I quote from a post I came across ,by chance ,on an other site
I fully agree with it
Pure common sense and real experience in game programming
The best review about 3dgs I have ever read
quote
....Think about that because the bottom line is when you pick an engine you hope to produce a finished release product that is stable. When I look at engines I look at that first. Not a feature list. Next I look at ease of use that matches the team I Have. .....
...In fact I would absolutely hate to code a complex RPG game in C++ or Visual basic even if I knew both very well as the code management would be a nightmare.....
This is why you do not see too many finished projects that are complex with engines that do not have a scripting language....
Another plus when looking at engines. Ease of use....
Yes the editor is somewhat old world. But it is very powerful. The scripting tool is tied into the editor, and console feature in the running engine allow easy script testing. It's a subset of C++ and very powerful, modular and compact for easy code management. Doing anything complex this is paramount.
People knock the modeler MED. Why? s it lacking? not really. It just lacks the 'auto click' features of high cost modelers like MAX. Here a button press does lot of things automatically. In MED you can get th same results, it's just your have to do the work. That's really the difference in a nutshell. Own Max and know it? Then use it. There's importers for every modeling program out....
A lot knock the BSP culling system. Sure it's old hat. But it renders extremly fast. It great in underground levels. It's great because you get good results on old systems.And recenty in 6.40.5 Occlusion Culling has been added and octree is in beta. Decent outdoor levels are possible, you just need to know what your doing. There's nothing stopping you from saving the level in .wmp file format and using MED to convert the whole level to a model. You simply have to flag the model in WED for polygon detection and use c-trace and c-move in your player biped scripting. And these model levels render instantly and I have achieved 60 FPS with 500,000 polys in view without any LOD added yet, Which by the way is also built in the engine ready to use. (In fact the next update will generate the 3 lower poly models from the original automatically..
This is now my 3rd game project with 3dgs. My other 2 were released and stable. Spam1024 being the biggiest and most downloaded and I never got one single end user complaint related to the engine but rather all driver issues or end user fault. Not one. It's rock stable.
In the end you want to get done. Low cost engines will always have you compromising on features. All have ups and downs. But in the end if you can't see release day with a finished stable product then I really don't care what the engine you use can do or can't do. It's not a good choice.
... INDY teams fail 95 percent of the time. Most times because thier too busy ew'ing and aw'ing to concider any of the above.
The old adage of working at your level is good advice if you want to suceed. You level up skill wise fastest that way too. Jumping over your head in engine choices will almost always end up in frustration and failure. Stay at your skill level and your skills increase faster, and the games you make increase in quality as you go.
Most INDY developers don't understand this. They rush to devmasters and look for the latest and greates in hopes in competing with the big corporate giants. You can't and your not. Focus on the gmeplay, not the pretty pictures. I play more Xcom in dosbox than I do Oblivion. There's a reason for that and it has not a damn thing to do with shaders, particles, culling systems or whatnot.
Unquote
The complete post is available on TV3D site, along with their usual fanatic comments