This my idea. I like the way you're doing it now so my suggestion would be just to automate that same process. First, you could set a minimum number of posts before the page would be split. Once the minimum number of posts was achieved the script would start checking the number of new lines (\n) in the page. So let's say you set the minimum number of posts to 20 and the number of new lines to 150 and the page had <!-- msg: 19; lines: 163 -->
The next post would automatically trigger a new page.
What exactly is the problem with "the (1)(2)(3) systems"? The only problem I see with them is that sometimes a visitor will not spot the (1)(2)(3) list, and never reads anything but the first page. Or do you have any other objections? Personally, I'd find it annoying not to have some kind of reminder at the top that informs me that there indeed are other parts. In fact, in an ideal world, even the entry in the thread list would offer a choice to pick any subthread of a set directly.
If you use a system that flattens those subthreads into the main forum list as Key_Master suggests, then the main trick will be to autogenerate the "continued in..." and "continuation from..." posts at the top and bottom of each, and possibly a "part x" title extension. The "contination from..." post could actually include the original starting post, and an automatically updated list of all other subthreads. Maybe the "reply to this thread" function should also display a notice when someone replies to one of the earlier subthreads, telling the poster that his reply will actually go to the end of the last one. Not knowing anything about the implementation, I assume you'll have to add some flag to each thread, so that the system can keep track if it is part of a sequence.
Ok, I won't code it, so I'll stop whipping up potential features now... ;)
>>What exactly is the problem with "the (1)(2)(3) systems"?<<
One of the problems I have with these systems is that you have to load the first part to get to the rest, so you may not really be saving any load time (assuming a dial-up connection, where time is a factor).
If an auto-split, I feel it would be nice to keep the parts as if separate threads on a particular forum, as we have here now, rather than the (1)(2)(3) arrangement. The nice thing about the manual splits we've had on this board is that a person can choose better break points than an auto-split. Of course this takes someone's time and effort... and Brett probably has better things to do. ;)
What I have decided on is the (1)(2)(3)... system. There are a couple of differences between my system and some of the other ones out there.
You will not have to load the first part of a thread to get to the last section. Additionally, the number of msgs before splitting will be user controllable up to 200 msgs per thread. I may roll that back to 100 in an effort to control bandwith.
>a person can choose better break points than an auto-split. Of course this takes someone's time and effort...
It's sometimes important, Robert. Even in the same thread, covering a particular topic, there's a natural breaking point where one logical sequence of thought transitions into another.
Example: the Google Update [webmasterworld.com] thread for February was looked at throughout the day and again at length to decide on where to split. Good reason for the first post in Part 2 being where it is - it's something many new people want to know time and time again, and there's the explanation for them, right where it's easy to see.
A lot of what's in a thread like that is time-relevant, but that information has enduring value for users, so the first post is where it serves the best function, rather than lost in the middle. Some do take a lot of thought and re-reading and it's not impossible that with some it's worth waiting until a logical point comes a little further along.
There are split points that would not be good, and I hope if that happens with automated splitting there will be a way to over-ride and change it.
>will be user controllable
I can't envision how that will work. Who gets to choose, and how much controllability is there?