featuring: writing a PL implementation (as it is sometimes called, a "compiler") is hard, I should write a PL for that
1
1
ᵉᵈᵈʸᵇ
@eddyb_r
the worst part is when you think you have a cool trick but you've never heard of it being in the literature, and the likelihood of outright inventing something new is pretty small as an amateur... so there's a nagging feeling that you might regret later doing it at all
1
1
ᵉᵈᵈʸᵇ
@eddyb_r
(in the sense of it being potentially unsound, or at the wrong level of abstraction I guess) there's a reason some people insist you should prototype everything in a proof assistant... but that brings me all the way back to "compilers are hard to write" :(
2
1