This was posted elsewhere...

Hi KLiK,
This bug was fixed in the RC4 release in 2005. Unfortunately, we are using the RC3 release of the mvnForum software. We adapted the forum software to run on 2 servers, which made it difficult to upgrade (and broke some features). Looking over the release notes on the mvnForum website, it looks as though the most recent release has finally implemented multi-server support using 'whirly-caches'. However, we have so much to do that it seems unlikely that we will be able to change the forum software. Maybe we will find some free time next year, by which time there will probably be a new release for us to look over.

Lawrence

I used to be consultant for interactive (CICS) mainframe systems and have a fair amount of PC experience - 20+ years more or less. I don't understand why installing new forum software is such a major production. What would be the problem with creating a couple of virtual machines on one of the servers and loading the software for an in-house trial. Just give the people at WCG access to it and see how it runs. If it seems to work, you could then run it in parallel with this software on a beta basis. When all the bugs are worked out, just throw the switch and redirect traffic to the new software.

This isn't rocket science guys.

And if you don't like the idea of running things in a virtual machine, I sure IBM has a couple extra server blades lying around that they could spare.

You don't have to dedicate anyone to it full time, just make the software available to the people on staff who want to play with it. Anybody who is really into the technology will jump at the opportunity, even if they have to invest some of their own time. But if you never give them the opportunity, well . . .

Granted this software works, but it is antiquated and doesn't have even some the basic features that people expect. When people are shopping around for a dc project to work on, a lot of them are going to look at the community behind it. If they see that your running old software that is 2 or 3 whole number versions behind what is available, they might think twice about joining.

Finally, if other groups can manage to upgrade their software, what's WCG's problem? Most of these other groups are non-profit too and run mainly on the sweat of graduate students. So go down to the slave market and buy a couple grad students. As long as you feed them pizza and beer and give them a nice futon to sleep on, they'll do whatever you want.

Not questioning anyone's dedication. Just saying, make things available to the people who want to play with it - see what happens. No promises, no timetables, just throw it out there. If people don't have time, then nothing happens. If they do or need a change of pace, at least they have the option. Even if nothing comes of it, when WCG IS ready to upgrade, you already have people on staff who know what they're getting into. That alone is a big leg up for any migration.

You're right about the numbers game - forum regulars versus grid volunteers. But it's just like any other entity that has to deal with the public - a lot depends on appearance. Let's say you have to choose between 2 online stores for a purchase. You don't know anything about either one except that their both considered to be reputable. One has buggy software that makes it difficult to find what you want and to place your order. The other one has a slick web site which makes it easy to browse and which has other features like wish lists, user reviews, one-click purchasing, etc. Which one are you going to buy from?

There may only be a couple hundred people who post regularly, but whenever I look to see who is on line, there are often more "guests" than logged in users. The point is, appearances count for a lot and in the long run, any investment you make in providing a clean, sophisticated appearance will be of substantial benefit.

Finally, is it really necessary to sync the forum database with the other 2? I'm not saying don't authenticate users, but if the forum database isn't always in sync with the others, would that really be so bad? If it would make the transition easier, it's something to consider.

It's more the "logged in" thing - when logged into the website, you're also logged into the forum. You may have noticed a few glitches with that feature....

And I rather suspect that WCG are not able to release all the back-end website code because it is owned by United Devices.

From my point of view, the sooner they rip that cruft out, the better!

And the forums do work, remember. It's not unusable. Most casual users are unlikely to spot a problem, particularly if they are only reading, not posting. And many of the "guests" will be forum regulars who are temporarily logged out....

If one of the 'problems' is the 'linking' of the forum with the operational side of crunching, such that if you are logged into one then you are logged into the other....then why not simplify and 'ditch' that feature?

To my way of thinking if the crunch side goes down and takes the forum with it, then users cannot log into the forum to see what is happening, which you could do if they were seperated.

Having forums that do not rely on another part of WCG would mean that they can be updated easily and tested easily without also having to test that there has been no impact on the 'crunching' side....and vice versa!