Chris Taylor is back (TA, Sup.Commander)

Why so needy for something new when the "old" works just fine?


Although I guess in that sense, let's all go back to playing TA or SupCom.

yeah thanks x'D

Until May this year WebGL was painful for any advanced browser based games as Artillery games found out the hard way after 3 years of development.

WebGL is based on OpenGL ES 2.0 which was based on OpenGL 2.x functionality.

Chrome 56 released in January 2017 now supports WebGL2 which is based on OpenGL ES 3.0 functionality. Firefox 51 also added WebGL2 support in January.

Artillery Games were probably just a couple of years too early. They needed WebGL2 and WebAssembly for performance. Operating in a secure browser sandbox also limits direct access to CPU / GPU shared memory.

iOS and Android natively support OpenGL ES 3.x if you want maximum cross platform performance on mobile.

Most iOS game developers have abandoned OpenGL ES for Metal 2 (same with OpenGL on macOS) simply because of the huge performance gains over OpenGL ES / OpenGL.

Epic's Zen Garden was the original iOS Metal 1 demo from 2014 and is now available on WebGL2 and WebAssembly:

• https://www.unrealengine.com/marketplace/epic-zen-garden
  
• https://s3.amazonaws.com/mozilla-games/ZenGarden/EpicZenGarden.html (use chrome)

That kind of demo from 2014 was not possible on the web until this year.

The future of WebGL and web based gaming will depend on which proposal wins out for the WebGL Next API.

• https://github.com/KhronosGroup/WebGLNext-Proposals/tree/master/Obsidian-Mozilla
• https://webkit.org/blog/7380/next-generation-3d-graphics-on-the-web/

EDIT: Fixed Chrome 56 release date.

What specific WebGL 1 limitations were an issue to them? There certainly are WebGL 1 demos that are pretty cool as well. Dunno about the specifics of the Zen garden thing though. Can't run it here right now.

Out of my memory I can only remember that, compared to WebGL 2, it is missing a way to render to multiple targets at once, making deferred rendering inefficient/impossible. But does one really need deferred rendering that badly if one only wants simple comic like graphics? I don't know. But I'd guess not.

Getting enough js performance to make enough draw calls may be an issue, looking at it the instanced rendering extension actually had not as much support as it has now back when they tried to make their thing. So I guess that could've been at least one source of a bit of headache.

Sadly that blog post about their in browser issues I read is nowhere to be found anymore, but my memory only contains "meh networking",
"meh mouse cursor handling" and "meh browsers change beneath our feet too much".

It probably comes down to how much of a step back in terms of graphical quality they were willing to make.

Did Atlas die to technical issues in the end? Without any knowledge of the project expect "some sort of RTS in the browser made by people who are fond of competitive Starcraft" I'd guess it might have died to user interest being too low? People who like that kind of game just continue to play Starcraft 2. I certainly do and didn't care one bit for Atlas, except as being interesting from a technical standpoint.

On iOS back in 2013 the big advantage of OpenGL ES 3.0 over OpenGL ES 2.0 was reducing the number of draw calls, improved texture handling and the updated shader language. Some important extensions moved into core like instanced rendering that could be combined with new extensions like UBOs. The private tech talk for developers was pushing multiple render targets and framebuffer fetch for single pass deferred rendering of lighting / shading effects.

It wasn't my area of focus or knowledge. Everyone was very excited about how you started to have feature parity with consoles and desktops. It was the first time you could deliver a console like visual experience on mobile. Metal 2 makes it more like a desktop visual experience.

The Artillery blog is still available via: https://web.archive.org/web/2016083...:80/2015/09/artillery-native-game-client.html

From what I read they did not die due to technical issues. Funding and dwindling player interest are often mentioned.

It did look like they had burned through a lot of development time overcoming technical issues in the browser before making the switch back to a native client for performance.

Starting now with WebGL2, WebAssembly and the huge increases in JS engine performance over the last 3 years might have been a different outcome for them.

I guess there's some truth in saying "it's the browsers's faults"

I don't see why not afterall. If we're making clean code then we're not replacing functionality that should be handled by something other then our code in the best of worlds.

in this case it's even impossible to hack the cursor lock functionality in.

it HAS to be the browser suppliers who add game-mode functionality to their browsers.

I guess not having to worry about doing things different then they are used to from the bigger platforms is a more important plus than I might realize. Interesting.

But the iPhone X's specs are comparable to the Galaxy S6 which came out a few years ago. >.>

Not in sustained real world performance especially with the GPU and that is what really matters especially with advanced games.

We'll find out in November how much of difference. Most people seem to be waiting for the X.

The S6 swapped to a Samsung Exynos 7 Octa 7420 for non CDMA devices instead of the SnapDragon 810 due to overheating issues in Qualcomm's rush to compete with the 64 bit A7. The S6 benchmarked well in multi-core but could not deliver sustained real world performance under load and the S6 is not even competitive now.

The S8/S8+ even with high performance game mode enabled will drop frames consistently under high load.

The Google Pixel shows what can be done with highly optimised performance on Android with lower specs. We'll soon see what the Pixel 2 and Pixel 2 XL can do.

... meh ... just gonna wait for him bringing "something" out ... after kings and castles, wildman and taylor joining wargaming there hasn't been anything noteworthy from him since .. so consider me very skeptical ...

Chris Taylor reveals Kanoogi cloud gaming platform and Intergalactic Space Empire RTS

https://venturebeat.com/2019/04/24/...-platform-and-intergalactic-space-empire-rts/

https://twitter.com/kanoogi
https://main.kanoogi.com/

^^^^
still aplies ... not gonna get hyped from some interview that doesn´t show any gameplay ..
also a webbrowser based game ? ...... :/ ...... .... eh ....

https://main.kanoogi.com/signup.html