[cap-talk] David Wagner's Google techtalk is now up!

[Sandro Magi]

Jonathan S. Shapiro wrote:
> A bit of follow up on this. Recently, Scott Smith, Swaroop, and I were
> discussing the typed I/O problem, and the apparent need for checked
> dynamic upcast in any sensible static language. We identified a set of
> criteria under which the result of a checked upcast may alias the
> untyped (really: vector of bytes) state originally delivered by the I/O
> system.

I don't completely follow the above, but it sounds like "Typing Dynamic
Typing" [1].

Sandro

[1] http://people.cs.uu.nl/arthurb/dynamic.html