Author: Robert Hyatt
Date: 07:52:40 06/03/05
Go up one level in this thread
On June 02, 2005 at 16:36:55, Matthew Hull wrote: >On June 02, 2005 at 14:38:28, Robert Hyatt wrote: > >>On June 01, 2005 at 08:16:06, Matthew Hull wrote: >> >>>On June 01, 2005 at 04:37:04, Stan Arts wrote: >>> >>>>Hi Ian, thanks for the link. >>>> >>>>It seems like a large (and unreadable) program for it's day. (before my time, >>>>i wasn't even born yet.) >>>> >>>>Is there a working executable available? (Can't get it to work with FreePascal, >>>>because of those intra-procedure goto's. ..but can hardly blame a compiler for >>>>that. Procedure parameters are still modern though in Pascal, if we mean the >>>>same.) >>> >>> >>>Got it compiled with gpc (GNU pascal compiler) and it "runs", but the move >>>syntax is some kind of screwed up descriptive notation. It could be the OCR >>>process made some workable typos or some of the program is missing. You can't >>>actually enter a valid move, so it makes a move for you. It moves instantly and >>>you can't tell if it is doing any kind of meaningful search. There is no >>>obvious command table. >>> >>>I can't make heads or tails of it. One would need to study it a little to fix >>>it up. >>> >>> >> >>One point. Back when that was written and published in Byte, the common move >>input format was "English Descriptive". The old P-K4, N-KB3, kind of move >>syntax, and it probably would not accept lowercase either... :) > > >I figured that out pretty quickly as it gave moves in that format. Also, >someone posted a useful README file which provided code corrections for >lowercase and ASCII board representation. However, there still seems to be a >bug that prevents it from parsing user supplied moves, so it assumes you want it >to make a move for you, which it does. I've never been able to enter a move >successfully. It's probably easily solved, though. But I might rather study a >more modern program. > >It is cool, though. > Another issue. CDC was a 60 bit word, and they used their own unique character set for representing characters, based on a 6 bit encoding. I'd bet that Atkin did some trick on converting the input into some internal form they could use. For example, to convert an ASCII zero (0) to binary, just subtract 0x30, or AND with 0x0F, or whatever. The CDC characters are nowhere near ASCII nor the old EBCDIC standard. I had to use a CDC once to input moves for Cray Blitz, and I had to rewrite the move input code because the CDC could not handle upper/lowercase whatsoever, yet Cray Blitz needed Bc4 and the like. I modified the code on the Cray to recognize UBC4 as "uppercase-B" "lowercase-C" "4" which was a kludge that worked. Probably you are hung up there somehow... > > >> >> >>> >>> >>>> >>>>Greetings >>>>Stan >>>> >>>> >>>> >>>>On May 31, 2005 at 21:30:22, Ian Osgood wrote: >>>> >>>>>Over on news:comp.lang.pascal, someone has unearthed the Peter Frey and Larry >>>>>Atkin Chess 0.5 source that was published in Byte Magazine back in 1978. >>>>>They've OCRed the source and managed to get it running on the GNU Pascal >>>>>compiler (which still supports that ancient ISO Pascal dialect, including >>>>>procedure parameters and intra-procedure gotos[?!]). >>>>> >>>>>Might be an interesting walk down memory lane... >>>>> >>>>>http://www.moorecad.com/standardpascal/misc.html >>>>> >>>>>Ian
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.