I use dynamic .so/.dll hot-deploy, which can reload native code in microseconds without disturbing state.
My game can render 600 non-instanced skin meshed characters on a 10W Jetson Nano.
I'm also not trying to say you can't have fast iteration time in other ways. Just that the statement about the browser was incorrect.
You need fast turnaround AND performance at the same time.
You are sinking your cost into a dead-end proving the articles point.
I've been making games professionally across a wide gamut of technologies for over sixteen years and that's simply not true. Further there are loads of games written in JS that prove it's not true! Believe it or not I've actually written game code in languages SLOWER than JS for shipping titles!
Krunker (https://krunker.io/) is a browser based multiplayer shooter.
CrossCode (http://www.cross-code.com/en/home) is another pretty famous one that is also on Steam and very popular.
Slither.io (http://slither.io/) and really the whole genre that these sorts of games became.
The whole Godot Engine IDE (https://editor.godotengine.org/releases/latest/) has a web version.
Game Dev Tycoon (https://store.steampowered.com/app/239820/Game_Dev_Tycoon/) is made with JS.
Curious Expedition (https://store.steampowered.com/app/358130/Curious_Expedition...) was written in CoffeeScript which is compiled to JS.
And so on, and on. There are literally loads including umpteen multiplayer ports of Quake 3 and that sort of thing. Throw in all the JS13k games. Numerous games people have published using the Unity web exporter and so on. All the little web games people have released on itch.io, all the stuff made with Twine and Bitsy. And on, and on.
The idea that JS is too slow to make games with is pretty silly and smacks a bit of willful ignorance.
These prove the point, they don't scale or deliver anything that will dislocate roblox. Even wormate (which is way better than slither does not run properly on my android pad)...
> Javascript is not usable, you wont be able to make anything useful if that is your underlying VM.
And yet all these useful games exist all targeting JS backends. Some of which have made oodles of cash.
Of that list I've been making games for dot big bang for the last three years.
For other slower languages I worked in for shipping games there is EVE Online a large part of which is written in Python and then APB which had a lot of code written in UnrealScript for Unreal Engine 3 both of which are very much definitely slower than JS.
Edit: Another really cool browser based project is this fight simulator: https://www.geo-fs.com/
The V8 is an incredibly good VM. It has compute performance comparable with Rust in Debug mode for a single core, or comparable to the JVM with a single core. That’s plenty for a great many game genres, and for ALL of the game genres that I am personally interested in playing.
CrossCode, Game Dev Tycoon, Curious Expedition, Screeps, FoundryVTT
I feel somewhat confident that there are others, but the devs of these games were somewhat public about what technologies they used. I can easily imagine other very popular games like FTL, Darkest Dungeons, Slay the Spire, etc having made use of browser tech, even though I have no idea if they did or not.
The fact that I bought these games (in some cases on multiple platforms), spent countless hours playing them, and never had a complaint about performance, should be evidence enough that browser tech is plenty good for delivering high quality games that people are happy to pay good money for.
Heck, the fact that it was possible to do this three decades ago ought to be evidence enough that you don't need something better than V8 (or even a potato) to make great games that "scale" in terms of number of players and in terms of purchases.
If you mean VM performance, you can always play the Computer Language Benchmarks Game: https://benchmarksgame-team.pages.debian.net/benchmarksgame/...
You'll find that of course languages like C++ and Rust can perform better. But is that level of performance necessary to pull off a great game with a big paying audience in 2021? Absolutely not.
Eventually you won't be able to afford anything beyond 10W! Javascript is not an option on 10W for action MMO games with real character graphics.
We are talking "metaverse" levels of complexity, and all the games both of you list are very simple compared to that.
It sounds like you are working to enable VR based massively multiplayer games with very low energy/compute requirements. I absolutely applaud you for that. I think it’s awesome. It just doesn’t match your original claim, that performance better than js is necessary for creating a viable game in 2021.
Possibly you meant 3D/VR MMO, but that’s not what I took you to mean when I first responded to your comment about js as a viable engine for communal game building.
I do think your low powered engine project is very cool though.
The 3D action part of this equation is very important, without it you cannot trick the reptile brain that you are actually there, 3D MMORPGs don't trigger this at all, f.ex!
Also if you use that excuse to keep building wasteful software you are not part of a solution.
Likewise I’m skeptical in such an energy scenario we’d have affordable internet access or supply chains to build and equip people with low power computers.
Let’s hope we don’t have to find out!
What I'm saying is that today we use the internet for mostly non interesting stuff when we all can be gods and create interconnected worlds with our minds!
Post link: https://news.ycombinator.com/item?id=28723689
Game link: https://personal-1094.web.app/scummvm.html