Taluntain
Sun, 11th Nov '07, 11:32pm
As most of you probably know, the message board software switch we've gone through recently has been planned for a long time. In fact, it's now close to 5 years since I've started looking for new message boards software to switch to that could better handle the size that our message board community has been back then.
At the time, the switch would not have been that easy, and it would also have meant losing a part of our content (or having it imported broken beyond repair). The importer to vBulletin (the message board software we've now switched to) at the time was very buggy, and I've read enough reports of botched imports to dissuade me from doing it until the importer was patched up.
In the meantime, I've upgraded the server several times, so it was able to cope with our ever-growing message boards better, and I've also upgraded the old message board software and applied a score of modifications to it to keep it reasonably up to date. For a time (several years, in fact), this was good enough. During this time, I've been checking up on the progress of the vB importer regularly and lobbied for inclusion of certain features missing from vB that would have made our import easier and more complete. Some of those suggestions have been implemented; some haven't to date, despite assurances to the contrary years ago.
The largest obstacle on the road to the vB switch in the last 3 years has been the inability to import the overwhelming majority of the polls made in our old message boards software. Ironically, the polling system on our old boards to date remains significantly superior to that of vBulletin (and pretty much every other message board software I know of), whereas vB's at the time was basic, to say the least. It hasn't changed much by today either, regrettably.
Consequently, conducting an import would have resulted in hundreds of broken polls all across the boards, and an inevitable loss of a part of the board's content and history. Call me sentimental, but that simply was not acceptable to me.
So I have waited and waited in the hopes that we'd eventually be able to switch over without the loss of our polls, but it became clear last year that given the state of vB's polling system, that would simply not be possible. Also in 2006, Infopop, the makers of our old message board software, have stopped supporting UBB.classic completely (development itself has stopped even further back), to the point of removing all mention and manuals for the software from their site. Imagine my joy at spending hours tracking down manuals to the software I have paid for to be able to fix a few issues and transfer our boards to a new server.
At that point it became clear that like it or not, I had to move our community to a new platform, because in case of an emergency or a problem with the old software that I could not handle myself, I would have no one to turn to for help.
Still unhappy with the prospect of losing over 1200 polls on the boards, I have started thinking of ways to get around the fact that the polls couldn't be imported into vB in a working state. I came to the conclusion that writing the poll results out and saving them into the poll threads themselves would have been a good solution that would at least preserve all of the legacy poll content. With that in mind, I have hired a web developer who has come up with a perfect working solution based on my concept. A little too perfect, as we would find out...
We have had a final version of the converter ready more than a year ago, and all the testing on the old boards has shown that everything converted beautifully. All the polls' results would simply show up in the poll threads themselves instead of in a pop-up when clicking on the "Results" button, and would be written in the poll threads as unchangeable archives. On UBB's side of things, this has worked without a hitch, so we've assumed that it would work fine once we imported such poll threads into vB later on.
As (bad) luck would have it, when I was starting to get ready for the switch over a year ago, a whole slew of misfortunes and personal issues have zoomed in on me in real life, and it's taken a good part of the year to get it all sorted out. During this time, the possibility of my committing to a long-term project that I knew our message boards switch would be was quite literally zero. So I put it all off until a more suitable time...
...which came about 4 months back. I started to catch up on everything regarding the board switch again (a great deal changes in a year in software that is actively being developed), and set up a vB board for testing. It took a few dozen hours of trial imports and information gathering for me to fully understand every aspect of the importer application that would bridge the old and the new message boards, and to get a first complete trial import done.
This first import has exposed literally dozens of issues that needed to be fixed prior to the final import (some would be impractical or impossible to fix afterwards), one of which was the fact that our neat solution to the problem of carrying the polls over didn't really work. Or rather, it did, but HTML had to be enabled board-wise for the poll results to show properly and not as chunks of code. On UBB's end this was not an issue because we would write the poll results directly to the thread files, bypassing UBB's "HTML off" setting. UBB wouldn't "know" that it was displaying HTML from the poll thread files, so it's simply shown it all as if HTML was turned on board-wise. However, HTML would import into vB as actual HTML, and the only practical way to show it properly on vB was with the HTML setting set to "on", which, for security reasons, is not a sensible option.
So, all the hours and custom coding spent on making the poll results look good and be easily customisable went right through the window. In the end, we've had to resort to plain text poll archives with simple board formatting, meaning we've had to strip off all the fancy poll layout design from the poll results. Resulting, of course, in yet more work for less of a result. That's Murphy's for you...
Once this was finally sorted out, it was "simply" a matter of importing, testing, fixing, and repeating. About 6 times altogether, with each process taking about 8 hours altogether. Why so many times? Because every testing revealed new "bugs", issues and problems, and because it takes a while to go through 7 years' worth of posts to make sure that everything in them that could or did break during the imports is properly accounted for and cleaned up after the import.
As you can probably gather from this long-winded documentation, the fact that our final import and the switch as a whole has gone so smoothly is not due to the fact that it was an easy or quick process, but due to literally months of prior preparation to make sure that we don't encounter any nasty surprises in the end. And, to be honest, what I've written here isn't even the half of it, but I don't want to bore everyone to tears, so I'll try to get to the point now. :D
My reason for writing all this is to explain why I'm putting out a call for support (the first one in well over a year). Besides the huge time toll that this project has taken on me (and will continue to for quite some time, since we're nowhere near finished yet), it has also been a substantial financial drain. The direct costs I've incurred were for the new board software itself, some extra software that will be implemented in the near future, and the cost of a hired coder to help with the polls and a few other issues. All in all, this amounts to about $400 spent on getting us to the new board platform. In my opinion, every cent of it was worth it, because what you're seeing on the new boards now is just the tip of the iceberg. Many features remain disabled for the moment because turning it all on at the same time would have been simply overwhelming (for the users, as well as your humble board management), so there's much more to look forward to in the near future, and not only from the message boards alone.
A few new perks for the SPS account holders have already been implemented, and several more are still on the way. The new board software will allow me to integrate many site features better and more completely than in the past, so it is not "just" a new board - it is the basis for most of the future site development and integration - all long overdue.
You can contribute to Sorcerer's Place in many ways; by helping out with the running costs via getting an SPS account (http://www.sorcerers.net/Supporters/index.php), some other means (http://www.sorcerers.net/Main/helping_sp.php), or simply by visiting regularly and inviting your friends to join you in our community.
Anything that you can do to help will be greatly appreciated, so thank you in advance!
At the time, the switch would not have been that easy, and it would also have meant losing a part of our content (or having it imported broken beyond repair). The importer to vBulletin (the message board software we've now switched to) at the time was very buggy, and I've read enough reports of botched imports to dissuade me from doing it until the importer was patched up.
In the meantime, I've upgraded the server several times, so it was able to cope with our ever-growing message boards better, and I've also upgraded the old message board software and applied a score of modifications to it to keep it reasonably up to date. For a time (several years, in fact), this was good enough. During this time, I've been checking up on the progress of the vB importer regularly and lobbied for inclusion of certain features missing from vB that would have made our import easier and more complete. Some of those suggestions have been implemented; some haven't to date, despite assurances to the contrary years ago.
The largest obstacle on the road to the vB switch in the last 3 years has been the inability to import the overwhelming majority of the polls made in our old message boards software. Ironically, the polling system on our old boards to date remains significantly superior to that of vBulletin (and pretty much every other message board software I know of), whereas vB's at the time was basic, to say the least. It hasn't changed much by today either, regrettably.
Consequently, conducting an import would have resulted in hundreds of broken polls all across the boards, and an inevitable loss of a part of the board's content and history. Call me sentimental, but that simply was not acceptable to me.
So I have waited and waited in the hopes that we'd eventually be able to switch over without the loss of our polls, but it became clear last year that given the state of vB's polling system, that would simply not be possible. Also in 2006, Infopop, the makers of our old message board software, have stopped supporting UBB.classic completely (development itself has stopped even further back), to the point of removing all mention and manuals for the software from their site. Imagine my joy at spending hours tracking down manuals to the software I have paid for to be able to fix a few issues and transfer our boards to a new server.
At that point it became clear that like it or not, I had to move our community to a new platform, because in case of an emergency or a problem with the old software that I could not handle myself, I would have no one to turn to for help.
Still unhappy with the prospect of losing over 1200 polls on the boards, I have started thinking of ways to get around the fact that the polls couldn't be imported into vB in a working state. I came to the conclusion that writing the poll results out and saving them into the poll threads themselves would have been a good solution that would at least preserve all of the legacy poll content. With that in mind, I have hired a web developer who has come up with a perfect working solution based on my concept. A little too perfect, as we would find out...
We have had a final version of the converter ready more than a year ago, and all the testing on the old boards has shown that everything converted beautifully. All the polls' results would simply show up in the poll threads themselves instead of in a pop-up when clicking on the "Results" button, and would be written in the poll threads as unchangeable archives. On UBB's side of things, this has worked without a hitch, so we've assumed that it would work fine once we imported such poll threads into vB later on.
As (bad) luck would have it, when I was starting to get ready for the switch over a year ago, a whole slew of misfortunes and personal issues have zoomed in on me in real life, and it's taken a good part of the year to get it all sorted out. During this time, the possibility of my committing to a long-term project that I knew our message boards switch would be was quite literally zero. So I put it all off until a more suitable time...
...which came about 4 months back. I started to catch up on everything regarding the board switch again (a great deal changes in a year in software that is actively being developed), and set up a vB board for testing. It took a few dozen hours of trial imports and information gathering for me to fully understand every aspect of the importer application that would bridge the old and the new message boards, and to get a first complete trial import done.
This first import has exposed literally dozens of issues that needed to be fixed prior to the final import (some would be impractical or impossible to fix afterwards), one of which was the fact that our neat solution to the problem of carrying the polls over didn't really work. Or rather, it did, but HTML had to be enabled board-wise for the poll results to show properly and not as chunks of code. On UBB's end this was not an issue because we would write the poll results directly to the thread files, bypassing UBB's "HTML off" setting. UBB wouldn't "know" that it was displaying HTML from the poll thread files, so it's simply shown it all as if HTML was turned on board-wise. However, HTML would import into vB as actual HTML, and the only practical way to show it properly on vB was with the HTML setting set to "on", which, for security reasons, is not a sensible option.
So, all the hours and custom coding spent on making the poll results look good and be easily customisable went right through the window. In the end, we've had to resort to plain text poll archives with simple board formatting, meaning we've had to strip off all the fancy poll layout design from the poll results. Resulting, of course, in yet more work for less of a result. That's Murphy's for you...
Once this was finally sorted out, it was "simply" a matter of importing, testing, fixing, and repeating. About 6 times altogether, with each process taking about 8 hours altogether. Why so many times? Because every testing revealed new "bugs", issues and problems, and because it takes a while to go through 7 years' worth of posts to make sure that everything in them that could or did break during the imports is properly accounted for and cleaned up after the import.
As you can probably gather from this long-winded documentation, the fact that our final import and the switch as a whole has gone so smoothly is not due to the fact that it was an easy or quick process, but due to literally months of prior preparation to make sure that we don't encounter any nasty surprises in the end. And, to be honest, what I've written here isn't even the half of it, but I don't want to bore everyone to tears, so I'll try to get to the point now. :D
My reason for writing all this is to explain why I'm putting out a call for support (the first one in well over a year). Besides the huge time toll that this project has taken on me (and will continue to for quite some time, since we're nowhere near finished yet), it has also been a substantial financial drain. The direct costs I've incurred were for the new board software itself, some extra software that will be implemented in the near future, and the cost of a hired coder to help with the polls and a few other issues. All in all, this amounts to about $400 spent on getting us to the new board platform. In my opinion, every cent of it was worth it, because what you're seeing on the new boards now is just the tip of the iceberg. Many features remain disabled for the moment because turning it all on at the same time would have been simply overwhelming (for the users, as well as your humble board management), so there's much more to look forward to in the near future, and not only from the message boards alone.
A few new perks for the SPS account holders have already been implemented, and several more are still on the way. The new board software will allow me to integrate many site features better and more completely than in the past, so it is not "just" a new board - it is the basis for most of the future site development and integration - all long overdue.
You can contribute to Sorcerer's Place in many ways; by helping out with the running costs via getting an SPS account (http://www.sorcerers.net/Supporters/index.php), some other means (http://www.sorcerers.net/Main/helping_sp.php), or simply by visiting regularly and inviting your friends to join you in our community.
Anything that you can do to help will be greatly appreciated, so thank you in advance!