On Tue, Nov 25, 2008 at 22:44, Roger Hicks <[EMAIL PROTECTED]> wrote: > On Tue, Nov 25, 2008 at 23:32, Charles Reiss <[EMAIL PROTECTED]> wrote: >> That's what I'd expect in this form given that expecting people to >> know how to program usually isn't considered unreasonable. But, of >> course, there are practical problems with doing any-language-you-want >> (do you have a Hypertalk interpreter? Z80 assembly? VisualWorks?), so >> things probably need to be more restricted in practice. Probably the >> closest canonical example of a contest is the (much less theoretically >> interesting) http://www.cs.ualberta.ca/~darse/rsbpc.html . >> > Why not have the competing programs communicate via an HTTP post?
That's a viable solution of course. Pros: Less contestmaster work. More programming environment choice. Cons: Making sure all programs are available at the same time. People going against the spirit of the game could use manual intervention to change strategy (timeouts can disincentivize this). Harder to run a whole lot of rounds (which would give a clearer idea of winner). Neutral: Likely encourages more complex solutions (using large external libraries, large datastores, etc.) -woggle
