smart@ditmela.oz (Robert Smart) (02/03/89)
I am amazed that there don't seem to be protocols and programs for supporting game playing over an internet network. The sort of thing I envisage would allow a client program to connect to a game control server. They could then start a game, or join an existing game. The protocol would include the ability to send move information to the control server, receive state-of-game information from the control server, send text information to opponents, opponents and partners, just partners [this last not allowed in some games, e.g. Bridge!]. Given the appropriate control server it should then be easy to modify programs like the Sun gammontool and chesstool to provide nice human interfaces to various games. I really feel that something like this would be extremely worthwhile in demonstating the sort of things that you can do with networks. I try to convince people that interactive computer networking doesn't have to mean that you use the network to log into an interactive account on a remote machine. In fact logging in over a network tends to be both a horrible security problem and a very inefficient use of network resources. But when you try to describe other ways that a network could be used interactively, there aren't a lot of examples: remote disk and file access and the talk/phone system (which isn't used a lot). I am sure a system such as I propose would inspire people with interests in worthwhile activities like remote education, learning through serious role playing and computer conferencing. Anybody else interested? Any existing work? Should we write an RFC? Bob Smart, CSIRO Division of Information Technology, Australia smart@ditmelb.oz.au (or smart%ditmelb.oz.au@uunet.uu.net)
cheriton@PESCADERO.STANFORD.EDU (David Cheriton) (02/06/89)
VMTP (RFC 1045) is intended support the type of communication required for distributed games (as well as other uses). We are in the middle of modifying mazewars to use VMTP and IP multicast, and will announce it when available. We are not planning to solve all the directory issues raised by needing to register games, locate a particular game, etc. I think a network directory service that could record this information is the major missing piece. X.500? David Cheriton
CERF@A.ISI.EDU (02/06/89)
There have been some interesting dynamic games run on low-delay nets, mostly LANs (e.g. XEROX PARC has several, but I don't remember the names - one involved a chase though a maze of corridors and you got zapped if a big, CBS-like EYE was caught staring at you - your opponent). What might be interesting is a delayed/deferred kind of game which you could essentially join or leave at any time, catch up on, as in computer-based teleconferencing, etc. Maybe some sort of group adventure? I suppose some board games would work if the delays were reasonably short. Chess obviously works fine in delayed mode. Vint
david@ms.uky.edu (David Herron -- One of the vertebrae) (02/06/89)
Ah, but there are some such games. For instance, empire. In any case, my impression is that not much noise is made lest management take a dim view of this sort of use of the network. -- <-- David Herron; an MMDF guy <david@ms.uky.edu> <-- ska: David le casse\*' {rutgers,uunet}!ukma!david, david@UKMA.BITNET <-- Now I know how Zonker felt when he graduated ... <-- Stop! Wait! I didn't mean to!
dwaitzma@bbn.com (David Waitzman) (02/06/89)
In article <4051@ditmela.oz> smart@ditmela.UUCP (Robert Smart) writes: >I am amazed that there don't seem to be protocols and programs for >supporting game playing over an internet network. The sort of thing I >envisage would allow a client program to connect to a game control >server. They could then start a game, or join an existing game. The >protocol would include the ability to send move information to the >control server, receive state-of-game information from the control >... >Anybody else interested? Any existing work? Should we write an RFC? >... May I suggest that new work in this area exploit IP Multicasting? Many-player games (and conferences) certainly could use multicasting, and multicasting certainly could use games to become more widely valued. Please see RFC-1054 and RFC-1075 for more information. thank you, -david (dwaitzman@bbn.com)
jg@jumbo.dec.com (Jim Gettys) (02/06/89)
In article <[A.ISI.EDU].5-Feb-89.12:53:46.CERF> CERF@A.ISI.EDU writes: >There have been some interesting dynamic games run on low-delay >nets, mostly LANs (e.g. XEROX PARC has several, but I don't remember >the names - one involved a chase though a maze of corridors >and you got zapped if a big, CBS-like EYE was caught staring at >you - your opponent). There are quite a few such games for X these days, including a re-implementation of MAZEWAR (which actually precedes XEROX PARC; I played it in 1972 or 1973 at MIT on IMLAC displays). The possibly dubious claim was made to me that it had been banned from the Internet because of excessive use between MIT and Stanford, which was skewing network use statistics. (Maybe someone out there actually knows....). There is also xtrek, and at least one other empire like game, which are network based. Only mazewar is really distributed; the others involve invoking a single server which then use multiple machines for display, which X facilitates. I guess I know X has succeeded by the proliferation of games for it... :-) Jim Gettys
VAF@SCORE.STANFORD.EDU (Vince Fuller) (02/06/89)
The two most memorable games from the Xerox PUP days were MazeWar (which Vint Cerf described) and AltoTrek, in which you controlled a ship on one of three teams and went around blasting members of the other teams and planting your bases around various planets (which could later be used for purposes such as spying and self-distruction, if other players approached). Xtrek is a game similar to AltoTrek which is implemented using the X protocol (which is layered over TCP/IP) and I believe an almost identical copy of MazeWar has also been implemented using X. I believe there are also others of this sort: Xtank and Xconq (multi-player Empire), though I haven't seen them myself. --Vince -------
vjs@rhyolite.SGI.COM (Vernon Schryver) (02/06/89)
In article <4051@ditmela.oz>, smart@ditmela.oz (Robert Smart) writes: > I am amazed that there don't seem to be protocols and programs for > supporting game playing over an internet network.... > Anybody else interested? Any existing work? Should we write an RFC? > > Bob Smart, CSIRO Division of Information Technology, Australia > smart@ditmelb.oz.au (or smart%ditmelb.oz.au@uunet.uu.net) A standard would be nice. It should support not only "low-speed" games such as chess or bridge, but also those which require at least a couple dozen updates/second, such as Silicon Graphics' "arena" and "dogfight." "Dogfight" has been a very useful tool for these past several years. It has sold lots of systems, and, since it was converted from XNS-multicast to UDP-broadcast, killed a number of obsolete machines which happened to be on the same networks. (We're working on IP-multicast so "dog" can get across gateways as well as be friendlier to old stuff.) Vernon Schryver Silicon Graphics vjs@sgi.com
philipp@physicsa.mcgill.ca (Philip Prindeville Comp Ctr) (02/06/89)
I would be more inclined to think of something like Project Athena's Zephyr notification system, but that could be out of lack of respect for X.500. See the Winter 1988 Usenet proceedings for a description or ftp it from athena-dist.mit.edu, in pub/usenet (I think)... -Philip
kent@WSL.DEC.COM (02/07/89)
There have been some interesting dynamic games run on low-delay nets, mostly LANs (e.g. XEROX PARC has several, but I don't remember the names - one involved a chase though a maze of corridors and you got zapped if a big, CBS-like EYE was caught staring at you - your opponent). I ported Mazewar to X and UDP a while back. Other people have built distributed games on top of IP (xtrek and xconq come to mind). I'm not sure that a centralized server/protocol would have done me a lot of good -- there's a lot of traffic and the extra hops can become significant. In Mazewar, one of the games acts as a central server for people who join the game and does "rebroadcast" of some of the packets, but there are also some short circuits for efficiency. I only use broadcasts to try to find an existing game. Of course, if there had been an existing protocol implementation, I might have built on top of it, and not noticed any loss in game responsiveness. What might be interesting is a delayed/deferred kind of game which you could essentially join or leave at any time, catch up on, as in computer-based teleconferencing, etc. Maybe some sort of group adventure? This sort of game has been built, too. Peter Langston's empire for Unix, built starting in the late 70s, has contributed to more than one academic suicide. (it's a highly detailed world conquest game, based on a board game played at Reed College.) I've lost track of the game in recent years, but he developed it actively for quite some time. As many as 20 players could play in a single game, playing moves when they wished. Games sometimes lasted weeks on end... chris
griefer@ibmarc.uucp (Allan D. Griefer) (02/08/89)
You might look at DREGS, a distributed games package from the University of Wisconsin.
tmac@SOL.ENGIN.UMICH.EDU (02/09/89)
I haven't actually played xtrek (though we have it) but xconq I have played over the network and it does provide an intersting effect to the game. I wouldn't mind seeing more like this. Plus, building it using the X toolkit provides (though xconq could use a little improvement) a good user interface. _Tom
CERF@A.ISI.EDU (02/10/89)
Chris, By academic suicide, I take it you mean failing to pay attention to studies, etc. Is Langston reachable? The distributed game sounds worth writing a paper or at least a column. Vint
kent@WSL.DEC.COM (02/10/89)
Vint, By academic suicide, I take it you mean failing to pay attention to studies, etc. Yup, exactly. Someone wrote to me and said that a modern version of the game is available on ucbvax.berkeley.edu. chris
sean@ms.uky.edu (Sean Casey) (02/11/89)
In article <[A.ISI.EDU].9-Feb-89.18:36:57.CERF> CERF@A.ISI.EDU writes: >Is Langston reachable? The distributed game sounds worth writing >a paper or at least a column. Langston is not really reachable, but UCSD empire is only based on Langston empire. It is growing into another creature altogether. You can find it on ucbvax.berkeley.edu in /empire. There are scads of multiplayer games using Berkeley INET sockets, you just have to look hard. Sean -- *** Sean Casey sean@ms.uky.edu, sean@ukma.bitnet *** Who sometimes never learns. {backbone site|rutgers|uunet}!ukma!sean *** U of K, Lexington Kentucky, USA ..where Christian movies are banned. *** ``Fate? I thought you said Freight.''
werme@Alliant.COM (Ric Werme) (02/12/89)
With the proliferation of X, I imagine a feature of interactive games would be a common user interface throughout the net. Keep in mind the possibilities of allowing individual sites or users to develop their own interface. Once upon a time, on an ARPAnet long, long ago (1973?) the Friday hackers forum at MIT start a discussion on a network-wide Star Trek game. I was C-MU, I think we had input from Perdue and Stanford. The idea was to inspire work on network graphics, a task that was languishing because everyone's system was different. We toyed around with the idea of developing Star Trek Protocol (STP) and servers on each machine that were identical. Each server would be responsible for some of the postional calculations, communicating with the other servers, and with local game clients. This would allow each site to write custom clients and encourage people to come up with really good, efficient, and informative interfaces for the game players. (And to develop automated players if there weren't enough real people around or if you wanted some extra help.) It had a lot of possibilities, but none of us had the time to devote to it and nothing came of it, as far as I know. Someone at PARC in a more authoritative position put forth a very similar proposal, but one aimed more at network graphic protocols than my interest in specialized clients. In 1974 I started fading from the ARPAnet. First to DEC, then to a dot matrix printer companies, then to starting a company to write a music teaching program for Apples. Now to Alliant and the Usenet. Maybe I'll make it back to the Internet someday. But I still won't have any time for STP! -- | A pride of lions | Eric J Werme | | A gaggle of geese | uucp: decvax!linus!alliant | | An odd lot of programmers | Phone: 603-673-3993 |