- From: Jens de Smit <jens.desmit@surfnet.nl>
- Date: Mon, 23 Aug 2010 11:30:19 +0200
- To: Thomas Wrobel <darkflame@gmail.com>
- CC: "Hermodsson, Klas" <Klas.Hermodsson@sonyericsson.com>, Henning Schulzrinne <hgs@cs.columbia.edu>, 전종홍 <hollobit@etri.re.kr>, Matt Womer <mdw@w3.org>, "public-poiwg@w3.org" <public-poiwg@w3.org>
On 21/08/2010 22:08, Thomas Wrobel wrote: > 2010/8/20 Hermodsson, Klas <Klas.Hermodsson@sonyericsson.com>: >> On Aug 20, 2010, at 9:14 , Jens de Smit wrote: >> >> On 19/08/2010 09:49, Hermodsson, Klas wrote: >> I think two levels (i.e. [criteria]<>[data]) is too simplistic. I would like to see a three level approach: >> [criteria]<>[representation]<>[actual data] >> >> I'm not really seeing this (yet). They way you put it, isn't the >> representation implicit in the type of data that's being linked? As in, >> if the [actual data] is X3D we're dealing with a "visual" representation >> (of the subtype "3D model") and if it's an OGG container with a Vorbis >> stream inside it's an "aural" representation, etc. >> >> I may be using some terms that are not really suitable above. Let's take a concrete example to illustrate: >> >> - Company A has a sign with their logo on outside their stores >> - When this logo is detected the company wants a spinning sphere with the logo on to be displayed while a music piece is playing >> - If you select/activate/click this spinning sphere the latest ad is played back as a video >> >> Criteria: if computer vision detection of the logo occurs (criteria expressed in suitable markup language) >> Representation: a spinning sphere + music (layout and resources of this representation expressed in suitable markup language) >> Actual data: the video ad (content in specific format stored reached through some URI and over suitable protocol) >> ...Maybe what I call representation is what other people call data? Note that both representation and data may need "layout" markup to explain how it should appear in our real world. > > Yes, I see both of them as data. The "representation" is, presumably, > a 3d file set spinning and a wave file of some sort. > When I say [data] <> [criteria] I am enclosure of those formats. > > The fact that that data can, in tern, act as a trigger to further data > (when the user clicks it) imho doesn't change the fact the original > information popping up is also itself data formats that have been > linked to the trigger. So to me your scenario is two separate > [data]<>[criteria] associations. > > I think the difference here is the first (the automatic appearing of > the sphere+music), is a passive/auto triggered event, and the "user > selects" is a manual one. > You could just as easily have a video file as the "representation" in > this scenario, that then pops up a big 3d file for the "data". So I'm > not really seeing the separation myself. I fully agree with Thomas here that these are two separate [data]<>[criteria] associations. In this use case they happen to be related in some way but they could just as well occur separate from eachother. > I do think there is variably presentation issues, but I see this akin > to feeding device's different data based on their specifications. Akin > to the @media in css. So, as part of the criteria, we should also be > able to specify its intended for a certain device. Having multiple representations associated with a single entity where the use agent selects the most appropriate to display is occuring in a lot of places. An ancient example is the "alt" text attribute for the <img> tag, more recent ones are the built-in support for multiple formats in the <audio> and <video> tag and the use of full-detail and reduced-detail 3D models in Layar. With the amount of (current) variety in input and output capabilities of networked devices this seems like something we should consider as well (but I think we already had some consensus on that :) Regards, Jens
Received on Monday, 23 August 2010 09:30:52 UTC