View Full Version : Any possibility the game can be updated to auto-switch hosts if they quit?
Haoshiro
16 Mar 2007, 11:36
There are many times game hosts will quite the game while it is still in progress. Maybe they lost, maybe they think they will, etc... but the result is the same, everyone gets kicked out when the host leaves. Often times I don't even think the host realizes he is killing the game for everyone else.
Any chance you (as in, Team17) will update the game to automatically switch hosts if a host leaves? Even if this requires the game to be pause during the switch, I'm sure lots of players would appreciate the feature.
If that's just not at all feasible, restricting the host from leaving could be another option. Perhaps if they choose quit/exit they could be given a message that it isn't allowed at this time, and tell them why. That wouldn't prevent dash exits, but it still might help if hosts can't be switched.
Switching hosts (propagating) is not supported and is quite a task I'm told, so it's unlikely to be developed, sorry. The game is peer to peer and not ad-hoc like FPS titles. I've had relatively few issues on this front and if you want to ensure that the game finishes, host a game yourself :)
quakerworm
16 Mar 2007, 18:09
The game is peer to peer and not ad-hoc like FPS titles.
never heard of an fps with an ad-hoc connection. they usually require a server, and that means client-server connection. pure p2p networks do not require a host. if a node drops out, the network can continue. that's why most (all?) ad-hoc connections are implemented as a type of p2p. so i'm really not sure what you are talking about.
from everything that parsley had said about the network, i think that switching the hosts could have been relatively easy. it just would have been a pain to test it properly. i'm sure that t17 simply did not anticipate the kind of network they ended up dealing with, resulting in host-switching being low on priority list. now, it would require a major overhaul of the networking, and it's probably just not a viable option in terms of costs.
if you want a real answer, i'd wait until parsley gets here (if he does) and actually explains it.
I meant that it wasn't a typical drop-in, drop-out server model (like those in FPS). I don't believe it's straight forward to handle host migration and I don't believe it'll be updated to support it.
You arguing about how easy it is, how you'd do it, how it should be done or whatever are quite moot.
quakerworm
16 Mar 2007, 19:18
depends on what you mean by host quitting in fps. if you mean that the server remains, as in the case of a dedicated server, then yeah, it's easy. but if the server stops, you can't migrate. server holds all the essential game data that clients simply have no access to. since most of the time the host is running the server, host quitting stops the server, and there is nothing you can do about it.
with worms it is easy in the sense that the network model is the kind that you want if you want to change hosts in the middle of the game. if the host quits, each node still has all the info needed to continue the game. it's just a matter of re-establishing a host. that might not be easy in itself, but it's not much more complex than hosting the game in the first place.
thomasp
16 Mar 2007, 19:46
with worms it is easy in the sense that the network model is the kind that you want if you want to change hosts in the middle of the game. if the host quits, each node still has all the info needed to continue the game. it's just a matter of re-establishing a host. that might not be easy in itself, but it's not much more complex than hosting the game in the first place.
Simple answer to this: OK then, you code/sort it :p
Somehow, I think Team17 are better judges of what is easy and what isn't easy when it comes to worms coding.
quakerworm
16 Mar 2007, 21:22
Simple answer to this: OK then, you code/sort it :p
sure. just send me the 360 developer package and the source codes.
worMatty
16 Mar 2007, 21:36
He asked for that
sure. just send me the 360 developer package and the source codes.
And you've 24-48 hours to turn it around and test it so that it won't break when 100,000 games a day are going through it, under a variety of network instances, under all the regulations that MS stipulate for their services. Suffice to say, it's not going to happen.
Haoshiro
16 Mar 2007, 22:54
Well I'd be interested in what parsley has to say about it, but seems pretty obvious this will be a no-go. Thanks for the reply though!
Simple answer seems to be that this would have needed to be a feature planned and designed at the very beginning, and would be a huge and nearly impossible task at this stage. I figured as much, but thought I'd ask anyway!
I have been doing most of the hosting now, didn't realize my NAT was strict and wasn't able to host without getting the proper ports forwarded. It's all good now! :)
I have been doing most of the hosting now, didn't realize my NAT was strict and wasn't able to host without getting the proper ports forwarded. It's all good now! :)
Good to hear you've got it sorted out. Have fun! ;)
quakerworm
16 Mar 2007, 23:13
i could do it in 48 hours. i know what i need to change. i know where to find it. i do not need to modify much in the way the network behaves, so it will not create any problems with ms standards. i'd only need to make sure not to introduce new bugs, and that i can test for. obviously i cannot guarantee that it will be bug-free any better than anyone else, but it will be ready for certification process.
the problem is that i can say all that, and there is no way i can demonstrate it. i highly doubt that you would allow me access to the source, much less provide me with the development tools.
Haoshiro
16 Mar 2007, 23:32
i could do it in 48 hours. i know what i need to change. i know where to find it. i do not need to modify much in the way the network behaves, so it will not create any problems with ms standards. i'd only need to make sure not to introduce new bugs, and that i can test for. obviously i cannot guarantee that it will be bug-free any better than anyone else, but it will be ready for certification process.
the problem is that i can say all that, and there is no way i can demonstrate it. i highly doubt that you would allow me access to the source, much less provide me with the development tools.
And no offense, that is exactly why you can say all that.
Considering you haven't ever seen the any of the code and have no idea what state it is in or how it was designed, I'd have to bet against you on this one. It could take more then 24 hours just come to grips with the way networking is handled and intergrated into the game. Now your statement is very believable if the code was all your from the start. ;)
worMatty
16 Mar 2007, 23:50
Wow, quaker, you're certainly ambitious. However, to reiterate Hao's point, without knowing the exact details you cannot guarantee you could do it in two days. You probably think you're capable as you think you have the knowledge needed to program anything you want but you don't have first-hand knowledge of the protocols set in place by Microsoft, or the requirements of the development process. I admire your confidence in your abilities but, rather than accuse Team17 of being sloppy, you would be more impressive if you trust what they've told us and inquire privately about working with them, to convey your true wish to help improve the game, in a polite manner.
He reminds me a bit of K^2.
He reminds me a bit of K^2.
You're not the only one to say that.
Even Halo 2 had this, when the host quit it found a new server within the party.
quakerworm
17 Mar 2007, 00:13
Considering you haven't ever seen the any of the code
18 hours to study code, 6 hours of sleep, 12 hours to write new code, 12 hours to finish debugging. it can be done.
worMatty
17 Mar 2007, 00:15
Aren't you supposed to be off curing cancer in India or something? I swear I saw you on Channel 5 a few weeks ago.
Can't we just accept the fact that if the host quits it's because he's losing and he's afraid to lie down and take it?
Haoshiro
17 Mar 2007, 00:20
Aren't you supposed to be off curing cancer in India or something? I swear I saw you on Channel 5 a few weeks ago.
Can't we just accept the fact that if the host quits it's because he's losing and he's afraid to lie down and take it?
I think it's mostly unfair in Ranked matches, whoever would have won doesn't get the chance to, so they've wasted their time and the rankings still consider it a game played...
worMatty
17 Mar 2007, 00:24
Do they? Since the end of the game never came about and the results weren't transmitted to the server?
Team17 cannot fix idiots. A disconnection penalty would be sublime but I'm not sure if that's how it is at the moment. And if any host disconnects because of internet connection problems then that's more incentive to stop playing the game, get a weekend job and pay for a better ISP. Everyone wins.
Haoshiro
17 Mar 2007, 02:43
Do they? Since the end of the game never came about and the results weren't transmitted to the server?
Team17 cannot fix idiots. A disconnection penalty would be sublime but I'm not sure if that's how it is at the moment. And if any host disconnects because of internet connection problems then that's more incentive to stop playing the game, get a weekend job and pay for a better ISP. Everyone wins.
Well it seemed like it did, of course I don't have direct access to check so I can't be 100% sure.
It would actually be nice to have more details on what effects ranking, as far as when a game gets marked as played, etc.
I think hosts do get a penalty for quiting, although I'm not sure they are warned of this first. Again, not 100% sure of any of this.
Wormetti
17 Mar 2007, 03:14
If you try to quit, whether you are the host or client, it warns that it MAY effect your trueskill rating. I recommend adding a warning about the game ending for everyone if the hosts quits since not all hosts know that.
thomasp
17 Mar 2007, 11:42
You're not the only one to say that.
Run an IP check then ;)
SirMossy
18 Mar 2007, 19:08
lol Quaker should just post the code if he's so confident and Team17 can cut n paste.
quakerworm
18 Mar 2007, 20:57
a) need sources. b) need to be able to test the code.
18 hours to study code, 6 hours of sleep, 12 hours to write new code, 12 hours to finish debugging. it can be done.
1: Considering that you have absolutely no experience in programming a game for an xbox360; somehow, I rather doubt that you can both learn enough code so that you can not only reprogram the game, but also be able to spot every part that needs to be changed, in only 18 hours.
2: There's no way that you're be able to go 24 hours with only 6 hours of sleep and still be able to work efficiently!
3: There's no way, whatsoever, that you'd be able to debug the entire thing, singlehandedly (which is very hard to do on a multiplayer game) in 12 hours! And that's not even getting close to the fact that you'd need at least 3 Xboxes and 3 broadband connections to do so!
quakerworm
19 Mar 2007, 03:53
Considering that you have absolutely no experience in programming a game for an xbox360
and you seem to have absolutely no experience in programming. no offense. a c++ code is a c++ code. but even sans compiler, i am familiar with ppc assembly, so that wouldn't matter. what matters are the libraries, and i don't need to touch them. there is already net code in place. i don't need to write my own code for communication between the systems. it's already there. all i need is to modify some addresses, perhaps close and open some connections. all done with functions already written by t17. anything else, i can fake on the spot.
There's no way that you're be able to go 24 hours with only 6 hours of sleep and still be able to work efficiently!
48 hours, and how do you think i made it through college?
There's no way, whatsoever, that you'd be able to debug the entire thing, singlehandedly in 12 hours!
i don't need to debug the entire thing. only the new code. since i'm not writing the net backbone, there aren't nearly as many things that can go horribly wrong. and even if host migration fails often, as long as it doesn't cause a hard lockup, it's an improvement. so basically, i need to make sure that a) my code doesn't leak, and b) that i call pre-existing net functions in the same manner as they are already being used. that way, the odds of new bugs being introduced are absolutely minimal, which is all that anyone can do. the rest is up to the certification process.
Babaloo2U
19 Mar 2007, 05:36
48 hours, and how do you think i made it through college?
Agreed. BAS Computer Science was a pain in the pooper but it was worth it in the end. ;) I miss my 48 hour days. Lol.
parsley
19 Mar 2007, 11:23
quakerworm: please stop. You do not know anything about Worms. You do not appear to know anything about Xbox Live. I am not going to comment on anything you have written (because that would be cruel), but your baseless boasting is beginning to grate. It is also painfully wrong. So, please. Stop.
The original question was answered in the second post. Thread closed.
vBulletin® v3.8.6, Copyright ©2000-2013, Jelsoft Enterprises Ltd.