• Mark Evans (11/11) Mar 02 2003 Maybe they are not 'over the top' for D after all. Some excellent perfo...
• Mark Evans (10/10) Mar 02 2003 For additional reference on related concepts.

Mark Evans <Mark_member pathlink.com> writes:

Maybe they are not 'over the top' for D after all.  Some excellent performance
gains were achieved with continuations in the Mach kernel.
http://research.microsoft.com/research/pubs/view.aspx?tr_id=84

More editorial comments and links here.
http://rover.cs.northwestern.edu/~surana/blog/archives/000080.html

The Oz book also treats concurrency at some length, though it does not favor the
term 'continuation' per se.
http://www.info.ucl.ac.be/people/PVR/book.html

Continuations also facilitate what some call goal-directed evaluation in
algorithm work.

Mark

Mark Evans <Mark_member pathlink.com> writes:

For additional reference on related concepts.

http://www-106.ibm.com/developerworks/library/l-pygen.html
http://www-106.ibm.com/developerworks/library/l-pyth7.html

"Continuations are hard to sell. Coroutines and generators, and especially
microthreads are easier. All of the above can be implemented without having
explicit continuations. But when you have continuations already, you find that
the step to these other structures is quite small, and continuations are the way
to go. So I'm going to change my marketing strategy and not try any longer to
sell the continuations, but their outcome. Continuations will still be there for
those who can see the light."