(An Enterprise Proposal, something I've been kicking around awhile, now re-iterating)
I grew up in the 80s, largely in Michigan, athletic culture, laid back. We got a PC XT with monochrome (amber and black) monitor. I played Falcon. Have been obsessed with simulation since then. Really it's the only thing about computers that really excited me. Contrarywise, I thought and think AI is bullshit and hype, obscuring some genuine utility in the neural nets, viewed as maps of data terrain. Roger Penrose nailed that one, also Noam Chomksy in "Language and Thought" (Anshen Lecture Series) makes similar, and powerful in generality arguments. I was never into the web hypes, other than the utopian library of alexandria type ideas -- I remain committed to that -- I'm talking about the hijacking for corporate commmercial, oligarchical, society-parasitizing/predating interests. Commerce is good of course, and internet has been great for that. Large players with economy of scale are good. Cultural central management, along with outright oligarchical annhiliation of healthy markets, is insane, indicating correction.
I digress.
Point being, commerce is cool, the web is cool, but it's not that interesting, and, outside of the techniques and tools of networked, distributed systems. (See my draft paper "Engineering Critical Network Applications", by the way, centered around network simulation.), I'm not super interested in the web-services and corporate IT software projects.
Crypto seems to me a small problem consuming the cognitve resources of a society of smart-boys who think a math-trick is the path to bling. The problem is sound money and good book-keeping, necessary to a civilazation but not sufficient. It's a small problem and I am not impressed with the hype.
So, for me as a software engineer, as a computer dude, in terms of pure computing, (excluding embedded, robotics, avionics, things not purely computing) there is only one thing that really interests me, and that is simulation.
Simulation is beautiful because it is the purest science; paradoxically the modelling as transcription from canonical reference theory (continuous) to computer code (discrete), together with the poverty of computing power measured against the resolution of reality, makes this a plastic and creative art at the same time. One has to choose, on a meta-scientific level, what to put in and what to leave out of the computed state, the state model. This is clearly going to be, at the high-level, an intuitive, painterly process. Where our statement doesn't apply are hyper-specialized engineering simulations, covering in highest resulotion possible a confined scope -- this or that piece of equipment subjected to this or that physical force, dynamic load or whatever. We are discussing things more like sociological simulations, simulations of a human being, of a battlespace, or even just a platform in the environment. These are massively complicated situations as compared with any computer power that could be thrown at them.
If the above wasn't clear: Simulation is basically transcribing the work of some actual scientist into computer code that does a good job performing a discretized model. Thus far we're just transcribing "science", mathematical models; we're doing applied maths and software. We run into a problem however with our computing power, as compared with reality. At this point an intuition comes into play which is artistic, of course the esthetic sensibility is attuned to a natural philosophic approach to a particular problem in the world that we want to study. What is at issue is what exactly constitutes the model, what gets left out and what gets put in.
The basic plan is to start with a combat flight-simulator, on the model started by Falcon and continued with Digital Combat Simulator (DCS). A wrinkle in the development concept is that the code will be built by a distributed team of free-lancers, working on an open-source project. The game will be open source. A vulnerability in the business plan is the non-ownership of the code, and even the mercenary nature of the workforce, volatile, it's hard to put a handle on what capital such an enterprise could realistically lay claim to, even considering a successful product. It could be forked, people could leave en-masse to start their own thing.
At the end of such a process, Terebinth is just the center of gravity for a collective that can be orchestrated to a variety of custom simulation projects. The combat flight sim / battlespace sim for the civilian market is simply an open piece of culture, developing the developers and developer culture.
Most of the code of Terebinth will be proprietary and secure within constraints largely up to the client. The only open-source compoonent is this first project, we have called Peregrine, after Falcon. Incidentally, the first model would be F-16A.
The timeline for the first phase I guess at 2-3 years. A totally arbitrary guess. The budget I'd scrawled arbitrarily at 50 million, but this may be high.
For a networked multi-player battlespace simulator or a networked game, the current topology is highly inefficient of computing power. Given a network with latency minimized under frame rate, networked, both centralized and distributed, computation of physical and graphics state --yes could send individual frames over the network. At present, the games on the consoles are limited by the computing power of the console. With a faster network, we could leverage a warehouse full of computing power, memoizing for efficiency across clients.
Hence, while Peregrine project must necessarily prepare a desktop prototype, it is my belief that the flagship product will only be available on high-performing networks, in terms of latency if not so much throughput.