Chairs summary of the WAP - W3C workshop on position dependent information services

It is evident that we all consider this topic to be important - and moreover, that in the trials that have been done. users consider it to be important, too. As a mattter of fact, it position information may be paramount in creating contextualized applications, and so be the killer application for the semantic web.

However, today the market is fragmented to destrution. As a content provider, I do not know which way I should turn. The fact that there are 20 different data formats for this information is not alleviated by dressing them all up in XML. We must do something more radical.

I would propose we take a page from the book of the cooperation between the WAP Forum and the W3C: We create a "GML Basic", an XML/RDF module that contains the parameters that fufill the basic needs such as latitude, longitude, time, etc. Other more specific parameters, such as altitude, route, etc can then be added using XML namepaces. We would be independent of positioning method and transport protocol. Merely creating a diagram describing the overlap between organizations would be useful.

But it is evident that geographical data are not "normal" data. For one, they change all the time - and so are more akin to sensor data than static profiles. How to cache them to achieve network efficiency is a protocol question. But they are also descriptions of information objects, which are presented to the end user - wether as route plans or maps, wether as tourist guides or as Chinese restaurant menues - they are all object descriptors, and so metadata. Of course, what is the object and the description will depend on the users view of it. Creating guidelines for metadata deployment is something that will facilitate the deployment and construction of these services tremendously.

At the same time, fragmentation is unavoidable - at the transport level. We want the same dataformat to be used in many different architectures and over many different bearers. This will require different protcols to transport it. We should help organizations like the WAP Forum and IETF how to use and the common data format in their protocols.

A complication in that regard is the privacy - an absolute requirement for the system to be used. But the privacy is intimately intertwined with security - and security is dependent on the transport protocol. Still, techniques like P3P can alleviate this by abstracting the interfaces.

The next step will be to set up a mailing list to discuss a common element sets ("GML Basic") 2D/4D/3D. We should also discuss the overlap diagram and guidelines for metadata deployment.

We will also discuss possible future workshops on the mailing list.

Possible organizations that could participate in this effort are W3C, WAP Forum, EBU, IETF, OGC, ERTICO, ISO TC204. But "invite your friends" - we need the discussion to be as open as possible, involving as many groups as possible.