The state of InforML

Still slogging along on the development of InforML. I've finished modifying a good portion of the system to collect constraints over labels instead of trying to resolve them eagerly. However, my early tests of this lead me to notice a difficulty in generating constraints. This of course lead me to a rather interesting research problem that I can't explore just at the moment.

So instead, I am trying to make the best of my current design by being very careful about α-conversion. Specifically, I am moving the internals over to use de Bruijn indicies. (Interestingly, last time I checked Wikipedia they didn't have an article on de Bruijn indicies, but it would seem Kaustuv created one just last Wednesday.) This, of course, has required painful amounts of boilerplate code. I'm almost at the point where I can pretty-print the result of my global renaming pass to verify that it works correctly. Switching to de Bruijn indicies also resolves my previously sketchy treatment of internal versus external names in modules nicely.

In any event, I should get back to writing code.

Leave a Comment