[cap-talk] definition of the term "safe language"

[Sandro Magi]

On 09/04/2010 2:15 PM, Sandro Magi wrote:
>> Not according to your definition perhaps, but according to Pierce's it does.
>> Java's floating point abstraction admits undefined behavior.
> 
> So does non-deterministic choice, but admitting non-determinism does not
> by itself make a language unsafe and you have not explained how it
> could. There are many safe languages with non-determinism.

Correction: Java's FP abstraction admits non-deterministic behaviour,
not undefined behaviour. This is a critical point which I assumed was
understood but forgot to mention explicitly, and important to the
overall reason Java remains safe. More when I respond to David's e-mail.

Sandro