W3C home > Mailing lists > Public > public-declarative3d@w3.org > August 2011

Re: [AR Standards Discussion] Getting started with the W3C AR Community Group

From: Rob Manson <roBman@mob-labs.com>
Date: Mon, 22 Aug 2011 21:50:42 +1000
To: Philipp Slusallek <slusallek@cs.uni-saarland.de>
Cc: Thomas Wrobel <darkflame@gmail.com>, public-ar@w3.org, "discussion@arstandards.org" <discussion@arstandards.org>, "public-poiwg@w3.org" <public-poiwg@w3.org>, "public-declarative3d@w3.org" <public-declarative3d@w3.org>
Message-ID: <1314013842.31109.1460.camel@robslapu>
A strong +1 to getting to the point where people can just focus on D

I am also a big fan of the Web Based AR running "within" a browser.
This doesn't stop native apps benefiting from the standards like that
from the POI WG and DEC3D...but to me it's the "within" a browser that's
most interesting.  Especially from a distribution/market size point of
view.

That's what I'd like to see the ARCG focus on.

The rest of the standards discussions seem to clearly belong in a cross
SDO group like ARStandards.org.


roBman


On Sun, 2011-08-21 at 21:01 +0200, Philipp Slusallek wrote:
> Hi,
> 
> I might have a very naive view of AR here but to me it consists of four
> main pieces:
> -- (A) Input: Obtaining information about the real world and the user,
> via special devices (position, orientation, movement, audio, video, ...)
> and AR specific processing such as computer vision and others. This
> could be the main area covered by the AR-CG.
> -- (B) Descriptions of links between real and virtual worlds (POI-WG, I
> believe)
> -- (C) Output: Representing a 3D environment (made up of data from the
> virtual and real world) including the interaction descriptions attached
> to the objects and the scene as such. This is what we are targeting in
> the Dec3D CG.
> -- (D) Application: Program logic that takes the input from (A),
> retrieves appropriate virtual data (B), adds some application specific
> data and logic, and displays it with suitable user interactions using
> (C). If (A-C) are doing a good job, a simple AR application can be
> almost trivial.
> 
> Of course, you can make the application logic arbitrarily complex, but
> this should (ideally!) have little to do with (A-C). My point is that in
> the two CGs and the POI-WG we should target to make (A-C) as easy to use
> and as comprehensive as possible (in an incremental way, starting with
> the basic stuff), so that people that want to use AR can focus on (D).
> 
> I am a big fan of creating building blocks and this may be a really nice
> split into separate domains of concern and would clearly outline  areas
> of collaboration between the groups.
> 
> While I can see many AR implementations (Web and non-Web-based), I would
> still argue for the Web browser as our main target for a very simple
> reason: Its an open platform on which we can easily combine our forces
> to make the best and fastest progress. In particular, it has a huge user
> base already and a highly capable runtime environment (it certainly not
> perfect, but should easily cover at least 80% if not more of all use
> cases). With Mozilla and maybe also Google now being interested in Dec3D
> and WebCL getting quite some attention, we seem to be making progress in
> getting the necessary capabilities into the browsers as well.
> 
> Separate non-Web AR browsers would have to replicate some of the same
> runtime capabilities (including JS, Ajax, WebSockets, Worker Threads,
> etc.). I am sure there are people that can do a much better job than our
> current browsers, but I am not sure that the market is there (except in
> niches) to support this next to a quickly moving and broad browser-based
> AR community into which many of the millions of Web developers can
> easily integrate.
> 
> The money should be made with the apps and not with the infrastructure.
> 
> 
> Best,
> 
> 	Philipp
> 
> Am 20.08.2011 18:10, schrieb Thomas Wrobel:
> > On 20 August 2011 16:40, Rob Manson <roBman@mob-labs.com> wrote:
> >> I'm not sure where the discussion around defining a specific
> >> implementation comes from.  Personally, I've never proposed that in any
> >> way and the points both Blair and Thomas make about this seem logical
> >> and obvious to me so +1 to that.
> > 
> > Well, there could be varying ways to do a in-website AR browser, but
> > thats still just one possible way to make a AR browser.
> > Thus if the group was to focus on " Web Standards based model" that is
> > at the very least a sub-set of possible implementations.
> > 
> > eg. If the web model proposes using WebGL, that makes sense for
> > javascript based browsers designed to run on webpages.
> > 
> > However, standalone ar browsers (or hybrid browsers) would have no
> > need of that. They could use DirectX, OpenGL/ES, or any other 3D
> > solution they wish. Dictating they have to use the "web standard" to
> > render their 3d wouldn't serve any benefit as all are capable of
> > producing the same visual result - which is really all that matters.
> > 
> > Thats what I meant by that definition focusing on a specific
> > implementation. I should have really said type of implementation I
> > guess.
> > 
> > Note; I'm not saying theres anything wrong with defining a specific
> > implementation either. At least defining what technologies can be used
> > to make it possible and easy is a must.
> > --
> > The AR field and task is so big subdivision seems sensible to me. So I
> > think each groups goals should be precisely defined.
> > 
> > Theres at the very least in my view a few overall separate tasks;
> > 
> > a) Defining the data standard to store AR data. (that is, the physical
> > links between real and virtual data, as well as a few
> > standard/recommended formats for this data to be in).
> > 
> > b) For web based AR browsers there needs to be a look at precisely
> > what existing things can be used, seeing if they are suitable as they
> > are or need extensions, and if necessary defining new things.
> > 
> > c) Overall promotion and branding of AR, as you say, engage in the
> > larger community.  Theres also issues regarding Patents that could
> > effect AR quite negatively in the future. (Apple recently successfully
> > patenting ARs use on transparent displays, for example, could cause
> > problems for HMDs)
> > 
> > Those are very rough and of the top of my head.
> > There might be more, or different divisions. But really I am just
> > urging precise definition and separation of the tasks that need to be
> > done in different groups.
> > 
> > -Thomas
> > 
> > 
> > 
> > 
> >>
> >>
> >> On Sat, 2011-08-20 at 08:43 -0400, Blair MacIntyre wrote:
> >>> I'd agree with Thomas here;  we clearly don't need yet another group
> >>> of people trying to solve the whole problem.
> >>>
> >>> As an example:  I obviously have an interest in the web spec, since
> >>> that's what we've been implicitly create as part of our Argon work;  I
> >>> would agree that the implementation is a completely separate issue, as
> >>> it's quite easy to imagine very different implementations of a browser
> >>> that render our channels.
> >>>
> >>> BTW, I also think that there should NOT be an all-encompassing
> >>> standard;  building on other W3C standards where ever possible should
> >>> be a goal, I'd think.  For example, 3d data formats are separate, and
> >>> there is no need (at this point) to have a standard.  X3D has not
> >>> gained traction, and there may be other approaches that are lighter
> >>> and may be more suitable for a "baseline".  Similarly, 2D content
> >>> could be adequately handled by HTML5.  There are already working
> >>> efforts for video access, native code and local device access, and
> >>> other issues relevant to AR.
> >>>
> >>> The real question, thus, is WHAT is AR-specific?  That's what the
> >>> group should focus on.
> >>>
> >>> On Aug 20, 2011, at 5:41 AM, Thomas Wrobel wrote:
> >>>
> >>>> Id just point out, if you are focusing on Web-based AR, that thats an
> >>>> AR browser implementation solution - so you shouldn't also cover the
> >>>> standard for the data itself, as they are two very different things*.
> >>>>
> >>>> (Just as HTML specification specifies how html code should be
> >>>> displayed - it doesn't say what languages and technology's the browser
> >>>> should use to do that. Browsers can thus be coded in many languages,
> >>>> and use all sorts of techniques to display the same results. AR
> >>>> browsers should be the exact same).
> >>>>
> >>>> The discussion of the data standard and code to display that standard
> >>>> are thus two separate discussions, and the goal should be quite
> >>>> explicit on which it aims to do.
> >>>>
> >>>> [/2 cents]
> >>>>
> >>>> -Thomas
> >>>>
> >>>> * with the possibly exception of the 3D format, as web-based tech
> >>>> would limit that to certain types, while non web based browsers could
> >>>> support anything. Thus the non-ones should conform to the web standard
> >>>> 3D anyway. (which I think was more heavily towards being X3D - which
> >>>> as long as it serialises nicely I see no downside to using in any
> >>>> scenario). In either case, this would be a job for the data-standard to only
> >>>> choose formats both lisence free and suitable for web use.
> >>>>
> >>>> On 20 August 2011 04:43, Rob Manson <roBman@mob-labs.com> wrote:
> >>>>> Hi all,
> >>>>>
> >>>>> the W3C AR Community Group has been established and is now open for
> >>>>> people to join.  Great work on proposing the group Ya Knygar.
> >>>>>
> >>>>> Now I think it would be good to make some clear plans about what the
> >>>>> goals of the group are and what the scope of our activities are.
> >>>>>
> >>>>> From my perspective this would simply be:
> >>>>>
> >>>>>        "The development of a Web Standards based model
> >>>>>        for Augmented Reality"
> >>>>>
> >>>>> If you have a proposal for an alternate goal/scope then please submit it
> >>>>> and we can run a poll to select what the group runs with.
> >>>>>
> >>>>> Also, I don't think this group is going to work if we just automatically
> >>>>> make everyone who joins a co-chair 8)  At the moment everyone who has
> >>>>> signed up has been made chair.  I'd rather see us first establish the
> >>>>> goals for the group, then run a poll to decide how the group will be
> >>>>> managed and who the chair/s are.  We don't need to be too formal...but a
> >>>>> little structure would be good I think.
> >>>>>
> >>>>> We will also need to clearly define how this groups is different from
> >>>>> the existing AR related groups that have formed already.  I think the
> >>>>> goal I've proposed above does that (e.g. focus solely on Web Based
> >>>>> AR) ...but more discussion is obviously required.
> >>>>>
> >>>>> So, please join the group and get involved in this important discussion.
> >>>>>
> >>>>>        http://www.w3.org/community/ar/
> >>>>>
> >>>>> There's a lot happening and a lot of APIs that will directly impact the
> >>>>> future of a Web Based AR are being defined right now. So now is the
> >>>>> perfect time to get this up and running.
> >>>>>
> >>>>> roBman
> >>>>>
> >>>>> PS: I've cc'd all the related groups I'm involved in to encourage anyone
> >>>>> with a stake in related technologies and APIs to join this group.
> >>>>>
> >>>>> PPS: I've also cc'd in the W3C Community people as I think this
> >>>>> discussion is as much about Community Group process as it is about the
> >>>>> content of our group.
> >>>>>
> >>>>>
> >>>>>
> >>>>> _______________________________________________
> >>>>> Discussion mailing list
> >>>>> Discussion@arstandards.org
> >>>>> http://arstandards.org/mailman/listinfo/discussion
> >>>>>
> >>>> _______________________________________________
> >>>> Discussion mailing list
> >>>> Discussion@arstandards.org
> >>>> http://arstandards.org/mailman/listinfo/discussion
> >>>
> >>>
> >>
> >> _______________________________________________
> >> Discussion mailing list
> >> Discussion@arstandards.org
> >> http://arstandards.org/mailman/listinfo/discussion
> >>
> > 
> 
Received on Monday, 22 August 2011 11:51:06 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 22 August 2011 11:51:07 GMT