cognitive psychology. PhD 
2/ Compared to other formal methods, type systems provide a nice sweet spot: * lightning fast speed * convenience (no type annotations w/ inference, no manual proofs to write) * confidence, via soundness
Show this thread
-
Yeah, I really like the idea of allowing the user to pop the hood. Could be fun to try a naive thing like just show a graph representation of the constraints. Probably horrifyingly large for real programs, but maybe you can prune? And then link to the source somehow too?
Yeah, that approach turns it into a data vis problem. Probably want some kind of Google Map-esque zooming approach, viewing at different layers of abstraction. Very similar to how hw engineers use visual layouts for designing/debugging whole chips vs. individual components.
-
-
At OPLSS, I was talking to someone about usability of theorem provers. IIRC they said when their prover failed, they would open the hidden text logs to see the proof tree it spit out and find the offending node. That was the signal for where to start fixing the proof.
-
More generally, we **have** to move beyond ASCII diagnostics. We can push termcolor, ncurses, etc. to their limits but eventually we will need interactivity, direct manipulation, functional reactivity, etc. in our error messages.
End of conversation
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.