I'd say that the CGOS server has been an invaluable spur to development, since
it does allow fairly easy testing against the competition.
What Don seems to be proposing is a way of standardizing the hardware - all
programs run on the same platform.
It seems that this would require an organization to set up a farm of identical
Go playing machines, each running a different program.
Don raised the question about what a contest between the single-threaded
ManyFaces versus multi-threaded Mogo would prove. Fotland says his program does
not scale to multiple cpus; Mogo does. While it is interesting to know how the
two programs fare on identical hardware, it is also interesting to find out
whether multiple cpus can lead to a higher standard of play.
My guess is that if single-thread programs were routinely slaughtered by an
8-cpu version of Mogo, developers just might seriously consider the nontrivial
task of rearchitecting their wares to be more scalable - especially since so
many consumers already have dual-core cpus, will soon have quad-cores, and can
expect 8-way machines to be widely available in a year or so. It is a big task,
no doubt of it; but failure to adapt will lead to loss of market share to more
nimble competitors.
____________________________________________________________________________________
Building a website is a piece of cake. Yahoo! Small Business gives you all the
tools to get online.
http://smallbusiness.yahoo.com/webhosting _______________________________________________
computer-go mailing list
[email protected]
http://www.computer-go.org/mailman/listinfo/computer-go/
