MMOs ที่ใช้เบราว์เซอร์ (WebGL, WebSocket) [ปิด]

คุณคิดว่าเป็นไปได้หรือไม่ในทางเทคนิคที่จะเขียนไคลเอนต์ MMO สามมิติที่มีคุณสมบัติครบถ้วนด้วย Browser JavaScript - WebGL สำหรับกราฟิกและ WebSocket สำหรับระบบเครือข่าย?

• คุณคิดว่า MMO ในอนาคต (และโดยทั่วไปเกม) จะเขียนด้วย WebGL หรือไม่
• ประสิทธิภาพ JavaScript ของวันนี้อนุญาตหรือไม่
• สมมติว่าทีมพัฒนาของคุณคือคุณในฐานะนักพัฒนาและผู้สร้างโมเดลอีกคน (ศิลปิน) คุณจะใช้ไลบรารีเช่นSceneJSสำหรับเกมหรือเขียน WebGL โดยตรงหรือไม่ หากคุณต้องการใช้ห้องสมุด แต่ไม่ใช่ SceneJS โปรดระบุว่า

UPDATE (กันยายน 2555): RuneScape ซึ่งเป็น MMORPG ที่ใช้เบราว์เซอร์ 3D ยอดนิยมที่ใช้ Java Applets จนถึงขณะนี้ได้ประกาศว่าจะใช้ HTML5 สำหรับลูกค้าของพวกเขา ( แหล่งที่มา )

Java (ซ้าย) และ HTML5 (ขวา)

อัปเดต (มิถุนายน 2013):ฉันเขียนต้นแบบต้นแบบของ MMO ที่ใช้ WebGL / WebSocket: https://github.com/alongubkin/xylose

Do you think it is technically possible to write a fully-fledged 3D MMO client with Browser JavaScript - WebGL for graphics, and WebSocket for Networking?

Yes, absolutely. There is no reason WebGL or WebSocket technology would prevent you from making a 3D MMOG client, or any game client for that matter.

Do you think future MMOs (and games generally) will wrriten with WebGL?

Yes. I believe within the next five years, most 3D browser games will be written using WebGL. The reason is simple - WebGL is the only standardized 3D technology that will have implementations available in every major web browser (Chrome 9, Firefox 4, Safari 6, and Internet Explorer via Chrome Frame).

Does today's JavaScript performance allow this?

Yes. JavaScript performance in modern browsers has increased to the point where 3D game development is feasible. For example, see the Three.js project.

Let's say your development team was you as a developer, and another model creator (artist). Would you use a library like SceneJS for the game, or write straight WebGL? If you would use a library, but not SceneJS, please specify which. Thanks!

Use a library to save time. There is no reason to write your own WebGL graphics code unless an existing library is missing features that you need. Even in that case, it would probably be more time efficient to extend the existing library.

For my project, I am using GLGE as it supports many different graphics effects and is constantly being updated with new ones.

There are already several Browser based MMOs out there. They usually don't use 3D graphics though.

I'm a bit sceptic when it comes to WebGL. The current crop of browsers don't support WebGL in their normal release builds. You'll have to get special builds or mess with config files, which is more complicated to the average user than downloading a plugin.

Speaking of plugins: Several people mentioned Unity, but don't forget Flash. An upcoming release of the Flash player (code-name molehill) will have GPU-accelerated 3D. My guess is, that flash will provide hardware accelerated 3D graphics in the browser before WebGL takes off.

Technically possible? Yes. But why bother when web solutions like Unity are available?

Future MMOs written in WebGL? No. Ok maybe but there are web based plugins like Unity that do fine right now.

Today's JavaScript performance allow this? No. Ok yes if you keep your game simple.

Would you use a library like SceneJS [...] or write straight WebGL? Neither. I would use a third party application such as Unity.

This is not feasable right now (early 2011). You can not create a real-time, 3D MMO in native, browser technologies (meaning no plugins).

I won't speculate about the future. This is how it looks right now. Assuming you target only modern browsers (IE9, FF4, etc). If you plan to support IE7 then you are crazy. That will never work.

Things missing:

• full control over keyboard and mouse (ability to set mouse position, key interception)
• fullscreen
• 3D

Areas / solutions with serious problems:

• WebSockets now deactivated in the few browsers that supported it
  • without it: now low latency, bi-directional client-server communication
• audio problems
  • codec mess, you need ogg and mp3
  • common frequences not supported
  • concurrancy
  • delay
• canvas performance (in some cases doing divs & css transformations is faster)

Yes, but with major caveats.

You won't have full keyboard control, and for many control freaks, this could be a deal-breaker. The F-keys for example will likely remain out of reach for quite some time.

You're not going to be able to hit most browsers, either. WebGL isn't really well-supported yet, and websocket support is hit-and-miss these days, with some browser vendors enabling it for one version, then disabling it for the next.

But you can put together something simple, sure. Throw in some compelling gameplay and it won't matter that the graphics are a little low-key. Runescape started off pretty simple, and grew into a complex, fun, incredibly popular MMO, back when everyone said that an MMO created in a browser plug-in was a pipe dream.

An MMO built withing the current limits for browsers that adapts as things change would certainly have a leg-up by the time that browsers are far enough to support a "AAA" MMO.

Absolutely, yes.

WebGL is included in nightly builds of both WebKit (Chrome/Safari) and Gecko (Firefox). WebSockets is currently disabled due to security issues, but there is no reason why it shouldn't be re-enabled once the issues are resolved. In the meantime there is always HTTP.

Will all games be written this way? No, but a significant number will be. Once WebGL is in final versions of Chrome and Firefox the penetration will already be higher than that of native plug-ins like Unity.

JavaScript performance is absolutely up to writing games. You probably won't see a Crysis level of graphics for a while, but remember with WebGL, the number-crunching is offloaded to the GPU.

For a small 2-man team you will need to be prepared to do a lot coding which ever engine you use. There currently aren't any WebGL/JavaScript engines that actually cover much of the game code. SceneJS, three.js and similar are wrappers over the graphics layer, but provide nothing for collision, physics, audio, resource-loading, network or tools which make up the bulk of a product like Unreal Engine, or Unity.

It's early days at the moment, I'd expect to see more middleware emerging over the next year.

I don't think many future MMOs will be written with any client-side browser technology since most MMO developers will want to use the same language across the client and the server, and most game developers are fluent in non-web languages like C++. There's far too much client-code involved to make it very practical in Javascript anyway, except for quite trivial simulations. There are also other limitations imposed by running in the browser which are unsavoury for a typical game developer.

I'm sure there will be some games like this, however. They're just not going to take over from WoW or similar games for quite a while. They will probably be made more by web developers getting into games rather than games developers getting into the web.

The question you should ask yourself is why will game developpers want to depend on web technology they do not master and have no control over, current langages allows you even in a web browser to make awesome games ?

Yes, but the technology still needs to stabilize and there will likely remain performance trade-offs vs native applications.

Unity is good. See also Shiva and the Prime Engine.