The W3C and CTIA jointly sponsored a stakeholders meeting of the POWDER Working Group at AT&T's Innovation Center in Washington, D.C. on 10th July, 2007. The mobile industry was well represented together with policy professionals from other major telcos, content managers and consumer advocacy groups. This was the first time that the Protocol for Web Description Resources had been presented to a wider audience and was designed to show the new technology, what it can do that can't be done, or can't be done so well, today, and to encourage others to test its capabilities later this year.
The event began with a brief overview of the key features of Description Resources, namely:
Critically, Description Resources (DRs) are resources themselves, they have an identity (a URI) and can therefore be the subject of further statements that support or contradict the assertions. Thus the trustworthiness of DRs can be assessed. Essentially, if you trust the person or organization that created the DR, and can verify that it was indeed created by them, then you can trust the description.
Several of the Working Group members presented their own use case and expectations for future implementation of POWDER. These ranged from mobileOK and improved search, through trustmarks for Web accessibility and privacy to child protection. A number of implementations are under development and these were demonstrated. The MedIEQ project is developing a system for the semi-automated identification of medical information on the Web that can be assessed by qualified experts who can then use POWDER to declare their professional opinion. In another case, user-generated tags are being linked to DRs that give them unambiguous semantics suitable for machine-processing.
The participants in the meeting raised many important questions. Would POWDER be suitable for other use cases such as declaring the copyright protection (or lack of it) on groups of resources? Yes. Does POWDER depend on the willing participation of a chain of people from content provider to end user? Again, largely, yes. Hence the emphasis throughout on the benefits and motivation of self-interest. Content providers that take the trouble to, for example, meet Mobile Web Best Practices or that take significant steps to ensure the security of their customers' privacy should be recognised by end users if not directly then through the tools they use on the Internet.
Another question was "haven't we been here before with PICS and P3P?" There are, of course, distinct similarities with these technologies. All are concerned with user empowerment and all establish a machine-mediated communication channel between the content provider and the end user. POWDER offers richer semantics and greater flexibility than either PICS or P3P, thus, it is believed, making it of more direct use to other players in the chain such as search engines, aggregators, portals and content adaptation engines through which many users access the Web. The cost, ease of implementation and ultimate benefit of POWDER to everyone in the value chain is clearly an important factor in the design of the technology.
The big issue discussed at the meeting though, inevitably, was trust. Can you trust a Description Resource enough for it to be useful to billion-dollar brands? The Working Group members were at pains to make it clear that this was recognised as being a fundamental question and one that lies at the heart of the work. Authentication of the data is clearly critical. Who is publishing the description? Did they really say it? When? Would they still say it now? Much was made of the fact that DRs can be the subject of support and contradiction by users (because DRs have a URI and can therefore be the subject of descriptions themselves). Would a billion-dollar brand trust the wisdom of group think? The answer can only be 'it depends.' i.e. it depends on what the original DR says and the criticality of the claims made. Claims such as 'this is a cool website' and 'this nuclear device is safe' are clearly of a different order and require a different order of authentication. POWDER can express the collected wisdom of Web users where appropriate but, importantly, is designed first and foremost to allow organizations to put their name to descriptions and for others to be sure that those descriptions really are the opinion of the named organization. Therefore the trustworthiness of a DR should be equal to the trustworthiness of the organization that created it.