3 ms·
At a higher level, here's the real problem. Consider the web today. We have web pages that display some text and a few pictures. Yet viewing one requires downl
by Animats 5y ago
At a higher level, here's the real problem.
Consider the web today. We have web pages that display some text and a few pictures. Yet viewing one requires downloading tens of megabytes of data, perhaps a hundred network transactions, a few gigabytes of RAM, a CPU or two running at upwards of 3GHz, a little help from a GPU, and several seconds of loading time. This is an awesome level of inefficiency.
Today's web browsing uses at least an order of magnitude more resources than it should based on what appears on the screen.
We cannot afford that level of inefficiency in good quality 3D worlds. Just doing a good 3D world with every known optimization is at the upper limits of what available hardware can do.
Watch the Unreal Engine 5 Matrix demo to see what's possible with the best known techniques.
Then read up on what's done to put those meshes into Nanite format, an appallingly complex, insanely optimized, and very clever representation. One full of internal offset fields, so hostile content would be a huge problem.
That's the problem with adapting web technology to virtual worlds. If you make the system as general, portable, and robust as web services, it's too bulky, too slow, and too low-rez.
- ethbr0 5y agoThe "web" today is effectively a basic virtual machine, and pages are containers deployed onto it. Which means you get all the bloat of re-shipping an OS (js libs), because the ecosystem as a whole would rather have diversity, speed of framework iteration, minimized developer time, and absolutely minimal client friction. And by and large, those have been good choices! Look at how far the web's come in just the last 30 years. But as you note, many of those goals are incompatible with game engines, which have seriously tight minimum performance requirements. So what do you think are the most important priorities for a 3D world format? I'm assuming performance (rendering and network) comes first, but what gets optimized for after that?
- Animats 5y agoFun. We know how to build user-created virtual worlds, but we don't know how to make them fun. I could say more on this, but not right now.
- toast0 5y agoAdding on, it's got to either be fun or useful, or it's going continue to linger in this sounds cool in books, doesn't make sense in real life (or movies, really, either). There's not a lot of usefulness either.
- ethbr0 5y agoI assume you mean fun in the sense of "not a pain in the ass" (specifically, in terms of common operations).
- scsilver 5y agoMaybe, but you can also end up with standard cached assets and people leveraging optimized frameworks like unity. Already the tooling is there, and I as a react dev am having an easy time building up a vr project, that I see as a next evolution of a website. Some concerns have to be spent on optimization, but 3d environments are a well practiced area, and we can get some great experiences for the same cost of some of these bloated Js websites.
- dataangel 5y agoThis is just because of where we currently are on the hardware curve. All of the relevant capacities are still growing exponentially. You can run any UE3 game on a cellphone. Just wait a couple generations.