Hacker News new | ask | show | jobs
by plopz 552 days ago
The slow progress of WebGPU and WebTransport has hurt my enthusiasm for games on the web. I was so excited 6-7 years ago but it feels like everything slowed to a crawl.
1 comments

If you want real games on the Web, the answer is streaming.

Even with WebGL 2.0, there is nothing at the level of iOS and Android OpenGL ES 3.x games, after a decade.

Additionally, browser vendors haven't yet provided any debugging tools.

> Even with WebGL 2.0, there is nothing at the level of iOS and Android OpenGL ES 3.x games, after a decade.

I'm not so persuaded the barrier here is as technological as this forum is predisposed to believe, although I will concede that the Resident Evil iOS battery melter has no web equivalent to date.

The real problem is the web audience is wildly different to other platforms, and has very different expectations which prioritize speed of loading and then extreme long form engagement with little threat. This has created a very different ecosystem, and one that when it encounters something technically impressive goes "oh nice" then moves swiftly on to something else.

For example, you could 100% do Minecraft on the web today, with P2P multiplayer and everything else, and it's kind of revealing that this isn't a huge thing already.

Infinity Blade, the game used by Apple to demo iOS OpenGL ES 3.0 in 2011, is the bare minimum.
But you absolutely could do Infinity Blade. No one does because it's not worth the effort. (I would argue this was true on iOS too - the games that made money did not look like Infinity Blade).

That recent Marble Madness a like https://news.ycombinator.com/item?id=42212644 was a far better fit for the audience on the web, and is not technically unimpressive, considering how smooth and responsive it is, along with the image quality.

And I don't have the same amount of assets, but in terms of rendering features this is more than Infinity Blade: https://luduxia.com/whichwayround/

Being worth the effort or not, doesn't change the fact it isn't available.

That recent example, was designed for desktop, for example, lacking gyro use, and doesn't respond well to touch.

That demo looks more like a PS2 kind of thing, 2000 technology.

I suggest you go and look at youtube videos of Infinity Blade, because that doesn't use physical lighting models or even have real time shadowing of any kind. It is all just big textures covering 90% of the screen.

You have very serious rose tinted spectacles.

This WebGPU demo I'm linking to proves that statement to be incorrect, it's far superior to iOS and Android OpenGL ES content: https://play.spacelancers.com/
If anything, it proves why it will never replace native APIs on mobile devices, it killed Chrome on my S6 tablet.
What “real games” on mobile devices are you thinking of? I assumed when you said real games you meant desktop games only hardcore gamers play.

So why does one game crashing on a tablet prove anything?

I think you’re right about browsers not providing enough graphics debugging tools… at least half the entire problem is browsers. They also don’t provide storage APIs that can deal with game assets, nor robust APIs for audio & controllers & peripherals. For better or worse, the current set of anti-cheat software for competitive games can’t run in the browser. The other half of the problem is distribution and ecosystem.

To a first approximation, around ~0% of the problem is WebGL, at least for mobile games, casual games, and most non-AAA games. The graphics is the one thing that’s more or less there and good enough, it’s everything else that’s missing.

I don't know, maybe something from Activision, Blizzard, EA, Tecent, RockStar,...

It proves how fragile the whole process is after a decade.

It proves there’s a bug in one indie web game, and nothing more. It proves nothing about the process or the ecosystem or what can happen in the future with APIs.

> Even with WebGL 2.0, there is nothing at the level of iOS and Android OpenGL ES 3.x games, after a decade.

The main thing it seems you’re confused about is that ES3 and WebGL2 are very similar, WebGL2 was designed to be compatible with ES3. Why do you believe that ES3 is far superior, and what features, exactly, do you believe ES3 has that WebGL2 doesn’t?