I wish I could get you to use Unity3D - imagine your game running in a browser with this kind of detail! http://unity3d.com/gallery/live-demos/tropical-paradise
 

I listened to a presentation about it a week ago. I was definitely interested, but there are a number of issues:

#1 - the IDE is only available on the Mac right now, and I don't have one (yet).

#2 - No source code.

#3 - Several key features are disabled under Unity3D's indie license, and are only available under the more expensive commercial license, which costs 7.5x as much.

#4 - I already have a lot invested in my Torque development right now - both in terms of cash (easier to walk away from) and time (not as easy). For a future title, sure, I'd consider switching. But I have to get AC and FK out the door first with Torque. By that time, maybe the above issues with Unity wouldn't be so bad.
 

Yeah, I must really love it. It is funny, though, how you create two problems for every one you seem to fix. Hopefully this LOD solution won't break the game down the line in ways I haven't thought of yet.

As for Unity...back when I started the project, I had to decide between Torque and Unity. I went with Torque for a few reasons, but I really liked Unity and I tried to talk myself into it. But at the time (3 years ago) Unity was still fairly new, and the community was not nearly as large and helpful as the Torque community. Plus, Torque already had a number of things implemented that I needed for the game, which meant it would be a bit easier for me.

That said, if I was now starting out with a new project, I think I'd go with Unity. But as TRC says, I'm just so invested in Torque right now that it doesn't make any sense to switch right now.
 

K, y'all are making portals into the anti-christ or something here. No they are not the optimal solution, but they do work pretty good, but you must take them into consideration when designing your structure.

In the case of the Pokmor Xang dungeon, I went into building it with the assumption that, since Map2Dif does generate a BSP, said BSP would be used to generate a PVS, a la Quake. As Jay mentioned, this is not the case. I did already know about portals, but (based in part on my experience with Quake 3) I assumed that they were only used to seal any outer openings or as "hints" to the BSP process. (BTW, in Q3, portals are automatically created for all doors -- when a Q3 door is closed, all geometry behind the door is culled; coolness!)

As I mentioned, the key to correct portal usage is: You have got to take portals into consideration when planning a structure's geometry!

By the the time we realized that there was no PVS being generated for Pokmor, it was 95% complete. And it had been build without portals in mind, so there were only two choices left: A) Completely rebuild it, so that portals could be placed correctly (Not an option!) or B) Place portals as best as possible. The portals in Pokmor did their job but... there too few and the placement of several of them was sub-optimal, i.e. very little geometry was actually being culled by them. But it was the best I could do "retrofitting" them in.

Well, lesson learned. Now that I better understand how the DIFs are processed, and how portals fit into that process, I'm certain that portals can be used with good success. In fact, I'd be willing to bet money that I could have resolved the portal issues in the Vespers3D structure, but it sounds like they have a viable solution already.

@Jay: "AC"? You're still planning on finishing it?! ROCK ON DUDE! =D
 

Yep, gotta agree with that completely. Your experience sounds very similar to ours -- we modelled our DIFs without portals in mind, and it became too unwieldy to try to retrofit. A complete rebuild just wasn't in the cards, and I think my modeller would have killed himself before getting all of them working the way we needed. A good lesson learned, yes.