Neoaxis or Shiva don't cost me a thing in order to use the full engine with more features (not mentioning the "hidden" restrictions like Com may have weighted bones but in Com it's just one vertex per bone although that might have changed in the meantime), more modern tools, a multitude of shaders that work out of the box on anything and especially when considering using Unity's free version you also have access to an asset store with lots of assets. Especially Unity's asset store is fantastic with entire game kits, pathfinding algorithms, editor extensions, particle effects or even visual scripting solutions...

From my point of view especially the tool situation and the shader aspect are a huge downside for 3DGS. The tools didn't really evolve and still are stuck in the Half-Life 1 era. We have been promised an overhaul which also is on the forecast page but who knows how it'll turn out in the end - I actually expect nothing close to what Unity or Shiva (2.0) offers. Shaders also always were terrible in 3DGS as things that simply worked out of the box never really existed and in order to really make good use of them you have to have an in-depth understanding and should be able to write them for yourself. I also don't know if shaders on the BSP level geometry ever started working the way you'd wish them to do...

In my opinion 3DGS simply started to miss the "Zeitgeist". Look at MED - that's a mid 90s mindset. Nobody needs such a modelling tool anymore and it still got quite some love when free solutions like Blender or Wings3D existed and were good at what they do. You nowadays should be able to import from whatever tool you chose for your work. I can't even say if it now does import models with bone animations or if you still have to animate in MED...


For the most part I fail to see real arguments here. Mentioning the lightmapping on terrains especially feels wrong as the 3DGS lightmapper always created rather unsatisfying results and I still remember people constantly asking to "outsource" the lightmap information so they could use a software like Gile[s] instead to bake their own actually decent looking lightmaps...

Talking about multiplayer capabilities also is a rather bad idea as the users in this forum did show that the 3DGS multiplayer system is in many parts a catastrophe that might in theory be able to handle some nice number of players etc. but fails badly on rather simple tasks when actually using it. What I especially remember is that you could compile whatever program of yours and make it try to connect to a 3DGS multiplayer server of someone else which results in a server crash. So in the end just as for the lightmap example the multiplayer component is there on the paper but is lacking in its execution. You still should be able to find topics about it and I think someone at a certain point started integrating RakNet into their project as the 3DGS internal system simply was unable to deliver good results...