So, this issue has apparently started brewing again, although people haven't been willing to say anything to my face about it. Long threads! They clog the server and cause errors! They should be locked and restarted! Negrek, why are you such a jerk, hanging around with all your long ASB threads that you blithely refuse to do anything about? How could you willingly detract from other people's forum experience like that?
That would be because I doubt "long" (greater than the thrillingly arbitrary 1000-post cutoff) threads actually have any significant impact on forum performance. And, in the interest of not having to repeat myself a ton, I'm just going to make a thread about it that people can froth over instead of addressing each person who brings it up individually.
How did the whole "long threads cause slowdown" thing get started, anyway? There was some nebulous muttering about it here, back when people were on a crusade against Forum Games. Somehow, "Well maybe long threads cause server overhead problems" went from being speculation to actual fact, and from actual fact to actually informing forum policy, without there being anything like proof brought up in the interim.
In fact, a simple Google search on "do long threads cause errors vBulletin" brings up no resounding "Yes! They're bad!" Responses tend to be mixed or outright negative, and the information available is contradictory and almost always hearsay. Ultimately, there is no consensus that long threads are more harmful than the same number of posts split up between multiple threads that indicates it actually has an effect that can be discerned from other factors that tend to come along with really long threads, i.e. lots of people online at once, lots of people viewing/editing a thread at once, and so on.
Unfortunately, Butterfree doesn't have error logging set up so she can see the number of 500 errors that occur, when, and what action was being performed at the time (I assume); otherwise, it would be fairly easy to benchmark this. We can therefore proceed only with indirect evidence. Most damning of all is the fact that periods of server fail come and go, while over time these threads have only been growing longer. The board was choking even before the Challenge Board hit 1000 posts. Then things died down for a while. Then, oh no! More server fail, and now the Challenge Board has more than 1000 posts! But then I didn't do anything about it, and the errors died down again. And now there are more errors, and the thread has even more posts! But, once again, no action is taken; the errors die down, and yet the thread has more posts in it than ever before. There is simply no correlation between the length of these large ASB threads and server performance.
Not only that, but even way back in the beginning, when forum games were the scapegoat, deleting them didn't actually stop the board problems. They died down a little while afterwards, yes, but the effect was not immediate, and post hoc, ergo propter hoc. Even moving to a new server, which should have led to more resources and therefore less of a problem for long threads, did not appreciably reduce spates of server fail.
So, there remains no evidence that I can see that long threads cause server traffic problems or errors. If you can provide some, I will listen to you and become a supporter of the "kill long threads" policy. If you can't, I will continue to ignore you, especially if you'd rather stew quietly than confront me about the issue.
Note that "but I always get more errors when I view the ASB Challenge Board!" is not a valid response for three reasons: a) confirmation bias b) anecdata c) lack of proper controls (could you be getting more errors there simply because you view it more than other threads?). If you can show that this is actually the case, i.e. with numbers in a controlled experiment, then it would be acceptable.
I dislike server fail as much as anybody else. If I thought locking long threads would actually do something to alleviate it, I would suck it up and deal with the inconvenience and just do so. However, I refuse to go out of my way to comply with a policy that has no sound basis in fact.
tl;dr if you want me to actually go to the trouble of getting rid of long threads, provide some actual evidence that they're bad for the board.
That would be because I doubt "long" (greater than the thrillingly arbitrary 1000-post cutoff) threads actually have any significant impact on forum performance. And, in the interest of not having to repeat myself a ton, I'm just going to make a thread about it that people can froth over instead of addressing each person who brings it up individually.
How did the whole "long threads cause slowdown" thing get started, anyway? There was some nebulous muttering about it here, back when people were on a crusade against Forum Games. Somehow, "Well maybe long threads cause server overhead problems" went from being speculation to actual fact, and from actual fact to actually informing forum policy, without there being anything like proof brought up in the interim.
In fact, a simple Google search on "do long threads cause errors vBulletin" brings up no resounding "Yes! They're bad!" Responses tend to be mixed or outright negative, and the information available is contradictory and almost always hearsay. Ultimately, there is no consensus that long threads are more harmful than the same number of posts split up between multiple threads that indicates it actually has an effect that can be discerned from other factors that tend to come along with really long threads, i.e. lots of people online at once, lots of people viewing/editing a thread at once, and so on.
Unfortunately, Butterfree doesn't have error logging set up so she can see the number of 500 errors that occur, when, and what action was being performed at the time (I assume); otherwise, it would be fairly easy to benchmark this. We can therefore proceed only with indirect evidence. Most damning of all is the fact that periods of server fail come and go, while over time these threads have only been growing longer. The board was choking even before the Challenge Board hit 1000 posts. Then things died down for a while. Then, oh no! More server fail, and now the Challenge Board has more than 1000 posts! But then I didn't do anything about it, and the errors died down again. And now there are more errors, and the thread has even more posts! But, once again, no action is taken; the errors die down, and yet the thread has more posts in it than ever before. There is simply no correlation between the length of these large ASB threads and server performance.
Not only that, but even way back in the beginning, when forum games were the scapegoat, deleting them didn't actually stop the board problems. They died down a little while afterwards, yes, but the effect was not immediate, and post hoc, ergo propter hoc. Even moving to a new server, which should have led to more resources and therefore less of a problem for long threads, did not appreciably reduce spates of server fail.
So, there remains no evidence that I can see that long threads cause server traffic problems or errors. If you can provide some, I will listen to you and become a supporter of the "kill long threads" policy. If you can't, I will continue to ignore you, especially if you'd rather stew quietly than confront me about the issue.
Note that "but I always get more errors when I view the ASB Challenge Board!" is not a valid response for three reasons: a) confirmation bias b) anecdata c) lack of proper controls (could you be getting more errors there simply because you view it more than other threads?). If you can show that this is actually the case, i.e. with numbers in a controlled experiment, then it would be acceptable.
I dislike server fail as much as anybody else. If I thought locking long threads would actually do something to alleviate it, I would suck it up and deal with the inconvenience and just do so. However, I refuse to go out of my way to comply with a policy that has no sound basis in fact.
tl;dr if you want me to actually go to the trouble of getting rid of long threads, provide some actual evidence that they're bad for the board.