Aaron Putnam Jul 27 1992, 11:16 am show options Newsgroups: rec.games.chess From: a...@cie.uoregon.edu (Aaron Putnam) - Find messages by this author Date: 27 Jul 92 23:07:19 GMT Local: Mon, Jul 27 1992 4:07 pm Subject: ICS Reply to Author | Forward | Print | Individual Message | Show original | Report Abuse The ICS formerly on eaststar.utah.edu has finally returned. The new address and port are: valkyries.andrew.cmu.edu 5000 The network there is extremely fast, and the machine isn't a snail either. I am considering a local ratings server, to nullify netlag caused by the RS connection, but I would rather access the current ratings if the connection is fast enough. A suggestion was made by trb on eve, to handle the ratings server connection with a separate process, but this presents alot of tricky problems (such as password verification). It's a great idea, if it's feasible to implement. I'll look into it. -Aaron / -Desmodus PS - Please send mail regarding the server to a...@cie.uoregon.edu Andrew Tannenbaum Jul 28 1992, 7:12 am show options Newsgroups: rec.games.chess From: t...@ima.isc.com (Andrew Tannenbaum) - Find messages by this author Date: Tue, 28 Jul 1992 18:35:50 GMT Local: Tues, Jul 28 1992 11:35 am Subject: Re: ICS Reply to Author | Forward | Print | Individual Message | Show original | Report Abuse In article <1992Jul27.230719.25...@nntp.uoregon.edu> a...@cie.uoregon.edu (Aaron Putnam) writes: > A suggestion was made by trb on eve, to handle the ratings server > connection with a separate process, but this presents alot of tricky > problems (such as password verification). It's a great idea, if it's > feasible to implement. I'll look into it. Let me make my idea clearer, as I mentioned it to Desmodus in a brief ICS "tell conversation." Right now, it is apparently a major problem that when a game finishes on the ICS, it takes about five seconds to adjust the ratings and perform the database update. In the current implementation, this locks out all other ICS interaction during this operation. If the update is a separate process, then it can happen concurrently (from a process scheduling viewpoint) to the normal match-maintenance functions of the ICS. I think Aaron is missing a point when he brings up "the tricky problems." He is talking about "verification problems," I assume due to his visions of people bringing up fake ratings servers. My suggestion is to have the ICS fork off a copy of itself to perform ratings service. You get concurrent processing, and no questions about verification. Andrew Tannenbaum Interactive Cambridge, MA +1 617 661 7474 Mike Moore Jul 28 1992, 9:29 am show options Newsgroups: rec.games.chess From: mmo...@mesa.dsd.es.com (Mike Moore) - Find messages by this author Date: Tue, 28 Jul 92 21:16:49 GMT Local: Tues, Jul 28 1992 2:16 pm Subject: Re: ICS Reply to Author | Forward | Print | Individual Message | Show original | Report Abuse In article <1992Jul28.183550.22...@ima.isc.com>, t...@ima.isc.com (Andrew Tannenbaum) writes: [... stuff deleted...] |> functions of the ICS. I think Aaron is missing a point when he brings |> up "the tricky problems." He is talking about "verification problems," |> I assume due to his visions of people bringing up fake ratings |> servers. My suggestion is to have the ICS fork off a copy of itself |> to perform ratings service. You get concurrent processing, and no |> questions about verification . |> Richard has what I believe to be an adequate solution to the verification problem already in place. This is how I recall it working, though I could be mistaken.... Each ICS administrator is given a password for their server. It is only with this password that the database can be modified. Each administrator is given a different password, and that password is only valid for their site. This is fairly simple, but as long as the sysadmins don't go giving out the password to someone with access to the machine ICS is run on, there shouldn't be a problem. So I guess the point is that verification should not be a problem regardless of whether or not the IRS (Internet Ratings Server) business is done as part of the ICS process or as a separate process, so long as it is on the same machine. There are of course always ways around it, but I hardly think they would be worth the trouble. -Michael -- --------------------------------------------------------------------- |Michael Moore \ "Look, don't you understand?", | |Evans & Sutherland Employee \ shouted Arthur, "That man | |University of Utah Student / wants to knock my house down!"| |Email : mmo...@dsd.es.com / ''Well, he can do it while | |Home Phone : (801) 359-6552 \ you're away can't he?'' | |Office : (801) 582-5847 x 4721 \ -The Hitchhiker's Guide | --------------------------------------------------------------------- Aaron Putnam Jul 28 1992, 9:45 am show options Newsgroups: rec.games.chess From: a...@cie.uoregon.edu (Aaron Putnam) - Find messages by this author Date: 28 Jul 92 21:37:43 GMT Local: Tues, Jul 28 1992 2:37 pm Subject: Re: ICS Reply to Author | Forward | Print | Individual Message | Show original | Report Abuse In article <1992Jul28.183550.22...@ima.isc.com> t...@ima.isc.com (Andrew Tannenbaum) writes: >In article <1992Jul27.230719.25...@nntp.uoregon.edu> a...@cie.uoregon.edu (Aaron Putnam) writes: >Right now, it is apparently a major problem that when a game finishes >on the ICS, it takes about five seconds to adjust the ratings and >perform the database update. In the current implementation, this locks >out all other ICS interaction during this operation. > If the update is a separate process, then it can happen concurrently >(from a process scheduling viewpoint) to the normal match-maintenance The game-reporting would be quite easy to implement. (It's already set up with a queue, it would just have to be separated). >functions of the ICS. I think Aaron is missing a point when he brings >up "the tricky problems." He is talking about "verification problems," >I assume due to his visions of people bringing up fake ratings >servers. My suggestion is to have the ICS fork off a copy of itself Fake rating servers would be impossible (except by doing some tricky internet manipulation, and I don't think anyone wants to risk prison in order to raise their blitz rating a couple hundred points). What I mean is that when someone is logging in, their password has to be verified. This would mean that some sort of 'event' record would have to be added to the player - again, not a difficult thing, but it would mean alot of recoding. If something like that were to be done, I would prefer to rewrite the entire game into an event-driven system, which I believe would solve practically all of the problems that exist currently. > Andrew Tannenbaum Interactive Cambridge, MA +1 617 661 7474 -Aaron Putnam Eugene Stark Jul 29 1992, 12:33 am show options Newsgroups: rec.games.chess From: s...@cs.sunysb.edu (Eugene Stark) - Find messages by this author Date: Wed, 29 Jul 1992 13:14:52 GMT Local: Wed, Jul 29 1992 6:14 am Subject: Re: ICS Reply to Author | Forward | Print | Individual Message | Show original | Report Abuse >would mean alot of recoding. If something like that were to be done, >I would prefer to rewrite the entire game into an event-driven system, >which I believe would solve practically all of the problems that >exist currently. > -Aaron Putnam YES! Joel Lecorre Aug 1 1992, 5:13 am show options Newsgroups: rec.games.chess From: joel.leco...@synapse.isis.org (Joel Lecorre) - Find messages by this author Date: 31 Jul 92 19:34:00 GMT Subject: ICS Reply to Author | Forward | Print | Individual Message | Show original | Report Abuse To: All From: joel.leco...@synapse.isis.org Hi everybody, This is one of my first messages on Internet. I keep reading the word ICS in many messages and am intrigued. What exactly is this organization? Who do I contact to register? Does one need special software to play? Is it accessible for me in Canada? Thank you in advance --- . OLX 2.1 TD . I'm not dishonest, I'm morally different