The new QuakeForge tree is an effort based on the QuakeWorld cooked up by a few developers who had all come to the conclusion that the current tree had some inherent limits we didn't like.
First, the build system. It's not very portable and lacks a key feature called header dependency. Without them you often times find yourself needing to recompile the entire project just to test a little change. I should point out that a compile takes about five to six minutes on my PIII and most of our developers do not have that sort of machine for development.
Second, there are tons of bugs in the source. Not just the ones we're responsible for - those are easy to fix! The ones buried deep in Id's code, often with two or three layers of workarounds on top of them to keep the problems from affecting most people in common situations. We have been adding to the workarounds because we didn't know the cause of the bug either. And then it hits one of us what the original problem was. Sometimes trying to fix these problems the right way in our current tree is very difficult.
Third, the merge. One of our goals from the beginning was to have a single binary which could act as a client, a server, or both depending on what the user needed, including for single-player and co-op games. At the time, merging the two source trees together seemed like the best plan of attack. It hasn't been, and it adds to the problem above as well as making bugs caused by us more difficult to track down.
No, we're not abandoning it. Far too many people want it, including many of the project's developers. How we're likely to get it into QuakeWorld's model is kinda complex. I can explain it most simply by saying "emulation". Not like what you do to get Mario on your PC, but rather having the server be smart enough to recognize the differences between a NetQuake game and a QW game and adjusting a few things as needed to make it all work.
That sounds like a royal pain in the ass, which is why we didn't try before. It's actually nowhere near as hard as we thought it would be. Ender and Tonik have both made major strides in this area already.
As QSG Standards version 1 call for the features Mega2k needs to be available, they will be. In fact, I've already implemented them in the new tree. Later this week I hope to be able to produce tutorials for how these things were done for people who might benefit from them.
Yes, the rumors are true. QuakeWorld Forever and QuakeForge are merging! The new tree is pivotal in this merge as trying to combine QuakeWorld Forever's features with the current QuakeForge tree would set back their development about two months. This is unacceptable for them and for us. The new tree is right from the beginning easier for the QWF developers to work with, making this very serious problem seem almost nonexistant. I'll post more details as soon as we have them.