Re: Categorization + Whether a POI must have location

You'd need to define or leave open space for "locationing data" other
then static co-ordinates.

Non static location/image based positioning is of high interest to
the AR community and a capacity in the spec for it would be most
needed imho. (even if only specific clients make use of it).

On a basic level you could have just a url to a image - telling the
client the POI data is to be positioned relative to every location
where it sees this image. This is pretty easy to add as a spec as its
just a image link....however, its rather unoptimised as it requires
the client process the image into a more recognisable form before it
can compare it to the world.
There may be, for example, data formats specifically for "tracking
points" of which would suit the spec and client engines better. (I
havn't looked)


On 26 April 2011 23:53, Nathan <nathan@webr3.org> wrote:
> Seiler, Karl wrote:
>>
>> I am beginning to feel like this is a recurring and circular discussion.
>> Things (people, staplers, cars) have highly variable locations and places (a
>> park, buildings, a store, a photo spot, a scenic drive) also have low
>> volatility locations.
>> Are we encompassing both or just places.
>
> What additional work or requirements are entailed by also considering a high
> degree of variance in location?
>
>

Received on Tuesday, 26 April 2011 22:05:17 UTC