[wilfra]

"we’ve learned that the successful teams...build on mature (read: “old, creaky”) technologies, because you want to spend time making the game fun, not trying to get the latest and greatest development tool to work."

That, rather painfully, makes a lot of sense. We're 18 months into building an HTML5 game (http://warsocial.com) using rails, node.js, redis, angular.js and a bunch of other 'latest and greatest' technologies. We probably could have gotten where we are now in a couple months if we had just gone with Flash + PHP or something, and put the rest of our time into making the game better.

[kevingadd]

New technology will always cripple a game project, yeah. The core task of actually prototyping and iterating on mechanics and design until you have something fun is just incredibly expensive and difficult; layering difficult technological challenges on top just makes it worse.

[tluyben2]

I like to prototype with a scripting language (LUA usually with ZeroBrane) and either Inkscape and/or Blender until mechanics / game design is fun. Not always but sometimes it needs to be almost completely rewritten to C++ or something else more performant than LUA (Cocos2D-x, Monkey language, Unity, Blender gamekit are my among my chosen poisons). Luckily most environments can be extended with a LUA engine, even if it's not supported natively.

[Stranger2013]

Whatever is the platform there is always a C/C++ for it...

[cglace]

Curious why node and rails? Is this a common combo?

[wilfra]

Node came later. Had we known we were going to use it, we probably could have avoided using rails altogether.

Would not recommend rails for game dev if you can help it. For a multitude of reasons - speed and environment headaches being the big two.