Author: Don Dailey
Date: 16:20:19 06/02/98
Go up one level in this thread
On June 01, 1998 at 18:05:50, Tom Kerrigan wrote: >I just realized that the CPIP protocol spec that I had on my Web page >was all but unreadable. > >I converted it to HTML and now I have a link to it in the same place. > >Hope this helps everybody! > >www.frii.com/~kerrigan > >Cheers, >Tom I support any effort to make a truly standard interface. Also it should be combined (either by you or someone else) with a quality autotester. There seems to be an autotester out there that is a kind of standard but it's limited to windows and is commercial I think. It's my understanding that it does not check for time control violations but this may be wrong, or perhaps it's been improved. Anyway, I strongly support your efforts and would like to see a standard that is non-commercial. And also cross platform. It will probably be a lot easier to do the unix tools but a real effort should be made to do this for all possible platforms. If we start with your protocol or something like it, probably the tools would be built. I'm not sure how Edwards stuff figures in to all of this either. A few years ago I built a good autotester that did not require any support from the programmers. It was for DOS and I basically had a memory resident routine that tested the graphical output for state changes 3 or 4 times per second. When it found and decoded a move, it would pass the info to the other computer via the serial port. The move would be applied by using a keyboard stuff routine to simulate physical keypresses. It was very reliable (after lot's of debugging) and used almost no additional resources. You only needed 2 computers. The overhead was a tiny fraction of a percent, it was almost immesurable and both programs had the same overhead although one was designated the master. When I got a new program, I would write a simple configuration file and this would make the new program testable. It was very cool. The configuration file would identify key pixels, and keystrokes required to operate various features. I think it would be desirable to have this for windows, and with the option for a program to donate this information to the interface. I don't know if programming this for windows would be as easy since you are probably more isolated from low level stuff. Just a few thoughts. - Don
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.