TIGSource Forums

O3D (http://code.google.com/apis/o3d/) is a Javascript API for creating 3D applications in the browser.

I've been following this project since the first official release, but I've started to "play" with it just today.
My first impression (as developer) is that the technology is interesting, especially for demos/prototypes, but it needs a better documentation/some tutorials and google should start to push it with some kind of marketing or it will remain just one of the many projects in google code.

Have you tried it? Any comments?

I ran into this this other day.  ...Talk about under the radar: all the top google results are actual pages related to the software, instead of articles on tech blogs.

I found the demos I tested were really slow, but that could be my connection speed too.  I think with some serious work this could end up pretty good, but yeah, they need to put a lot more into it.

I think a new browser plugin needs a lot of momentum to really take off.  If most people would have to install a new plugin to play your game, you probably should have just made it stand-alone.  So, yeah, hopefully if google is serious they will push it harder ... I hadn't heard of it at all before now.

3D in browsers generally requires plugins though, right?  I'm pretty sure I had to install one to play Unity games on Blurst.

Why choose O3D rather than unity or java ?

Yeah, I think that's what we're all saying.  It needs to be made better/given standout features.

I'm not sure about that... if somebody makes a really cool game, downloading a 1.5Mb plugin and installing it with 2 clicks is not a big deal for playing it, and you need to install the plugin once.

Maybe a problem for most developers could be the inevitable open-source model required by JS.
Anyway, if you consider it more like a tool for demos/prototypes than a real plugin for web-games, this is not a problem as well, from this point of view at least.


There are some pros like:

• it's free
• it runs on Windows, Mac, Linux
• it's not Java :)

Of course there are cons too, probably the lack of marketing/diffusion/interest is the main one at the moment.

I'm not sure
...are really valid arguments against Java. I don't know what your fundamental problem with it is, but y'know...it's a solid, well established programming language. I'd rather program for a JVM than a new javascript driven plugin. But yeah....hopefully there will be -some- open, established standard for web-based 3D soon!

I thought "zero install" was a supposed advantage.

O3D just makes me think of VRML.

In fact those are against unity... maybe next time I'll post two separate lists with a header :)
Anyway, it's a matter of personal preference, you prefer coding for a VM, I prefer coding for myself ;)


I think flash and java are the best demonstration that installing a plugin is not such a big deal.

Anyway, these are just personal opinions, I'm not trying to sell O3D here, I just wanted to discuss about the technology with other developers.

Hmm, I personally don't install plugins unless there's a really compelling reason to -- I still haven't installed silverlight and I don't forsee installing this one.  So, it's definitely an issue for me.  Perhaps it's less so for internet users 'in general' though.  Still I think there's quite a big difference experience-wise between 'user loads up a page and the game is ready to play' and 'user loads up a page and it asks them to install a plugin they've never heard of'.

Hehe, sorry ;p

I don't necesarily prefer coding for a VM, it's just that it's a lot more established and available than a fledgling plugin. I'm all for a more unified 3D-in-browser, but if I were to work on a project with this in mind, I'd rather ensure a more wide-spread user-base. But yeah...go Google.

Actually, what do you mean by 'coding for a VM' vs 'coding for myself'?  I'm not sure that's a coherent distinction.

(sounds kinda like ... programming your own brain ...)

No problem dude, I was joking  ;)


It was a kind of joke, cause Schtee in his first post wrote "I'd rather program for a JVM", but, I repeat, just a joke, don't take it too seriously :P

The neat part looks to be that it's programmed in JavaScript. Which is cool, because lots of people know JavaScript, and it fits with the whole webpage thing.

Though JavaScript in a VM is the way the language seems to be headed.

JamesGecko: Google's "Chrome" already uses a full JIT VM for Javascript, and Firefox will do the same whenever they get around to adding "SpiderMonkey" to the main release (I think this will happen when 3.5 is released). There's actually even a Javascript bytecode format defined now, called ABC, but I'm not sure if anyone's using it.

Yes, but I think people most often don't play 3D in browsers for this exact reason...

Also incidentally apparently Apple has added a 3D-in-Javascript API to Safari, but only for the iPhone?! (http://arstechnica.com/apple/news/2009/03/apple-holding-back-on-web-based-3d-graphics.ars)

Yes, and between them and some other ideas I hope an actual standard comes out of this rather than a thousand competing plugins.

VRML wasn't the worst idea but the hardware and infrastructure for web-3d didn't exist back then.  I think that most PCs have the bandwidth and horsepower for this to develop now.

Also I vote Adobe Atmosphere as having the worst possible 3d controls ever for such an idea.  Bad Adobe.  Wait, I take that back because that google virtual IM thing was even worse.

I've started to publish some basic examples on my website (http://www.m3xbox.com/index.php?page=c_o3d), they are very simple and loosely follow the NeHe lessons.

Another source for few examples is www.o3dexamples.com (http://www.o3dexamples.com/).

And, according to the latest blog post (http://o3d.blogspot.com/2009/05/gallery-for-your-samples-and.html) from the O3D authors, Google guys are setting up a gallery for samples and apps that should be similar to ChromeExperiments.