Indeed, the guarantees you point out are the reason that STM is a highly composable concurrency abstraction in Haskell.
Otherwise, you just don't get that win.
-David
In reply to Re^3: Where is concurrency going? Is Perl going there?
by erroneousBollock
in thread Where is concurrency going? Is Perl going there?
by erroneousBollock
| For: | Use: | ||
| & | & | ||
| < | < | ||
| > | > | ||
| [ | [ | ||
| ] | ] |