Well... you implemented a compiler. And hey, that's a great thing. I've implemented several myself. But I doubt that you proved that it was correct, or that it was a large-enough language that many people would want to use it directly. :-).

> I'm actually ignoring them on purpose to get a short-term win as you put it. Specifically, I'm expecting incremental, well-tested, inspectable compiler without expecting formal verification. I list it as necessary for full-trust but not necessary to get baseline up.

Okay. But as you note, you're also focusing on shorter-term wins... they're just different shorter-term approaches for for different goals. And that's okay.