Re: Is Location Critical to a POI

Hi all,

I've tended to be something of a quiet listener on this WG, mainly because I am currently creating a commercial tool that implements POI on the cloud and connects these to mobile devices, and have therefore had my head down working hard (www.look-here.biz - by no means complete yet...).

Just to give a summary of what I have in my POI technology:

Basic information (title, description, picture(s), sound)
Metadata (key, value pairs)
Location in time and space
Other bits and pieces related to the specifics of the point (such as radius of sensitivity, link to previous and next points, etc)
Ownership info (A layer owner can let others own points, or make the available to the general public).
Layer info (I divide points, for convenience, into layers)
Cost - each point can cost more for the people who lease it from the layer owner.

Each Layer also has similar characteristics (without the location and time).

I have many other aspects coming later, but the key point is that I see POIs as a new way of representing information that is 'all around us', but is also a connection between an advanced program on the cloud and an advanced interactive program on a hand-held device.  POI gives us a chance to use a bit of intelligence to try to reduce the stress of information overload, so in my programming, we strive for Calm Technology and Adaptive and Adaptable Interfaces.

The key is the paradigm shift from 'information on devices' to 'information all around us' where we, as users, don't really care (or even know) where it is stored, and our devices merely become portals - different information may look different depending on what device we use.

Further, there is no reason, in my mind, that a POI can't change its location and, say, move around with a person, or even under it's own programmed paths.

/Roy.
--
--------------------------------------------
Dr. Roy C. Davies, The VR Guy.
--------------------------------------------
Managing Director, LOOK-HERE IP Holdings Ltd.
Consultant and Managing Director, The Flexible Reality Studio Ltd.
Senior Research Fellow, Auckland University of Technology (AUT)

EMAIL: roy.c.davies@ieee.org, roy.c.davies@flexstudio.co.nz, roy.c.davies@aut.ac.nz, roycdavies@mac.com

LINKEDIN: http://nz.linkedin.com/in/roycdavies
MYSPACE: http://www.myspace.com/roycdavies
FACEBOOK: http://www.facebook.com/roy.c.davies
TWITTER: http://twitter.com/roycdavies

SKYPE: roycdavies
MSN: roy.c.davies@ieee.org
ICQ: 2557565
YOUTUBE: drroycdavies

PH: +64 (0)21 795294, +64 (0)9 8338360
WEB: www.flexstudio.co.nz, www.look-here.info, www.humanitycomputer.org, www.forwardthinking.org.nz, www.colab.org.nz

On 19/01/2011, at 6:47 AM, Andy Braun wrote:

> I agree with you. I am  not suggesting a separate definition of a person but rather a location-less POI that is tied by something like image recognition as you suggest. We are currently driving the spec to include a number of building blocks like a location primitive which is mandatory. I think this location-less  POI could cover a number of other scenarios like the how to make a conference call that Jay Wright at Qualcomm mentions in his talks. 
> 
> On Tue, Jan 18, 2011 at 11:26 AM, Thomas Wrobel <darkflame@gmail.com> wrote:
> Surely this would be covered be just letting any data be linked to a
> non-fixed location?
> 
> I certainly see value in tyeing POI information to non-fixed
> co-ordinates, even given by image recognition, markers, RFID or some
> other technology. But I'm not sure I see value it tieing it to a
> "concept" such as a person - because theres no way to locate a person
> without using some other technology.
> 
> Generally speaking, AR is about linking digital information to the
> real world. You can link something to a person already ver, say, an
> address book. But without a way to get this into the real world, its
> not really relevant to AR.
> 
> So I'd say "+1" to tyeing any data to generic moving objects....which
> could include people.
> So that would be [method of identification - image/RFID/marker] tied
> to [data/url etc], where the data or url could describe a person, or
> be a link to a profile online or an existing standard format.
> 
> But I don't think theres any benefit to having a separate definition
> for a person directly in the POI spec, as a lot could be said about a
> person and I think it would be making the POI spec too large to deal
> with it on its own. At the same time specifying a person doesn't help
> a device know when/where to display the information.
> 
> Better to focus on the "link" between the real and the virtual, and
> let the actual data being linked to be anything.
> 
> -Thomas Wrobel
> arwave.org
> ~~~~~~
> Reviews of anything, by anyone;
> www.rateoholic.co.uk
> Please try out my new site and give feedback :)
> 
> 
> 
> On 18 January 2011 17:03, Andy Braun <ajbraun@gmail.com> wrote:
> > I hope all are following the work on the POI wiki, in particular the Core
> > Draft ( http://www.w3.org/2010/POI/wiki/Core/Draft )
> > Clearly location is a critical and identifying piece of most POIs, however
> > there has been discussion around AR related POIs that not tied to a
> > particular location.
> > One such example is that of a person. A common discussed use case is to use
> > augmented reality to attach a person likes or dislike to their physical
> > being rather than just an online profile.  A person can easily be
> > represented by the existing spec with the exception that a person's location
> > is transient (certainly more so then a building). I would ask for feedback
> > from the larger group on how important it is that a POI is tightly tied to a
> > location as I think this is an important AR POI use case to consider.
> > Andy
> >
>