Closed Blaisorblade closed 10 years ago
I feel that language equality should be nominal.
I feel that language equality should be nominal.
Makes sense. In which case, Blaisorblade/pts@3e876fe9911e9a321b2adc22e5f58c5e08dfb481 is at least correct.
Closing this for now, I feel the current resolution is sufficient.
Blaisorblade/pts@3e876fe9911e9a321b2adc22e5f58c5e08dfb481 opens a bug's nest — how do we properly compare PTSs/languages, taking infinite languages into account? The solution I implemented won't work well enough for externally specified languages, but comparing the PTSs themselves won't work easily for infinite PTSs.
We need to choose among nominal or structural equality (named after nominal vs structural subtyping).