soggie said:
There are 2 common problems with 3d engines: (1) polygon limits and (2) camera controls. I'm sure all of you know the latter, which can't be solved even with fixed angles (especially when you have tall buildings that might get clipped in the camera's near clip plane).
That is a design problem, not an inherent issue with 3D. Seriously, tall buildings? Come on, what kind of programmer are you to cite such a non-issue? For every 10 3D games with overview camera that got it horribly wrong, 1 got it right. Most recent example that I've played might be StarCraft 2 where the I've yet to experience a single instance of obscurity due to tall buildings or other structures and SC2 was full of them.
Polygon limits, while being very capable of rendering ultra realistic scenes in games, doesn't not perform well on older generation hardware. The last factor is pretty crucial as I can't expect my target market (indie game consumers) to always have cutting edge hardware.
Unless Unity comes with Unity henchmen who oversee your development and put a gun to your head when you use low poly models and don't exceed a certain poly count, I don't see the problem here. Such a non-issue again. Besides, hardware-wise Unity is pretty flexible unless you start using the default bell and whistles features. User base of Unity doesn't exactly consist of the bleeding edge hardware owners.
Now about U3D's engine issues. Yes it is powerful, but without access to source code and a limited license (100,000 revenue before you have to upgrade the license?), that's a huge turnoff.
You only have to upgrade to Unity Pro from Unity at that point, where the latter is actually free. I find that pretty reasonable and agreeable. Some people have also been granted the source code; they have an address specifically for handling source code requests. I imagine it would depend on the commercial viability of your project, your experience with Unity and your ability to convince you have the skills to see it through. They are pretty much hush-hush about licensing conditions, though. They probably handle these on a project-basis.
Now in Unit3D it is always possible to do this by manually placing all these things in the editor, but what about reference grids, grid snaps, and so forth? All these are crucial to isometric map building and unless you want to place the world by hand, I don't see how U3D fits into the picture.
Invest in a month or two to develop your own editor -using Unity again- to do all of that just the perfect way for yourself, then. It's been done. You'll have to do that with any API anyway, otherwise content creation pipeline will be immensely cumbersome and what would take a few days to create a simplistic quest in, say, Bethesda's games could possibly take beyond a week. There are at least a half a dozen Unity projects where people have done exactly what you have described and not surprisingly at all since 2.5D is one of the oldest methods of achieving good looks with least resources and thus one of the most widely recognized way of achieving a goal in game development.
Imagine if you wanted to try a specific rendering method, and found out that U3D only supports a specific way of forward and deferred rendering.
How fortunate that it doesn't. Community is crawling with custom rendering and lighting techniques, based on who knows what, like obscure papers some Russian published in some obscure indie-dev platform and they work. Some of them quite very advanced too, mind you.
You'd be limited to playing around with shaders to achieve what you need, when using Ogre3D you could have dug into the source and replace what you need. Arguably, U3D's plugin system might be able to solve that, but I believe that's not worth the trouble when you have all the source code in your hand.
For an experienced programmer, I can't argue about the value of source code vs. low-level scripting (ie. that's a lot more than just plugins) and plugins.
Finally, we're talking about bulk here. I need a rendering library, not a full blown game engine. This is where Ogre3D comes in handy, because it does what it does best - rendering. I like specialized libraries like these.
Unity is only a full-blown game engine if you need it to be. There are as many people developing non-gaming middleware using Unity as people making games.
Even though I'm at best a novice at programming (I most certainly couldn't endure to start from absolute scratch with just a renderer; it's just way too fucking much work without advanced libraries that I can simply adapt to and work around, instead of studying and revising), I can safely say that your knowledge of Unity is terribly criminally limited and skewed. Unity isn't a mere plugin based game world editor with them shinies. That would be like saying Excel is MSPaint+ with more text. That's the impression you seem to have of it, anyway. I probably sound like a Unity fanboy, what with suggesting it every now and then in random threads when the subject comes up, but I only speak of what I know to be true by experience.
Also, not that it's much of an argument by itself but just to add weight to it anyway: when Google goes out and picks Unity out of a hundred similar software to provide native support in their OSs and browsers, you know there's something to it.