Honestly, I don't see any reason to buy Unity for simple indie projects - Personal Edition is quite enough. What features of PRO+ versions is really worth it to buy a subscription?
In a next few days I will try my best to test on all mentioned engines/frameworks.
Personal Edition can be licensed up to 100.000$ of gross revenue, after that you HAVE to start paying the monthly subscription. It is perfect for people working alone, but it is a indie teams killler to me. Also because you cannot do something like:
1) team member pay the license and sell the game so he is the only that have to pay..
because all team members have to have the same license (and the same apply to assets bought on asset store: 1 license per seat on same project, and all team members should have the same license)
Well, I'm plan to make it free, so Pro Edition doesn't make much difference from Personal Edition for me.
I will always keep in mind Unity, but come to it only after some experience in more simple engines (I mean, at least one completed full-featured game). I think it's important to make simple things first to earn your wings, and go hard later.
While I will concede that for a majority of indie titles, ready-bake engines are completely permissible, and viable options (And in this case I do indeed recommend you use an engine for your game). I don't think however, we should be so swift to refer to DIY engines as, "Reinventing the wheel." Because with all honesty it only takes one with enough intelligence, and ingenuity to realize it can always be done better, and sure enough we'll never be able to fathom on how we made do with the wheel ;)
Besides, starting from the API level will give you much better perspective as to how many subsystems that make up a game works under the hood which could serve as valuable knowledge that you can take with you to higher level engines, and frameworks.
Also, making a game at the API level is not synonymous with making an entire engine from scratch. Nowhere close.
Marcus Hansen
I still recommend GameMaker: Studio. Besides all of the above mentioned good things about it, there are others. The newer versions introduced "shaders" into the pipeline. Technically, you could use them for 3d, but the best use is for 2d stuff. I'm a fan of "normal-mapped sprites" which can set up really nice lighting effects with sprites. You can use normals you render yourself(if you are using pre-rendered 3d sprites for art pipeline), and there are programs like SpriteLamp(and another one I can't remember) that also help create normal maps for this. Of course there are plenty of other uses of shaders for 2d, but this is a big one.
The feature that isn't discussed that much is the basic simplicity of the software and the programming. It may be considered bad practice by many traditionalists, but there is great speed in the way things work with GMStudio. The resources can be accessed from anywhere. You don't have to worry about drag-and-dropping resources as is normally recommended with Unity(though with some code you can make that go away too). You still have the nice event system, which Unity also has. The GML language is somewhat C-based, and somewhat original. There are advantages to only having 2 variable types(doubles and strings). And yet, you can still have things like data structures, arrays, and working in binary. You don't have to remember where to type in a "new" statement like with C#, but you can still do the basics like function calls, code re-use, data manipulation, etc... that C# can do. And, you can get all that GML code compiled like C++ if you so choose(and buy, as the main disadvantage of GMStudio is that you have to pay to get the best goodies).
Basically, my point is that despite the history of GameMaker's previous versions being more hobbyist, being "toys" and what not, the newest incarnation is actually quite powerful and capable.
