Chris Taylor is back (TA, Sup.Commander)

Ah good to see they're factoring in throttling (for iOS at least). Promising news this generation in that regard.

I have other problems with Apple, but it tends to relate to the ecosystem and their tangled mess of authorising apps for the App Store.

Another off topic response to mobile challenges and why I would only target iOS for a PA live game / replay viewer...

The other issue with mobile games that push the hardware limits is that Android is a total crap shoot.

There are significantly more Android devices most of which are NOT capable of running advanced games. Only the latest flagship Android devices come somewhat close to iOS CPU / GPU performance.

The market size of capable Android devices is actually much smaller than iOS with the majority of devices sold now being low-mid tier devices driven by cost which make almost no $$$ for the Android device manufactures. They are basically fighting for scraps. Contrast that with iOS where every device sold is a premium device capable of running advanced games.

At the end of last year, worldwide market share for premium devices was around 70% for iOS. Samsung was the closest competitor with around 17%. In mid-2016 that was closer to 50% for iOS and 25% for Samsung.

The iPhone 8/8+/X upgrade cycle might be over 80% this year sucking all the profits out of the mobile industry for everyone but Samsung which at least has some premium market share and other revenue streams from chips, memory, NAND, screens, etc.

That makes it even harder for Android device manufactures to keep up with the pace of investment in CPU / GPU performance over the next 2-3 years.

guys let's not let phone wars seep into here.

But honestly now who thinks getting a simulated projectile RTS onto a phone is legitimately doable?

I personally don't but I'm starting to believe I could be persuaded.

for example they did a tablet and phone version of world of tanks but they removed the sim pro to make it happen

I guess the question is: how many projectiles? How many units?
I'd say something on the scale of TA would work. Cause potatoes can do it.
Performance wise I don't think it is an issue. If a Pentium 200 can do it, so can a modern smart phone. I am rather confused about UI issues and the "why would anybody even want that"-factor. But I have that about most things smart phone...

And on what phone? There is a massive difference between my ancient android brick and the newest iPhone.

That's the kind of response you expect from #idiot.

Take those hashtags to twitter, where they belong please

Yeah mikeyh wrote good arguments I can recognize that even though I'm an Android developer (and really hate ios). My apps are not very demanding and I agree that the iphone market is bigger for something that requires more CPU/GPU performance.
Iphones may be less cost-effective but there are no cheap Iphones.

If you were porting something like TA with a unit cap then you would target any iOS device since the iPhone 6 and any recent flagship device from the top Android device manufacturers.

ie what is considered the premium mobile device market.

For testing that would focus on iPhone 6 or newer and Samsung S6 or newer which will probably be most of your potential world wide market. China would be the exception to that.

CPUs / GPUs have not always kept pace with the rush to increase screen size and pixel density so you may find devices that do not perform as well as expected under heavy load.

For something like a PA replay viewer it's really only the iPhone 8/8+/X and in reality probably just the iPad Pro with it's 4GB RAM and extra GPU horsepower as a starting point.

World of Tanks Blitz?

That was originally developed in 2013 and they would have wanted to target the widest possible audience including older devices.

eg they are still supporting iOS 7 from 2013 and Android 4.0.3 from 2011.

For iOS that's all the way back to iPhone 4 which is not even listed on: https://browser.geekbench.com/ios-benchmarks

Yeah agreed. Google must do something huge to change that, and I think they are in the right direction with the new Fucshia OS and the HTC aquisition and the in house chip development (former appler arquitect). Still, I cannot use iOS and a daily driver, to me Android is so much better. I might buy the iPhone 8 only for gaming if the new arkit brings something really interesting.

The race to the bottom by Android licensees, Samsung's willingness to use / abuse / sabotage Android and Google's start / stop / indecisiveness around developing their own hardware in competition with their own licensees has left Android exposed in both the premium / high margin and low cost / low margin / high volume markets.

This article is a good summary of the HTC acquisition: https://arstechnica.com/gadgets/2017/09/motorola-redux-google-appears-set-to-buy-htc/ (Google might buy a dying Android OEM again)

I have an ARKit developers kit which will probably never get used now that I have stopped work on everything.

UPDATE: HTC not acquired... https://arstechnica.com/gadgets/201...ogle-to-acquire-part-of-htcs-smartphone-team/

@cola_colin fair enough I guess we can expect simpro on phone to be feasible (somewhat) ... let's move on to another contention point what techs he's thinking of when he sais the entire thing will be in a webpage. (please dear go no webgl ! can webgl even do simpro?)

Tatsu- it will be client server I'm sure of it... so it will be Sim Pro, however the host machine won't be doing the sim at all...

Maybe he'll do something similar to the web platform Project Mercury that was being work on at GPG before they joined Wargaming.

https://www.polygon.com/2013/1/29/3...emonstrates-new-mod-tool-and-platform-project

if you do that that you rely way too much on potentially very bad internet.

phones don't necessarily get internet that's as good as PCs. and even then we've seen that this can be one of the issues with that model even here with PA.

WebGL is awesome and a great thing to use. Has nothing to do with doing simulation stuff though. It's just OpenGL (ES) for the Web really.

Puting it all into a webpage has the great advantage that it doesn't require all this shitty stuff of dealing with various app stores and being at the whim of whoever runs that store.
It has the disadvantage of putting up with the whim of the main browser manufacturers though.

the thing is new techs just came out that are potentially massively more optimized and potent then webgl

I'm thinking of Vulkan

but I guess I'm forgetting that HTML5 can do everything webgl can (?!maybe?!)

Isn't Vulcan some attempt at reinventing the wheel with yet another graphics API that is supposed to be really performant and whatnot?

WebGL isn't as fast as desktop APIs and has only a limited featureset, but it specifically has this because it was designed to run on mobile devices just as well as on desktop systems. Even my android brick supports WebGL 1.

You don't need a super advanced high perf API if you make a game with simple graphics anyway. You want something more simple.
You don't want super fancy new tech either, you want stuff that is well supported. WebGL 1 is.

I dunno exactly what html5 contains and whatnot, but I think WebGL is basically part of the html5 "cool new shiny features in the web browser" push.
If you're referring to the html5 canvas API, than no, that is not as potent as WebGL.
The html5 canvas API is a simple "draw me a line" or "draw me a picture" API. Can be good enough for sprite based games, so you could implement something similar to original TA or Starcraft 1 on top of it.
But anything that involves more modern graphics effects, the kind where you do some math per pixel in a shader for fancy effects you need WebGL, as that allows you to make shaders and run them on the GPU. Basically, stuff like this is impossible on the canvas, unless you find somebody with a CPU that is not of this world.

In short: If he intends to do a multi platform in webbrowser RTS then WebGL is probably the only technology that can do it.


However he is not the first to try something like this. There was something called "Atlas RTS" that tried this and they had to move away from it (or gave up on the project completely by now?). There blog posts seem to be gone, but for a competitive MP RTS in the browser targeting Desktop the main problem was latency.
At least a year ago there was no way to access UDP networking from a browser. Everything you can do is based on TCP. Which is too much focused on reliability, causing occasional lags that just broke it for them.

Though they were clearly going the Starcraft-esque route of RTS, so latency was even more important to them than it would be to something like TA/SupCom.

the big difference is that it's cross-OS (and platform) directX is not.

also TIL, thanks.

yep.

I can still play vs the bots on my pc.

atlas is a real damn shame cuz they were promising to make their editor engine free to use.

Not as cross platform as WebGL

yeah it technically doesn't support phones right now... : /

I think it could though