ordy@cwruecmp.UUCP (Greg Ordy) (01/08/84)
This message is in response to the article '776@garfield.UUCP' by Sean Byrne, regarding the Unix game 'search', and myself. First of all, I am still at Case Western Reserve, hard as it may be to believe. 'Search' long ago had its day at Case, and is no longer played here, although available. It was ported to V7 and 4.1 BSD by Sam Leffler, formerly of CWRU, and UCB. There was a rumor of a 'distributed' version for 4.2 BSD, however I don't know anything more about it. It was thought to be a good network exerciser. The game was originally 'released' at a Unix Conference, several years ago -- perhaps as far back as 1979 or so. Although local modifications were made here since, I don't think that there were any 'major' modifications. The 4.x BSD versions are certainly up to date with respect to our version, and thanks to Sam Leffler are nicely organized. More interesting to me, however, are modifications made to the game by other sites. Over the years I have heard rumors of several very interesting enhancements to the game. I would like to ask people with knowledge of such enhancements to mail me a summary, and if they are interesting, I will post them to this group in several weeks. The only major enhancement to the game I would have liked to have made (but never had time) was a programming language for describing the motions and actions of the participants in a game. Given such a language (and execution) implementation, the actions of the aliens could be made much more interesting, and players could take part in a game without actually directly playing. This brings me to an interesting point, which is actually why I chose to put this message in net.games, rather than reply by mail to Sean. I am suprized to see that somebody has not (to my knowledge) created a game which has a programmable 'strategy' so that Usenet users could accumulate a pool of such strategies, and run competitions between literally hundreds of users, in search of the 'best' strategy. The closest I have seen is the game 'Bolo' (I think) by Peter Langston (Mr. Empire). It is difficult, however, to create such a game, which lends itself to programmed strategy specification, is nontrivial, and is interesting. Such a game would also be a large programming project (for a game), because it would require a game language compiler, and the actual 'game', which would execute an arbitrary set of strategies, in a common competition. For the purposes of debugging, as well as optimization (and FUN), the game execution phase would probably require a graphic interface so that the user could observe the players in operation. This idea is nothing more than an extension to the various automated schemes to play games such as 'empire' and 'rogue', but carried further.