Customisation and Personalisation through RDF
At a meeting in Luxembourg on 4th February 2003, a project with the following aims was launched:
To develop a new RDF application that:
- supports multiple content classification vocabularies;
- can be applied to multiple media types such as text, audiovisual materials and multimedia presentations;
- works with multiple device types such as desktop PCs, handheld devices, mobile devices and a wide range of consumer electronics such as MP3 and DVD players.
- supports both embedded and external markup, whether produced by the content provider themselves (first party descriptions) or by others (third party descriptions).
- provides a foundation for multiple customisation and personalisation strategies.
In short, the aim is to use XML/RDF to create a common platform through which metadata can be fully utilised across all digital media and device types for a variety of purposes. These include but are not limited to:
- parental control using a variety of classification schemes applied through a single delivery method
- automated locating of resources (i.e. search engines)
- educational use – for example the playback of clips from pre-existing DVDs with additional text, such as language variants, to a degree of granularity and flexibility greater than presently supported.
Doesn’t XML/RDF do this already?
Almost, but not quite, no.
An RDF description can be applied to “anything with a URI.” So it is possible to use RDF to describe, for example, a web page, an image, a digital television programme or a computer game. This description might include an existing classification system, such as the ICRA vocabulary or a film classification board’s system, but can also include other metadata such as the author, subject matter, copyright owner etc. Since RDF is based on the ubiquitous XML standard, this is all possible although it is simpler in some areas than others.
What’s missing is an agreed methodology that would bring together standards such as MPEG-7 with HTML web pages, MP3 files with mobile services etc. for these varied purposes. One issue, for instance, is how you link the resource to the description, and how you link a description to multiple resources. The latter case is important. As presently specified, there is no method through which a single RDF description can be applied to the whole of a website or a defined part of it in the way that, for instance, is possible with PICS. How would a digital TV access the metadata? How might it link to the EPG? What delivery method(s) would be appropriate for a small data file that could set a TV, DVD, PC, MP3 player or mobile phone according to the relevant classification system as chosen and controlled by a parent?
IA Japan, W3C etc.
A closely related project was launched by IA Japan in a meeting the previous week in Yokohama. Their aim is to develop and RDF system called “PICS2” and they’re seeking a new Working Group under the W3C. The focus is on fixed and mobile internet.
It is this project’s intention to integrate the two projects and to be fully engaged with the Word Wide Web Consortium. Both the Luxembourg and Yokohama groups are working towards providing material for consideration by W3C for a future Recommendation and clearly these efforts should be unified.
Project deliverables
The Luxembourg meeting identified the following list of initial objectives.
- Devise a workable specification for the linkage of RDF resources and descriptions.
- Explore thoroughly the possible uses of templates in the widest context. How device parameters might be stored and distributed using XML/RDF.
- Review authentication mechanisms defined for XML and assess suitability for use in the project area.
- Develop tools to aid the creation of descriptions, both for first and third party applications, using a combination of AI analysis and human review.
- Translate PICS into RDF. IA Japan’s work is known to be closely related to this – need more information from Yokohama meeting.
- Coordinate efforts with MPEG-7 standard
- Seek support and raise funding for the project.
These objectives should not be seen as the group’s overall objectives – these are starting points from which the group can move towards achieving the aims stated above. Importantly, it is anticipated that the project will demonstrate at least 2 applications using the newly devised system.
Original participants
The organisations who met to begin the project were:
Internet Content Rating Association (ICRA); UK & USA. Phil Archer
Centre National de la Recherche Scientifique (CNRS); Paris. Prof. Gian Piero Zarri
Università degli di Milano (UNIMI); Prof. Elena Ferrari, Andrea Perego
Yale University Law School, CT, USA. Ernest Miller
Brigham Young University (BYU); UT, USA. Prof. Alan Melby
Thus; UK. Clive Feather
National Centre for Scientific Research “Demokritos” (NCSR); Greece. Konstantinos Chandrinos
Additional participation
In order to achieve its aims, the original group will need to be augmented by organisations able to participate actively, bringing additional resources in terms of expertise and administration. Inevitably, some extra funding will also be necessary.
Relevant documents
RDF is defined by World Wide Web Consortium. See http://www.w3.org/RDF/.
There are other relevant activities at W3C, notably Semantic Web, Annotea, Timed Text, and SMIL, details of all of which can be found at www.w3.org.
The methods of linking resources to their RDF descriptions suggested by the project are likely to be informed by the methods adopted for P3P, as defined in section of that standard’s specification. See http://www.w3.org/TR/P3P/#Referencing
For background on MPEG standards, see http://mpeg.telecomitalialab.com/. The TV Anytime Forum’s work is also noteworthy. See http://www.tv-anytime.org/
Contact
For more information, please contact Phil Archer at ICRA.