For what it's worth, you almost fooled me into thinking mututally recursive variants _couldn't_ capture CFG structure, even though before this discussion, I had taken it for granted.
Another guy I knew messed up in the opposite direction, and tried to enforce well-typedness in his AST types in SML. I sort of instinctively figured that couldn't work, but I didn't really know what was missing till I learned dependent types. On Wed, Jun 3, 2015 at 8:17 AM, Jonathan S. Shapiro <[email protected]> wrote: > There is definitely a joke on me in this discussion. > > I started with an assumption about what an AST structure looks like, and of > course it turned out that the assumption was the source of my confusion > about type threading. Well, no surprise there. I suppose we all know that > assumptions are dangerous things. :-) > > Now I have a different problem. Thanks to Matt I no longer know what I want > my AST structure to look like or how to build a tool to describe and check > it. > > Thanks a LOT, Matt! :-) > > Oh well. I asked for it, didn't I? And as usual you folks delivered.... _______________________________________________ bitc-dev mailing list [email protected] http://www.coyotos.org/mailman/listinfo/bitc-dev
