The Potential of POWDER and the Future Prospects of ICRA labels
The concept of labels as a parental control feature has been around since the birth of the Web. Early work at the World Wide Web Consortium [W3C] created the Platform for Internet Content Selection [PICS], published between 1996 - 1997. The system was greeted with enthusiasm by the browser manufacturers with both Microsoft and Netscape including PICS-based content filters in their products that were reliant on ICRA's forerunner, RSACi. Microsoft has retained its PICS-based filter in Internet Explorer since that time, updating it recently in IE7 to make use of ICRA's current system [IE7].
Despite its early promise, labelling and filtering of websites based on those labels has not enjoyed traction and broad uptake. Modern filters are almost all based on large databases compiled by machines that crawl the Web looking for particular characteristics. ICRA labels are very often among the characteristics the crawlers look for but few filters make direct reference to them in their user interface. There are many reasons for this, some policy-related, some technical, but whatever those reasons are, it's clear that the Web has moved on and that ICRA, now part of the Family Online Safety Institute, must move on too.
Actually, FOSI/ICRA is not so much moving on as leading the charge!
In 2004 we began work to update our descriptive vocabulary and labelling technology. This latter initiative was an important input to the Quatro Project [QUATRO] that took up the call from the European Parliament for an 'interoperable system of quality labels.' ICRA played a leading role in that project, after which it chaired the Web Content Label Incubator Activity [WCL-XG] within the W3C and, from February 2007, the full working group for the Protocol for Web Description Resources [POWDER].
Participants in POWDER
A lot more than parental control
The first document to be published by the POWDER Working Group is its Use Cases and Requirements [UCR]. This is an updated and revised version of that produced by the WCL-XG and, importantly, covers a lot of ground. In brief, the use cases are as follows.
Profile matching
A description of online content is matched with the end user's profile. This breaks down into several sub use cases including adaptive search results, mobileOK, Web accessibility and Child protection.
Trustmarks
Assertions by independent authorities that online resources meet defined criteria. Again, this general use case breaks down into sub use cases that vary slightly in the detail of who is making the claim and who is authenticating it.
Rich metadata for RSS/ATOM
This use case stands alone and motivates the easy application of POWDER to these popular publishing standards.
Expressing opinions
POWDER will allow individuals to express their opinions on other people's work, either using controlled, that is, pre-defined, vocabularies, or user-generated tags which, optionally, can then be linked with other terms and definitions. Disagreement between individuals is, of course, expected!
Scalar classification
This use case requires the provision of Description Resources that in saying, for example, that a website is suitable for everyone over the age of 12 automatically also says it's suitable for people over 18.
Expressing Editorial Policy
Although it is implicit throughout all use cases, this one makes it explicit that POWDER must fit in with work flows and editorial systems of large scale content providers.
As is clear, the use cases are very broad. POWDER is about content discovery, personalization and trust; highlighting content that meets defined criteria. Alongside other initiatives looking at the important area of security [WSC], navigating the Web should be more rewarding and a lot safer.
Key features of Description Resources
As defined in the WCL-XG, and emphasized in the revised use cases, Description Resources (i.e. modern labels) have three key features:
- Attribution - who is making the claim and how can it be authenticated?
- Scope - what resources are being described?
- The description itself
There is a fourth key element that is implicit in the name Description Resource - the DR is itself a resource and therefore has its own identity on the Web (it has a URI). This means that the comments can be made about DRs, adding trust in, or detracting from, a particular DR depending on your point of view. These features, plus the easy generation of Description Resources and their support for user-generated tags, means that they're very much part of Web 2.0.
The future of ICRA labels
The Family Online Safety Institute will begin to use POWDER as its primary labelling technology as soon as the working group's documents reach a suitable level of maturity - expected to be around the end of 2007 or early 2008. Like our current labels, DRs will use RDF and there will be a good deal of similarity between the two. However, Description Resources will be better defined and subject to the full rigour of the W3C Recommendation Track [REC].
FOSI fully intends to work with others to make open source software modules available that will process Description Resources. These should be available in at least Perl and Java with support for the current system used by ICRA [RDF-CL] to enable backwards compatibility. These modules will be generic POWDER modules, designed to process DRs carrying any sort of description. They will therefore be suitable for the provision of services that inform end users about the presence of trustmarks and other forms of recommendation, as well as ICRA labels. ICRAchecked will continue to offer an authentication route for descriptions that use the ICRA vocabulary.
Subject to final confirmation, the expectation is that rather than asking each webmaster to host their own labels file, as at present, ICRA Description Resources will be delivered directly from our servers (we do this now with Label Assist, part of ICRAchecked). Webmasters will then include a link to their specific DR in the database. Importantly, other people will also be able to generate ICRA DRs for any website and we have plans to publish aggregated, that is, 'average labels.' Steps will, of course, be taken to counter spurious entries in the database, give different weighting to different labels and to highlight labels created by the webmaster.
FOSI remains 100% committed to the idea that the best person to describe content is the person who created it and the best person to decide what is and isn't suitable is the end user, or, in the case of children, the end user's parents. We believe that POWDER and the developments around its creation, will make that idea a reality.
Links & References
- W3C
- World Wide Web Consortium
- PICS
- Platform for Internet Content Selection
- IE7
- Using ICRA in Internet Explorer
- QUATRO
- QUATRO Project website
- WCL-XG
- Web Content Label Incubator Activity
- POWDER
- POWDER WG home page, includes links to group blog and public mailing list archive
- UCR
- POWDER: Use Cases and Requirements
- mobileOK
- mobileOK is a key use case. It is a trustmark that will allow content providers to claim conformance to Mobile Web Best Practices. Implementations of mobileOK are implementations of POWDER.
- Web accessibility
- Web Content Accessibility Guidelines
- WSC
- W3C Security Activity home page
- REC
- World Wide Web Consortium Process Document. This is the formal document that sets out how W3C arrives at its Recommendations
- RDF-CL
- RDF Content Labels: Schema Description
- ICRAchecked
- ICRA's Authenticated Label Service
