Good decision. Fun > realism.

Is scripting definitely coming back? I ask because I've made enough games in Javascript (where "real code" and scripts are the same thing) to realize that too much scripting freedom tends to make a mess of things. In my current game I only have "fixed functions" coded in C++, and my scripts are just declarative data in a simple text format.

If it only has a few features but they are implemented well, it will be fun. Simple as that.

I always refer to Antoine de Saint-Exupery:

"Perfection is Achieved Not When There Is Nothing More to Add, But When There Is Nothing Left to Take Away"

It is so true, but although very hard to get rid of all the stuff and ideas. But once you have overcome the desire to protect your time investement (it is a ugly trap), your project will get better

And if my implementation of a feature can't feel great, it gets dumped.

It is interesting to see, that you run through the same decision I made for my game. I don't remember who said it (I think it was the designer of original Sim City ): If it is fun, double it, if it not fun, half it.

since I'm bored of looking at coloured polygons now.

Maybe you should reconsider this Early game developer were able to create games solo or in really small teams due to massive restrictions. Restrictions will help you to find interesting ways to define content. Once you add texturing, normal mapping, complex materials etc. you will have the problem of generating content.

At some stage of my project we had it all in our game: normal mapping, displacement mapping etc. and suddenly we weren't able to generate content in satisfying quality and in time any longer. I dumped this approach and we do only simple textured models now.

Look at this low-poly art: http://kingpolygon.tumblr.com/

So, your approach has lot of potential, with some interesting technical improvement (pre-backed AO, SSAO, GI) your art will improve over night without improving the content creation at all.

Good luck, you are on the right track

Very solid advice Ashaman73. I totally agree with restricting the awesome realistic graphics for solo. My wife looked into some voxel editors and started producing models I'm proud to have in my game.

I recently had the privilege of talking to a pro game designers/producers. I was telling them a story about how I had stopped working on my space-trader game because I was pretty far along in the features, but I still couldn't see much fun. I switched gears to a tactical dungeon crawler and after I made decent progress, I definitely saw fun potential. He told me that idea was one he tries to teach to other indie game developers at conferences, and it's also one of the toughest to do.

Good decision. Fun > realism.

Is scripting definitely coming back? I ask because I've made enough games in Javascript (where "real code" and scripts are the same thing) to realize that too much scripting freedom tends to make a mess of things. In my current game I only have "fixed functions" coded in C++, and my scripts are just declarative data in a simple text format.

Yeah, plan is to have scripting working in a big way in this game, similar to how it worked in an old project of mine which was quite successful at the time.

Basic idea is you add physical objects to the level in the editor called ScriptGates, which you place on the map but don't appear in the game. Each has an input and one or more outputs and can either do a simple task - negate its input to its output, pass on its input to its output after a delay etc - or actually run a full Om script with access to its inputs and outputs.

Objects then like doors, say, have an input that opens the door when it is 1 and closes the door when it is 0. Or you can have an invisible polyhedron that, when entered, sends a 1 signal to its output and a 0 when you exit.

It was quite amazing to me how easily you could then create very complex level behaviour with this visual and visceral system.

Thanks for the other comments. Makes me even more certain I'm on the correct track here now.

Much appreciate you taking the time to confirm my thoughts

Objects then like doors, say, have an input that opens the door when it is 1 and closes the door when it is 0. Or you can have an invisible polyhedron that, when entered, sends a 1 signal to its output and a 0 when you exit.

It was quite amazing to me how easily you could then create very complex level behaviour with this visual and visceral system.

That sounds really slick. It reminds me of Creatures 3 which had something similar, only the objects WERE physical objects in the game. So you could say have a gun turret that fires when it gets a signal. Then you could hook up a motion detector which would send a 1 when it detected motion. You could also wire it up to an AND gate to another component that would detect when a selected creature type was in range. Then your friendly creatures could walk around unharmed, but when an enemy creature walked up your gun turret could light them up.

Objects then like doors, say, have an input that opens the door when it is 1 and closes the door when it is 0. Or you can have an invisible polyhedron that, when entered, sends a 1 signal to its output and a 0 when you exit.

It was quite amazing to me how easily you could then create very complex level behaviour with this visual and visceral system.

That sounds really slick. It reminds me of Creatures 3 which had something similar, only the objects WERE physical objects in the game. So you could say have a gun turret that fires when it gets a signal. Then you could hook up a motion detector which would send a 1 when it detected motion. You could also wire it up to an AND gate to another component that would detect when a selected creature type was in range. Then your friendly creatures could walk around unharmed, but when an enemy creature walked up your gun turret could light them up.

Yeah, the last project I developed this system on was a mix of physical and non-physical objects, so you'd have inputs and outputs on doors, spikes, fireball shooters etc, and also a collection of non-physical gates that were shown as physical objects in the editor but not shown in the game.

The non-physicals were things like a timer that sent its signal on after a delay, or a splitter that took an input and sent it on to multiple outputs, or a not gate that just negated its signal and so on.

So you could build really complex behaviour out of collections of these. Would love to implement the same thing in a 3D world this time, previous project was a 2D sidescroller but principle is exactly the same.