roBman (@nambor)

« Back to blog

A web standards based model for AR

It's great to see so much discussion of an open web based AR bubbling
around. I think it's clear now that there is enough of a framework
across the different web standards that we can map out a clear skeleton
for web based AR. So that's what I'd like to present here. This is the
outcome of an R&D project we've had working on defining a model for web
standards based AR.

The site itself has a few goals:
- visualise how all the standards fit together
- let you see how your browser rates against this ideal
- provide informative links about each of the tests
- open each of the tests for community discussion

Part of this project was also to find a short word or phrase that people
can use to more easily describe what this is. "A web standards based
model for AR" is a bit of a mouthful. And if you say Augmented Reality
instead of AR well then that doesn't exactly roll off the tongue and
into the mind. In aiming this concept at a broader market we've defined
the overall experience simply as "web3".

"But web3 means...?!" I hear you shout 8) I know a lot of people argue
the Semantic Web is web3. And others say the Internet of Things is
web3. And there are plenty of marketing campaign examples in there too.
But I propose that this particular model is a structural change similar
to the introduction of the AJAX based service driven web 2.0 world -
with the natural flow on cultural effects that a structural change
brings. A change big enough that it is a generational step and not just
an incremental improvement.

Have a look and see what you think?

http://isweb3here.com?

It's still a "work in progress" and some of the demos aren't fully wired
up yet. In fact if you can suggest a good javascript game that's
multitouch friendly please let me know. We're also putting together
some more tangible demonstrations of what web3 can do and we'll provide
these as videos so people that can't access these features can see what
they're missing out on.

My hope is that if you are interested in an open web standards based AR,
if you are a developer or a browser vendor or a standards organisation
then you'll join in and contribute to refining this model to the
point that it is a workable and viable reference people can rely on.
Some of the standards that we're currently testing for may change. But
I think the overall functional model will roughly stay the same.  If you have feedback or constructive criticism then please add a comment below, post a comment on twitter with the #web3 tag or post a comment to the ARStandards.org discussion mailing list.

So why do I think this is a significant change?

The iPhone4, iPad2 and many of the Android devices now ship with all of
the required sensors.

- Microphone
- Camera
- GPS
- 6 degrees of freedom gyroscope
- Accelerometer
- Touch screen

And the required core capabilities.

- Hi-res display
- Broadband wireless network
- OpenGLES
- Audio output
- html5

So all the hardware and operating system elements are now easily
available in common consumer level devices. Not just commonly available
but widely adopted too!

Once browsers commonly support the cluster of standards mapped out in
http://isweb3here.com then really integrated innovation can begin. This
innovation can occur at the software level alone so it can occur much
more rapidly. And it can now be developed by the much larger
javascript-aware audience. These are two major multipliers of change.

audiovisualiser is a simplistic example of this
http://files.robman.com.au/audiovisualiser/audiovisualiser.html
Requires Firefox 4+

JSARToolkit is a great example of this.
See this page if your browser supports it:
http://fhtr.org/JSARToolKit/demos/tests/test3.html
See this page for more info: http://fhtr.org/JSARToolKit/

Beatdetektor is an audio example of this.
https://beatdetektor.svn.sourceforge.net/svnroot/beatdetektor/trunk/core/js/beatdetektor.js
http://www.cubicvr.org/index.php?option=com_content&view=article&id=72&Itemid=87

Facial recognition and other CV functions have also been demonstrated.
Service based http://demianturner.com/2010/05/facial-recognition-api/
Browser based http://gigaom.com/video/firefox-demo-facial-detection-within-videos/

In the web3 world, whole new levels of functionality and control can be
established just by visiting a specific web address or physical
location.

Now this is no longer just speculation about the future. All of the
different standards used in the http://isweb3here.com test have been
demonstrated as working in different builds of different browsers...and
your browser is tested live for them just by you visiting the site. All
of the devices listed above have the sensors and capabilities required
to support web3. Now it just seems like a matter of time until they
come bundled in a single web3 experience by default.

I hope this test can speed up this process a little.

So is this another tool like modernizr or caniuse?
No, in many ways this is the complete opposite of that. They are great
examples of developers struggling to deal with the complexity and
fragmentation that is out there in the device/browser market right now.
http://isweb3here.com is an attempt to address the root cause for this
fragmentation. An attempt to abstract this into a more standardised and
open model.

There are lots of API/standards issues to be resolved in each of the
tests. There are lots of security questions and performance
optimisations that must be addressed in general. And there are a huge
number of user experience, privacy and control questions to be resolved
too.

But to me it seems that the overall pathway for web3 is now pretty
clear. What do you think?

Posted June 26, 2011