Cooperation with Mozilla on 3D for the web?
category: general [glöplog]
Hi everybody.
As some of you may already know, Mozilla is working with Khronos on a new low level API for 3D on the web (nicknamed Canvas3D), based on OpenGL ES 2.0. The problem is that we'll have to prove to people that this isn't just another attempt to do what ealier technologies have tried before: Just showing some 3D models. We have to get people excited about what can be done if you have a true low level API.
Typical web developers are not very familiar with low level APIs and what benefits they offer over high level abstractions, so right now the majority of people think along the lines of showing static 3D models and little else. They simply can't imagine putting a low level API to good use to accelerate well, everything: video editing, photo editing, data visualization, dynamic model generation...
So, we have to show them what can be done. We need showcase applications. And we need them to be exciting. And of all people we think that you folks probably know best how to create such applications.
What we would like to know is if there's any interest in working together on this. Maybe a Mozilla sponsored "3d-web" category for some contests? Can you imagine something like this working out? Can you see any problems such a cooperation might face?
You should know that I'm not a Mozilla employee, just an interested user who has talked about this with a few people over at Mozilla and they basically suggested I should ask you guys first. If you know any other places where we should ask for feedback, just post a link.
Thank you.
Hans Schmucker
hansschmucker@gmail.com
As some of you may already know, Mozilla is working with Khronos on a new low level API for 3D on the web (nicknamed Canvas3D), based on OpenGL ES 2.0. The problem is that we'll have to prove to people that this isn't just another attempt to do what ealier technologies have tried before: Just showing some 3D models. We have to get people excited about what can be done if you have a true low level API.
Typical web developers are not very familiar with low level APIs and what benefits they offer over high level abstractions, so right now the majority of people think along the lines of showing static 3D models and little else. They simply can't imagine putting a low level API to good use to accelerate well, everything: video editing, photo editing, data visualization, dynamic model generation...
So, we have to show them what can be done. We need showcase applications. And we need them to be exciting. And of all people we think that you folks probably know best how to create such applications.
What we would like to know is if there's any interest in working together on this. Maybe a Mozilla sponsored "3d-web" category for some contests? Can you imagine something like this working out? Can you see any problems such a cooperation might face?
You should know that I'm not a Mozilla employee, just an interested user who has talked about this with a few people over at Mozilla and they basically suggested I should ask you guys first. If you know any other places where we should ask for feedback, just post a link.
Thank you.
Hans Schmucker
hansschmucker@gmail.com
You can do pretty nifty stuff with Canvas 2D already. IMHO, what's more important to get people interested in Canvas for demos is getting sound support. The best 3D graphics are boring if there's no sound, and having to use flash for that sucks.
Oh, don't worry about that... You can use the video and audio elements as well... in fact it would be great if you did.
http://www.whatwg.org/specs/web-apps/current-work/#video
http://www.whatwg.org/specs/web-apps/current-work/#audio
http://www.whatwg.org/specs/web-apps/current-work/#audio
Given the breadth of creative stuff people are doing with canvas 2D these days (not specifically in the demoscene... the JS demoscene is a pretty lonely place right now), I'm fairly sure that as soon as a real-world non-beta browser ships with 3D canvas support (in a form that's backed by a web standards body), the web development world will be all over it.
My last couple of JS demos (Gallions Reach and Antisocial) have basically been one massive exercise in working around the lack of native 3D support, so when 3D canvas arrives it will make me go either "Hell yeah! Time to get cracking on some *real* demos!" or "oh well, that's my job done then. Time for people who actually know OpenGL to take over." Sorry, I'm fickle :-)
Also: yes, HTML5 Audio support in Mozilla would make me very happy indeed (and stop me from having to faff around with Flash plugins and sandboxing crap).
My last couple of JS demos (Gallions Reach and Antisocial) have basically been one massive exercise in working around the lack of native 3D support, so when 3D canvas arrives it will make me go either "Hell yeah! Time to get cracking on some *real* demos!" or "oh well, that's my job done then. Time for people who actually know OpenGL to take over." Sorry, I'm fickle :-)
Also: yes, HTML5 Audio support in Mozilla would make me very happy indeed (and stop me from having to faff around with Flash plugins and sandboxing crap).
Hehe, you're not the only one who as soon as Canvas2D came out tried to do 3D on it :) http://tapper-ware.net/devel/js/JS.tinyDim/
About Canvas3D... we're really not so sure that web developers will start shouting "hurray" as soon as they see it. There are some of us (you and me for starters) who like to toy with stuff like this, but most web developers have gotten so used to only working with established technologies that they're a pretty tough crowd to convince that something new is worth the trouble of a) learning it and b) figuring out fallback behaviours. We have to throw them a few bones. Show them that there's some real value to this. Only if we reach critical mass this way can we get other browser makers to implement it too and make it a universal solution. The standard is being worked on over at Khronos and it will be an official standard, but just because it's official doesn't mean that it will be universally accepted.
Just to be clear: There will be contests by Mozilla... we can't afford to not promote it: We just thought that it may be easier for demo sceners if we hosted them in your natural habitat. :)
I intentionally left out the technical stuff in my initial post, but I think there's some stuff that may be of interest:
Video and Audio are working on current Firefox and Safari nightlies.... although you'll sadly need different formats as Apple doesn't want anything to do with OGG: You need an MP4 version for them. Both implementations are relatively stable (although probably not this weekend as the media cache patches have just landed in Firefox).
The intersting thing about video, Canvas2D and Canvas3D is how it all ties together: You can take a video, render the frames to Canvas2D, alter them, and finally map them to a texture in Canvas3D!
And finally: If you want to try Canvas3D already, you can find the necessary addon and a rough API overview here:
http://blog.vlad1.com/2009/03/28/canvas-3d-extension-update-2/
About Canvas3D... we're really not so sure that web developers will start shouting "hurray" as soon as they see it. There are some of us (you and me for starters) who like to toy with stuff like this, but most web developers have gotten so used to only working with established technologies that they're a pretty tough crowd to convince that something new is worth the trouble of a) learning it and b) figuring out fallback behaviours. We have to throw them a few bones. Show them that there's some real value to this. Only if we reach critical mass this way can we get other browser makers to implement it too and make it a universal solution. The standard is being worked on over at Khronos and it will be an official standard, but just because it's official doesn't mean that it will be universally accepted.
Just to be clear: There will be contests by Mozilla... we can't afford to not promote it: We just thought that it may be easier for demo sceners if we hosted them in your natural habitat. :)
I intentionally left out the technical stuff in my initial post, but I think there's some stuff that may be of interest:
Video and Audio are working on current Firefox and Safari nightlies.... although you'll sadly need different formats as Apple doesn't want anything to do with OGG: You need an MP4 version for them. Both implementations are relatively stable (although probably not this weekend as the media cache patches have just landed in Firefox).
The intersting thing about video, Canvas2D and Canvas3D is how it all ties together: You can take a video, render the frames to Canvas2D, alter them, and finally map them to a texture in Canvas3D!
And finally: If you want to try Canvas3D already, you can find the necessary addon and a rough API overview here:
http://blog.vlad1.com/2009/03/28/canvas-3d-extension-update-2/