Michael Moore Apr 9 1993, 11:15 am show options Newsgroups: rec.games.chess From: mmo...@bambam.es.com (Michael Moore) - Find messages by this author Date: Fri, 9 Apr 93 17:45:40 GMT Local: Fri, Apr 9 1993 10:45 am Subject: ICS timestamping Reply to Author | Forward | Print | Individual Message | Show original | Report Abuse Hello everyone, I just thought I would bring up an idea for discussion here by all the ICS users and administrators. I think most people would agree it would be nice to have a timestamp on the moves so that net-lag would cease to be as great a factor. The problem with the idea is that you can't trust the clients to be honest about that timestamping unless you create a password to be sent with timestamp packets (encrypted of course) and only distribute binaries with this password compiled in. If users have access to the sources for the client, there would otherwise be nothing to keep them from altering the timestamping routine. Another idea is to have the client send normally to what I term a "tower" (like a transmission tower) which would timestamp its input and send it off to the destination. Then you are left only with the problem of trusting the "tower"... but that shouldn't be nearly as big a hassle since you only need one of them per "geographical" area.... and that means there will be a lot less of them than there would be of clients.... in fact you could probably eliminate trust entirely by building binaries of the towers with the password built in for the major Unix boxes (sun, hp, SGI, dec, IBM). Perhaps there is an even better solution which I am overlooking.... or perhaps users aren't really concerned with the time delays -- to be sure it would require some work to even get going... and if there is lack of interest it probably would not be worth that effort. Anyway, food for thought! -Michael -- --------------------------------------------------------------------------- |Michael D. Moore | home (801) 585-4009 | email to mmo...@dsd.es.com| |Evans & Sutherland | work (801) 582-5860x4721 | **The heart never lies** | --------------------------------------------------------------------------- emusser Apr 10 1993, 7:10 pm show options Newsgroups: rec.games.chess From: emus...@mac.cc.macalstr.edu - Find messages by this author Date: 9 Apr 93 23:54:13 -0600 Local: Fri, Apr 9 1993 10:54 pm Subject: Re: ICS timestamping Reply to Author | Forward | Print | Individual Message | Show original | Report Abuse In article <1993Apr9.174540.13...@dsd.es.com>, mmo...@bambam.es.com (Michael Moore) writes: > I think most people would agree it would be nice to > have a timestamp on the moves so that net-lag would cease to > be as great a factor. > The problem with the idea is that you can't trust the > clients to be honest about that timestamping unless you create > a password to be sent with timestamp packets (encrypted of course) > and only distribute binaries with this password compiled in. If > users have access to the sources for the client, there would > otherwise be nothing to keep them from altering the timestamping > routine. I think the idea of timestamp is a good one, but you are overlooking one crucial point, ICS is meant for fun recreational chess. Having to encrypt stuff and distribute only binaries is putting the ICS at a much too serious level. If you (not specifically the writer) can't deal with losing because of lag I don't think you should really be playing at all. Perhaps an opposite point can be made though: maybe we can just distribute the source code with time stamping included and trust people not too alter it, and those who do change it are taking winning to seriously. The whole ICS system is based on trust. Who knows who actually is on the other side of that connection, maybe a computer or maybe even a group of people. What I think should be done is to have the people who run the internet system eliminate the damn lag :-). To further the discussion of unfairness and cheating on ICS I would like to mention the clients (I being a maker of one of them). We could put in alot of code that tells you when you are making a really dumb move, or code that automatically calls your opponents flag (I understand xics or xboard actually does do this), or maybe even suggest a move. We could do these things but we realize it is totally unfair and very unlike real chess. I suggest the flag code of the X-version client be taken out since most users cannot do this (certainly not ascii users). I have seen it used many times as an excuse to why someone called a lagged players time. The client should just be a extension to the server to make chess playing as natural and as real as possible. It should not be made to make a player better than that player's actual ability. Eric Michael Moore Apr 10 1993, 8:35 pm show options Newsgroups: rec.games.chess From: mmo...@bambam.es.com (Michael Moore) - Find messages by this author Date: Sat, 10 Apr 93 16:52:09 GMT Local: Sat, Apr 10 1993 9:52 am Subject: Re: ICS timestamping Reply to Author | Forward | Print | Individual Message | Show original | Report Abuse In article <1993Apr9.235413.4...@mac.cc.macalstr.edu> emus...@mac.cc.macalstr.edu writes: > I think the idea of timestamp is a good one, but you are >overlooking one crucial point, ICS is meant for fun recreational >chess. Having to encrypt stuff and distribute only binaries is putting >the ICS at a much too serious level. If you (not specifically the writer) >can't deal with losing because of lag I don't think you should really be >playing at all. Perhaps an opposite point can be made though: maybe >we can just distribute the source code with time stamping included and >trust people not too alter it, and those who do change it are taking >winning to seriously. The whole ICS system is based on trust. >Who knows who actually is on the other side of that connection, maybe >a computer or maybe even a group of people. What I think should be done is >to have the people who run the internet system eliminate the damn lag :-). You have a good point. I would still suggest timestamping be implemented regardless because it only furthers the fun of ICS. I can guarantee you that when someone playing from Finland loses simply because of net-lag he is not going to be having as much fun as he should. On the other hand it might ruin the fun if people started suddenly having 20 second lags at crucial moments. That's why I suggested the "towers" because they can't be cheated. Assuming everyone plays "honestly", timestamping is good. And while I also believe much of ICS is based on trust, I also see no reason to encourage cheating. Ideally, one would remove as many opportunities to cheat as possible. Regardless, if no one really cares, then it's probably pointless to worry about it. - Michael -- --------------------------------------------------------------------------- |Michael D. Moore | home (801) 585-4009 | email to mmo...@dsd.es.com| |Evans & Sutherland | work (801) 582-5860x4721 | **The heart never lies ** | --------------------------------------------------------------------------- Matthew Hastings Apr 12 1993, 9:11 am show options Newsgroups: rec.games.chess From: hastings-matt...@yale.edu (Matthew Hastings) - Find messages by this author Date: 12 Apr 1993 12:07:25 -0400 Local: Mon, Apr 12 1993 9:07 am Subject: Re: ICS timestamping Reply to Author | Forward | Print | Individual Message | Show original | Report Abuse - Hide quoted text - - Show quoted text - In article <1993Apr9.235413.4...@mac.cc.macalstr.edu> emus...@mac.cc.macalstr.edu writes: > I think the idea of timestamp is a good one, but you are >overlooking one crucial point, ICS is meant for fun recreational >chess. Having to encrypt stuff and distribute only binaries is putting >the ICS at a much too serious level. If you (not specifically the writer) >can't deal with losing because of lag I don't think you should really be >playing at all. Perhaps an opposite point can be made though: maybe >we can just distribute the source code with time stamping included and >trust people not too alter it, and those who do change it are taking >winning to seriously. The whole ICS system is based on trust. >Who knows who actually is on the other side of that connection, maybe >a computer or maybe even a group of people. What I think should be done is >to have the people who run the internet system eliminate the damn lag :-). > To further the discussion of unfairness and cheating on ICS >I would like to mention the clients (I being a maker of one of them). >We could put in alot of code that tells you when you are making a really >dumb move, or code that automatically calls your opponents flag (I understand >xics or xboard actually does do this), or maybe even suggest a move. >We could do these things but we realize it is totally unfair and very >unlike real chess. I suggest the flag code of the X-version client be taken >out since most users cannot do this (certainly not ascii users). I Good idea! Actually, I didn't know this was part of the clients, but, I agree-it should be taken out. >have seen it used many times as an excuse to why someone called a lagged >players time. The client should just be a extension to the server to make >chess playing as natural and as real as possible. It should not be made >to make a player better than that player's actual ability. > Eric Another problem with time-stamping is that some of us don't use clients at all. I usually use NCSA Telnet on a Mac, connecting to valkyries directly, not going through any Unix machine here. If people in favor of a timestamp would like to also write a Mac client, that could use an Appletalk network... It's an interesting idea, but IMO a bad one-it makes the whole thing too complicated, and it isn't really the point of ICS. matulovic Michael Moore Apr 12 1993, 10:51 am show options Newsgroups: rec.games.chess From: mmo...@bambam.es.com (Michael Moore) - Find messages by this author Date: Mon, 12 Apr 93 17:20:13 GMT Local: Mon, Apr 12 1993 10:20 am Subject: Re: ICS timestamping Reply to Author | Forward | Print | Individual Message | Show original | Report Abuse In article <1qc43tINN...@MINERVA.CIS.YALE.EDU> hastings-matt...@yale.edu (Matthew Hastings) writes: >Another problem with time-stamping is that some of us don't use clients at >all. I usually use NCSA Telnet on a Mac, connecting to valkyries directly, not >going through any Unix machine here. If people in favor of a timestamp would >like to also write a Mac client, that could use an Appletalk network... >It's an interesting idea, but IMO a bad one-it makes the whole thing too >complicated, and it isn't really the point of ICS. >matulovic Well, not to keep dragging this on and on, but if you implemented the "tower" idea this would not be a problem. Whether you are using an X client, an IBM one, or just telnet, you still have to connect to valkyries somehow. All we do is change the place you're connecting to to somewhere a little closer. That machine timestamps what it receives and sends it on to the actual server. I think it would work out real well, especially for those not in the continental U.S. But it does require work, and if no one cares then I guess we can say "Why bother". As for the point of ICS, last I heard it was to play fun chess. If you can provide for more honesty in the times for games, without opening things up for cheating, and without penalizing those people who don't use a special "client", then why not do it? It will only increase the pleasure of playing chess when you don't have to worry about a nasty lag ruining your game... Heh.... sorry... I just like this idea... :) - Michael -- --------------------------------------------------------------------------- |Michael D. Moore | home (801) 585-4009 | email to mmo...@dsd.es.com| |Evans & Sutherland | work (801) 582-5860x4721 | **The heart never lies ** | --------------------------------------------------------------------------- stein Apr 12 1993, 7:53 pm show options Newsgroups: rec.games.chess From: s...@tellabs.com (stein) - Find messages by this author Date: Mon, 12 Apr 1993 23:15:41 GMT Local: Mon, Apr 12 1993 4:15 pm Subject: Re: ICS timestamping Reply to Author | Forward | Print | Individual Message | Show original | Report Abuse - Hide quoted text - - Show quoted text - In article <1993Apr12.172013.15...@dsd.es.com> mmo...@bambam.es.com (Michael Moore) writes: >In article <1qc43tINN...@MINERVA.CIS.YALE.EDU> hastings-matt...@yale.edu (Matthew Hastings) writes: >>Another problem with time-stamping is that some of us don't use clients at >>all. I usually use NCSA Telnet on a Mac, connecting to valkyries directly, not >>going through any Unix machine here. If people in favor of a timestamp would >>like to also write a Mac client, that could use an Appletalk network... >>It's an interesting idea, but IMO a bad one-it makes the whole thing too >>complicated, and it isn't really the point of ICS. >>matulovic > Well, not to keep dragging this on and on, but if you implemented >the "tower" idea this would not be a problem. Whether you are using >an X client, an IBM one, or just telnet, you still have to connect to >valkyries somehow. All we do is change the place you're connecting to >to somewhere a little closer. That machine timestamps what it receives ^^^^^^^^^^^^^^^^^^^^^^^^^ How would this address lag effectively? A user's connection to ICS could be quite lengthy covering multiple private networks and the public switched network. If the purpose of time stamping is to address impediments beyond the user's control then this suggestion of "somewhere a little closer" may miss the mark. I know my connections to ICS are quite unreliable, with lag possibly created anywhere in my circuit path. Moving the source of timestamps further towards valkyries doesn't strike me as very helpful. Ultimately, (however difficult this may be, and I don't think the security issues are insurmountable) upon further analysis we may discover that timestamps only have meaningful value when they are produced at the local host. As annoying as the lag may be, I do have points of internet access. How I get to those points is my problem. Don't get me wrong - I think time stamping would be great if properly implemented. However, each user does have some degree of control over their end-to-end ICS connection. The "tower" idea needs to be studied closely. After all, we wouldn't want to embark on a solution that might only eliminate lag for 10% of the users when an alternative might eliminate lag for 90%. Actually, although I've lost a fair number of games on account of lag, I'm sure I've won some on account of it as well. I wonder if it doesn't just come out in the wash? I mean we're not playing for money, right? - Hide quoted text - - Show quoted text - >and sends it on to the actual server. > I think it would work out real well, especially for those >not in the continental U.S. But it does require work, and if no one >cares then I guess we can say "Why bother". > As for the point of ICS, last I heard it was to play fun chess. >If you can provide for more honesty in the times for games, without opening >things up for cheating, and without penalizing those people who don't use >a special "client", then why not do it? It will only increase the pleasure >of playing chess when you don't have to worry about a nasty lag ruining >your game... > Heh.... sorry... I just like this idea... :) > -Michael >-- > --------------------------------------------------------------------------- >|Michael D. Moore | home (801) 585-4009 | email to mmo...@dsd.es.com| >|Evans & Sutherland | work (801) 582-5860x4721 | **The heart never lies** | > --------------------------------------------------------------------------- emusser Apr 13 1993, 4:17 am show options Newsgroups: rec.games.chess From: emus...@mac.cc.macalstr.edu - Find messages by this author Date: 13 Apr 93 00:03:28 -0600 Local: Mon, Apr 12 1993 11:03 pm Subject: Re: ICS timestamping Reply to Author | Forward | Print | Individual Message | Show original | Report Abuse In article <1993Apr12.172013.15...@dsd.es.com>, mmo...@bambam.es.com (Michael Moore) writes: > In article <1qc43tINN...@MINERVA.CIS.YALE.EDU> hastings-matt...@yale.edu (Matthew Hastings) writes: >>Another problem with time-stamping is that some of us don't use clients at >>all. I usually use NCSA Telnet on a Mac, connecting to valkyries directly, not >>going through any Unix machine here. If people in favor of a timestamp would >>like to also write a Mac client, that could use an Appletalk network... >>It's an interesting idea, but IMO a bad one-it makes the whole thing too >>complicated, and it isn't really the point of ICS. >>matulovic I am working on the Mac telnet problem now. The latest version of E-ICS has telnet support (if you would like a copy, please email me, its not on valkyries yet) but seems to die after awhile. Direct connections to ICS don't work at all but going through a host machine seems to work okay for an hour or so. > Well, not to keep dragging this on and on, but if you implemented > the "tower" idea this would not be a problem. Whether you are using > an X client, an IBM one, or just telnet, you still have to connect to > valkyries somehow. All we do is change the place you're connecting to > to somewhere a little closer. That machine timestamps what it receives > and sends it on to the actual server. > I think it would work out real well, especially for those > not in the continental U.S. But it does require work, and if no one > cares then I guess we can say "Why bother". There are two problems to the tower system: machines are needed to act as towers and the towers would probably add even more lag to the game. Other than that it was a good idea. Eric