I was wondering why uber team decided to develop their own engine when, for example, they could have used http://unity3d.com/ . Unity engine is one of the best and cheap 3d graphic engine out there. And its use would have allowed the deploy of PA on 10 platform automatically without any extra (almost) efforts. ? Cheers
A successful engine can be licensed, would be my guess. These guys are capable of making something that could stand the test of time. Think of where Unreal has gone since it was created. Also, its so much easier to cook in your own kitchen.
Well I'm assuming because they felt they could self-develop something better that they could get to do exactly what they want and easily expand upon for future games. Mike
Because Unity does not allow for the depth of control that you would need to create a game such as this and have it run decently on any pc. Unity is great for small studios wanting to create semi-simple games, but it's not very good for these kinds of ambitious designs.
No one wants to be tied to someone else's proprietary engine when you have the option and skills to use your own. I think Uber said they already had the engine anyway, or at least some of it.
If you want to pull everything out of a pc, you need to build your stuff from scratch. Unity would never be able to perform well enough to run a game like this.
They did a few months worth or preliminary work/research alongside of creating the Gameplay Visualization prior to doing the Kickstarter, but I don't think they had a functional engine, small test bed style aspects maybe. Mike
From the livestream, their 'engine' was more a set of libraries and code that could be used for just about any game, as opposed to an 'engine' like Unity (end to end system). And you have to consider - if you need to tear down a commercial engine like Unity and rebuild 90% of it to meet your needs... (new UI system, new graphics system, new physics, new pathfinding, new AI).. why NOT start with your own system first?
Unity is just a 3D engine, it can't do much other then render some pretty 3D stuff. What you need for PA is much more complicated. Client-server interaction, procedurally-generated planets, dynamic environment, dynamic pathfinding, ballistics, celestial mechanics, you name it. 3D renderer is going to be the least important thing here.
Unity is good engine, allow to make what you want, even tools. Great to make prototypes. But unity is for everything, and if something is for everything it means that it is for nothing. Its slow and not always allow you to do extreme, and cool things. Good that uber dont use it because they need something special. I know what im talking about, because i work in it every day, and i created "PA engine" on unity.
I've actually used Unity a bit; it is nice however look at a title like Kerbal Space Program (KSP). While a good title even it slows down on a high end machine once things get going. A custom created engine will be able to do what you want, exactly how you want it to do it. No extra overhead, no wasted effort. Besides, Unity does not support many of the things Uber is attempting to become the first to pull of successfully (like the player/server architecture).
A middleware like Unity3D is very unsuited for games like this. Too much original fancy stuff needs to happen to make it work, unity is extremely limited for deeply innovative games like PA. That said, if you can get away with using middleware as a developer, by all means, do. No need to write code that others have written before, and better, and for a longer time than you.
One of the things we've realized as a company is that we can differentiate our games on technology. Did you guys know MNC was the only game I've ever worked on where I didn't personally have a hand in writing the engine?
So what have you done in the mean time ? R&D ? And by the way, when you left GPG did you allready have in mind to make a game like PA or is it something that went later on yout mind ?
Coding isn't my primary job. I spend a lot of time doing things like business development. At the time it wasn't on my mind. After making a game it's nice to change gears for a while. SupCom was my life for a few years there.
As a veteran in the game industry and (as far as i'm concerned) as a recognized very good software engineer (to me, "Mavor" sounds like "Carmack" when i ear/read it ), Have you ever considered writting a book that would, at the same time, explain (from a technical perspective) the different challenges you had to face (how you manged to workaround them) and also relate these pieces of life ? A book might not be the good format, would also take too much time, would not even be something profitable, and i know you maintain such stuff on your blog [I know all this because as working in the Oracle DB work area, all good books written on the subject require at least a full dedicated year from their Author] But as i used to enjoy reading Michael Abrash books (relating sometimes his experience at Id software) I sometimes regret there are no such book anymore written by recognized veteran from the game industry combining technical topics and experience feedbacks as a software engineer in the game industry. Ok was off topic, but just to say i'm really facinated by all your stuff
Thanks for the comparison but I'm no Carmack. You have read my blog right? www.mavorsrants.com. That's my book. I'll keep writing more anecdotes there... Mike is a great writer and a brilliant guy. He's also been writing some interesting blog posts lately. In fact my blog started as a response to one of his posts. The first time he wrote about the augmented reality stuff I called him up and spent some time with him at Valve going over the general problem as it's an area of interest to me.