Protocol for Web Description Resources (POWDER): Primer

W3C Working Group Note — 03 April 2009

This version:
Latest version:
Previous version
Kai Scheppe, Deutsche Telekom AG


POWDER — the Protocol for Web Description Resources — provides a mechanism to describe and discover Web resources and helps the users to make a decision whether a given resource is of interest. There are a variety of use cases: from providing a better means to describing Web resources and creating trustmarks to aiding content discovery, child protection and Semantic Web searches.

There are two varieties of POWDER: a complex, semantically rich variety, called POWDER-S, and a much simpler version, just called POWDER, which is intended as the primary transport mechanism for Description Resources. POWDER-S can be generated automatically from POWDER.

Status of this Document

This section describes the status of this document at the time of its publication. Other documents may supersede this document. A list of current W3C publications and the latest revision of this technical report can be found in the W3C technical reports index at http://www.w3.org/TR/.

Publication as a Working Group Note does not imply endorsement by the W3C Membership. This is a draft document and may be updated, replaced or obsoleted by other documents at any time. It is inappropriate to cite this document as other than work in progress.

This is the third publication of the document as a Working Group Note by the POWDER Working Group

To send comments, please use the public mailing list (public-powderwg@w3.org), an archived mailing list. See W3C mailing list and archive usage guidelines.

This document has been produced as part of the Semantic Web Activity, following the procedures set out for the W3C Process.

This document was produced by a group operating under the 5 February 2004 W3C Patent Policy. W3C maintains a public list of any patent disclosures made in connection with the deliverables of the group; that page also includes instructions for disclosing a patent. An individual who has actual knowledge of a patent which the individual believes contains Essential Claim(s) must disclose the information in accordance with section 6 of the W3C Patent Policy.

Table of Contents

  1. What is POWDER?
  2. Why would I want to use POWDER?
  3. How does POWDER work in the real world?
  4. How do I use POWDER?
  5. What do I need to use POWDER?
  7. Examples
  8. Acknowledgements
  9. References

1. What is POWDER?

POWDER is an abbreviation for "Protocol for Web Description Resources." The goal of the working group has been to develop a mechanism that allows not only the provision of descriptions but also a way to apply them to groups of online resources and for the authentication of those descriptions.

The primary 'unit of information' within POWDER is the Description Resource (DR), which comprises:

There are two varieties of POWDER:

The simple version has relatively loose, human-readable 'operational semantics,' and is written in XML.

The semantically-rich version, known as POWDER-S, allows POWDER to harness the Semantic Web at large and encodes formal semantics that underpin the operational semantics.

There is a third, transitory version of POWDER, called POWDER-BASE, which is intended for processors and is not of concern for this document. A "Gleaning Resource Descriptions from Dialects of Languages" (GRDDL) transform, may automatically generate POWDER-S as RDF/OWL from a POWDER document. The details of this transformation are defined in the [FORMAL] document. XSLTs are available to support this. Alternatively, POWDER-S may be written directly.

There is no restriction on which form to use, but it should be noted that the simple version is intended as the primary exchange mechanism between systems. All POWDER tools should process POWDER. Using the POWDER-S form is optional, so that a processor may not necessarily understand this form.

POWDER-S is designed to facilitate incorporation of POWDER information in larger RDF-based systems and it should be noted that such systems will need to implement a Semantic Extension to do this (see How do I process POWDER-S).

Importantly, POWDER allows a variety of questions to be answered about a given Web resource or group of resources, without having to actually retrieve and inspect the resource(s).

At first a Description Resource is simply a claim: somebody is making some statement about a given resource, or group of resources. However, most users would have to trust the person that made the claim before deciding whether to trust the data. If a DR is made available directly by a well-known content provider that is trusted to uphold a certain level of quality, then the data might readily be trusted. However, this will not always be sufficient. Since a DR may be published by anyone, anywhere, to describe anything, an end user may reasonably want to query the cited author of the DR to ask questions such as: Did you really make that claim? And, if so, when? Would you make the same claim today?

For some situations this might still not be sufficient for the end user. To facilitate the further extension of trust a means has been provided to allow certification of DRs. A Description Resource that has been certified immediately gains in trust, through the verification by a third and trusted party of the original claims made by the DR author.

Through the combination of these tools various questions can be answered using a Description Resource, without having to retrieve the resource itself.

The following are examples of questions that could be asked using POWDER:

The POWDER Suite consists of the following documents, in order of recommended reading

  1. POWDER: Primer - this document.
  2. POWDER: Use Cases and Requirements - the premises upon which POWDER was created.
  3. POWDER: Description Resources - the definition and structure of Description Resources (DR).
  4. POWDER: Grouping of Resources - the method by which sets of resources may be defined to which DRs apply.
  5. POWDER: Formal Semantics - a detailed description of the formal semantics of POWDER-S and of the GRDDL transformation process.
  6. POWDER: Test Suite - provides data against which conformant POWDER applications may be tested.
  7. POWDER: Web Description Resources (WDRS) POWDER-S Vocabulary - the definition of the RDF vocabulary that is used to express POWDER-S.
  8. XML Schema for POWDER and POWDER-BASE - the namespace document/schema for POWDER

There are a variety of tools available to aid in the implementation of POWDER.

  1. POWDER Validator
  2. POWDER Processor - PERL based
  3. POWDER-S Processing kit
  4. POWDER Processor - PHP based
  5. TransOnto - a semantic POWDER Processor
  6. POWDER to POWDER-BASE transformation tool
  9. POWDER Grouping Tester

Details and newest versions can be found at the POWDER home page.

2. Why would I want to use POWDER?

The amount of information on the web grows continually. Conversely, the amount of time people have to devote to this information seems to be decreasing. To cope with these constraints (too much information and too little time), users need a system where more customized content is delivered to them in the most personalized way. Basically the average user wants to get online, find what they want and move on.

From an End User's Perspective, POWDER:

From a Publisher's Perspective, POWDER:

From a Service Provider's Perspective, POWDER:

Use Cases That Have Driven the Development of POWDER Include:

POWDER offers a much more dependable means by which to deliver the most relevant information without putting the onus on the user to verify and validate every aspect. Both information providers and information consumers are interested in getting the highest return for their individual investment of time and effort.


This is one of the most powerful overarching features of POWDER and one that has been an unresolved challenge until now. It allows one to offer semantically rich information about entire groups of online resources with both flexibility and precision in the way the group is defined. A processor may use a single POWDER document to extract information about many - perhaps huge numbers of - other resources. In addition, the maintenance of such Description Resources is simplified by support for defining many descriptions in a single document.

The methods for grouping resources range from the individual listing of IRIs, through the specification of things like domain names and paths, to the matching of IRIs against regular expressions. Requests made to a POWDER Processor, however, are always handled at the level of an individual resource. The RDF returned will be triples about that resource, allowing an application to analyze the data and decide how to act case by case.

N.B. POWDER uses the term IRI (Internationalized Resource Identifier [IRI]). The more common term URI is a subset of IRIs, so that all URIs are also IRIs.

Data retrieval efficiency

Traditionally, meta information is always linked to a single resource and is usually embedded within it, as is the case, for example, with the HTML META element where the attributes are chosen and the values are given by the author. When metadata is embedded in content in this way, the complete resource has to be retrieved in full to then determine if it is of interest or not.

POWDER describes online resources that may be of interest to a user or a system. The keyword is "may", because the requester can decide prior to resource retrieval whether the resource should be retrieved at all, based on information provided by a Description Resource. This makes information retrieval more efficient and precise by reducing network traffic and server load. From a user perspective it means greater personalization and less irrelevant content.

Profile matching

POWDER allows profile matching - that is, the retrieval of resources according to user preferences, device capabilities and current state at the time of content delivery. The use cases [USECASES] deal with adaptive search results based on context, suitability for mobile devices, functional user experiences based on capabilities, web accessibility, child protection and privileged content.


Resources and the DRs that describe them may be connected to each other in a web of trust. Partners in this web may be certification authorities that verify the truthfulness of claims made in Description Resources. Other partners may be repositories of thematically linked URIs, such as white lists for web sites that are recommended for children.

There are several possible models in which assertions and claims can be made, authenticated and reported to the end user. The Use Cases and Requirements document lists several use cases which have a number of elements in common but differ in details such as whether it is the content provider or the trustmark operator that makes the original claim, whether the data is stored on the trustmark operator's servers or alongside the content itself, and whether the trustmark operator provides the description or the authentication for a description.

Semantic Annotation

POWDER makes it easy for annotations to be created and published independently of the relevant content. Such annotations may cover large amounts of content as it is created with only occasional changes needed to the annotations. This matches typical content production workflows where different individuals are often responsible for the two tasks. In situations where annotation can only be done by the content author, all too often content is published without any meaningful metadata.

Cost-effective, trusted annotations have a significant potential benefit to content aggregators, search engines and related services.

Semantic Searches could return results filtered for location-based cultural parameters. Information provided on the Web can affect users in different ways depending on the context in which it is retrieved. For example, nudity on a medical website dealing with breast cancer may be fully appropriate if it is known prior to retrieval what the context of the content is. Content promoting discrimination may be appropriate for analysis purposes but may be served with additional contextual information in other circumstances. Semantic annotation also allows the disambiguation of terms such as 'football.'

Furthermore a POWDER document can be used as a set of instructions that allow a crawler to follow up on all links, pointing to DR in POWDER documents, to automatically create a triple store about those resources. This could be used in conjunction with XML-based site maps to create triples by processing those site maps.

In summary, POWDER empowers users to find more of what they ask for and less of what they don't.

3. How does POWDER work in the real world?

The previous section describes the benefits of POWDER and the reasons why it is a compelling method for defining relatively small amounts of data that can be applied to large amounts of content.

The following are brief examples of how POWDER applications work.


POWDER offers various possible methods by which trust claims and assertions can be made and used.

Visual notification

A user could install a browser plug-in designed to provide an indication (such as a visual logo or audible alarm) that a Web site is trustworthy or not. The plug-in would rely on various sources such as reputation and accreditation services to determine trustworthiness.


This can be particularly valuable in cases where certain sites are required by law or other regulations to provide a certain level of content or access. Once those sites have been approved and tagged as having met specific standards, an automated process can aid the discovery and regular review of the site's content and its DR for continued authenticity and validity. This could go as far as generating a report for the evaluating party.

Description authentication

A Web site operator submits their work to a trusted organization that can authenticate the work and provide a Description Resource that identifies the site as accurate or trustworthy.


POWDER enables search engines and portals, if they so choose, to provide customized links for users who prioritize compliance with particular accessibility checks. For example, an organization might review Web sites to determine their level of compliance with accessibility guidelines [WCAG]. The reviewed Web sites can be duly tagged as compliant using a DR supported by a test result expressed in [EARL]. The search engine/portal can then present or perhaps highlight the best sites based on a user's preference settings and the tags on the reviewed Web sites. The system can be highly granular so that the links presented to a user with limited manual dexterity (to resources that support keyboard navigation) would differ from those presented to a user with a visual impairment.


In a similar fashion to the accessibility case, POWDER may be used to identify resources that conform to W3C mobileOK [MOK]. A user wanting to browse the Web using his mobile phone may request that the search provider favor links suitable for display on his device. When the search engine retrieves a set of links from its index, it determines which have an associated mobileOK descriptor, and presents those before the remainder.

[Example: mobileOK]

Child Protection

Online child protection, as well as the continuation of offline child protection, is a priority for any responsible site or service provider, whether directed at children or not.

A service provider may include a feature that offers parents the ability to determine the type or level of content they would allow their child to view. In today's market, most, if not all, service providers do offer some degree of parental controls. A hypothetical Web site that sells an array of merchandise can use POWDER to accurately describe their content as either being child appropriate or intended just for adults. While most of their content and products are appropriate for all ages, there are also numerous pages showing "adult" toys. When a child whose setting allows him only to view content that is age appropriate accesses this online retailer through the family network he will only be able to view the content that is deemed age appropriate and not that in the adult category.

[Example: ICRA Labelling]

Functional User Experience

Web pages and whole Web sites containing any type of rich assets such as video/streaming video or audio can be tagged with that information using POWDER. A search engine, content aggregation or adaptation service can then determine whether a user is accessing content via a low or high bandwidth connection and return only those pages that contain assets and images that will be supported by that user's connection speed.

Privileged Content

A user pays an extra fee to his ISP in order to have privileged access to third-party premium content. When he accesses a premium page on one of these third-party Web sites via his ISP, the server is able to recognize him as a paying customer and deliver the content that has been described as premium by an associated Description Resource.

Additionally in that fashion a user may be informed that he will not be able to access said content because he is not a paying customer.


Sites can use Description Resources to communicate the subject matter of their content more accurately and with greater relevance to user requests.


Because the same term can suggest two distinctly different meanings, the judicious use of Description Resources can go a long way in making a site more valuable to a greater number of people. An owner of a global sports site can provide descriptors that are accurate enough so that users searching on the term "football" can receive the portions of the available content relating to the specific game that the user's location suggests is what they are searching for information about.

Distinguishing opinion

By developing an appropriate and detailed vocabulary, Description Resources can be used to identify the value of a site as being either of value in its own right or as an excellent example of something that is intrinsically bad (in the opinion of the DR author).

4. How do I use POWDER?

Usage of POWDER is dependent on the intended results.

The process is a sequence and can be stepped through all the way or terminated at the appropriate place to obtain the desired result.

For someone who wishes to describe content:

  1. Create one or more Description Resources within a POWDER document and make it available on the Web.
    This requires no more than the generation of an XML document that contains the necessary components of a Description Resource.
    No further steps are necessary, other than publishing, to create a Description Resource and thus offer this meta data publicly.

  2. Make the POWDER document discoverable
    This is no more than making sure that DRs can be accessed by the intended audience, most likely the Web.
    You should link from the described resources to the POWDER document.
    It is also possible to describe someone else's content.

  3. Add trust to your descriptions
    Trust is an integral component of POWDER documents. Choose a method, ranging from simply publishing your document, thereby making a claim, to providing fully certified POWDER documents with supporting resources to add trust to your DRs.

For someone who wants to access RDF triples obtained from POWDER documents:

  1. A POWDER processor must be used to read and process POWDER documents to obtain RDF triples that describe specific Web resources. Refer to the [DR] document for POWDER processor details.
  2. Now an RDF querying tool could be used to retrieve:

As an alternative, someone may want to draw inferences based on the data contained in POWDER documents.
Drawing an inference means deriving implicit information from known facts.
For example:

Therefore that particular resource conforms to accessibility level AA.

In order to do so, the following steps should be followed.

  1. Convert the POWDER document into a POWDER-S document (intended for semantic web usage, via the GRDDL transform defined in the Formal Semantics document [FORMAL]
  2. Now an inference engine must be used to retrieve the property of a given resource.
    For this step it is important to realize that an extension to RDF/OWL has to be implemented in the reasoning engine in order to support and utilize POWDER-S

Following the example above we could now determine which resources on example.com are AA conformant, as well as the conformance level of a given resource.

How do I create a Description Resource?

The first step is to create a skeleton POWDER document in XML and declare the namespaces:

<?xml version="1.0"?>

Any descriptive vocabularies may be used which are identified by a namespace. The POWDER namespace is required.

The "ex" namespace used above is a fictitious example to denote that any namespace may be used.

Next we need to say who has created the document. All POWDER documents have exactly one attribution element, and within that an issuedby element. This points to details of the person or organization that has published the POWDER document. Exactly which details is for the publisher to decide, but a name and a homepage address should usually be included, perhaps along with contact details. POWDER authors may use either the Friend of a Friend [FOAF] vocabulary or the Dublin Core [DC] to do this.

<?xml version="1.0"?>
     <issuedby src="http://authority.example.org/company.rdf#me" />

An individual or organization may publish many Description Resources. Therefore it is more convenient to define the profile information, describing that individual or organization, in a single file and refer to it from each POWDER document as it is created. The above example points to a profile as listed below. Notice the rdf:ID="me", which is referred to by the #me in the above example.

The publisher's profile would look similar to this:

<?xml version="1.0"?>
  <foaf:Organization rdf:ID="me">
    <foaf:name>The POWDER Company</foaf:name>
    <foaf:homepage rdf:resource="http://authority.example.org" />

A POWDER document will typically also include information about when it was created (our example was created on 14 December 2007).

<?xml version="1.0"?>
    <issuedby src="http://authority.example.org/company.rdf#me" />

Next an actual Description Resource (DR) is added. This example POWDER document will contain a single Description Resource, as the dr element.

<?xml version="1.0"?>
    <issuedby src="http://authority.example.org/company.rdf#me" />

The Description Resource itself must contain at least one set of IRIs. This is the scope of the Description Resource - i.e. what it describes. IRIs are a superset of the well known Uniform Resource Identifier (URI) with the expansion of also allowing international characters. If multiple IRI sets are included within a DR, then the scope is all the IRIs in all the IRI sets.

In this case, the scope is 'everything on example.com'. This is done using the includehosts element. There are other elements available which are described in detail in the Grouping of Resources document [GROUP]. All Description Resources must contain at least one iriset element, and this cannot be empty and cannot contain any elements from any other namespace.

<?xml version="1.0"?>
    <issuedby src="http://authority.example.org/company.rdf#me" />

The final key element of a Description Resource is the actual description. There are two ways of providing this.

A DR must contain at least one of either a descriptor set or tag set, none of which may be empty. Subject to that condition, any number of tag sets and descriptor sets may be included. The example DR states that in the opinion of the entity referenced by the issuedby element, all resources within its scope are described by all descriptive elements.

We'll add one of these to our example: a descriptor set. It may contain RDF/XML properties with literal values (including XML literals) or resources identified by an IRI. In addition, a textual and/or graphic summary that can be displayed to end users may be included in a descriptor set using the displaytext and displayicon elements as shown in the completed example below. Notice that the namespace used in the descriptor set is also highlighted.

Generic Example of a POWDER Document Containing a Single Description Resource.

This is a repeat of Example 2-1 in the Description Resources document [DR]

<?xml version="1.0"?>
<powder xmlns="http://www.w3.org/2007/05/powder#" 

    <issuedby src="http://authority.example.org/company.rdf#me" />



      <displaytext>Everything on example.com is red and square</displaytext>
      <displayicon src="http://authority.example.org/icon.png" />



Of course, much more complicated structures are possible than this simple example.

The following example uses an ordered list of DRs. In such a list 0 or 1, DRs will apply to a given resource. A processor will work through the list until the first match is found. This feature is designed to make it easy to add POWDER descriptions to existing content within established workflows. Effectively it enables a list of exception to be created, ending with the general case.

NOTE: The example below contains an abouthosts element which may be used by a processor to decide if a list, perhaps a long list, need be processed at all. Here we know that only resources on example.com are described in the ordered list, however it is important to recognize that this does not guarantee that all resources on example.com are described.

Example of a POWDER document which contains an ordered list

This is a repeat of from Example 2-6 in the Description Resources document [DR]

<?xml version="1.0"?>
<powder xmlns="http://www.w3.org/2007/05/powder#" 

    <issuedby src="http://authority.example.org/company.rdf#me" />






The example above encodes the following assertion made by the entity described at http://authority.example.org/company.rdf#me: all web resources with paths on example.com that start with /foo are blue, all other resources on example.com are red.

POWDER offers other methods of defining the scope of DRs that are designed to increase flexibility. For example, it is possible for the scope of a DR to be the union of two or more IRI sets, meaning that an IRI that is a member of any IRI set is described. The following example describes resources that are available from example.com where the path starts with /foo AND those at example.org where the path starts with /bar.

Example of a POWDER document with two IRI sets

This is a repeat of Example 2-8 in the Description Resources document [DR]

<?xml version="1.0"?>
  <powder xmlns="http://www.w3.org/2007/05/powder#"
      <issuedby src="http://authority.example.org/company.rdf#me" />


        <displaytext>Everything on example.com/foo, and everything on example.org/bar, is red and square</displaytext>
        <displayicon src="http://example.org/icon.png" />

How do I publish a DR?

Like any other resource on the Web, POWDER documents can only be found if you know where to look or you can follow a link from where you already are. Individuals or organizations whose activities include reviewing and describing online content will be able to set up and advertise the existence of a repository of Description Resources.

There are several methods of linking from a resource to a POWDER document that describes it.
First of all, HTML documents can include a link element in much the same way as is common for linking to CSS style sheets, RSS/ATOM feeds, etc.

<link rel="describedby" href="powder.xml" type="text/powder+xml"/>

The relationship type (rel) of "describedby" tells user agents that the file contains a description of the resource and the mime type tells it that it is a POWDER document. The value of the href is the IRI of the POWDER document.

Just as with stylesheets, it's important to include this link on all described pages, so it is best included in the document template.

When using HTML link elements in this way, authors should also refer to POWDER's profile document as shown below, if using an HTML version that supports it.

<html xmlns="http://www.w3.org/1999/xhtml">
   <head profile="http://www.w3.org/2007/11/powder-profile">
      <meta name="wdr.issuedby" content="http://authority.example.org/company.rdf#me"/>
      <link rel="describedby" href="powder.xml" type="text/powder+xml"/>
      <title>Welcome to example.com </title>
      <p>Today's content is ....</p>

Adding this makes the relationship type "describedby" clear and allows content authors to include data about the POWDER document within the described resource. As shown in the example above, this can include information about who has issued the DRs (issuedby), so that a user agent that recognizes and trusts that Description Resource author will be more confident that the data available is trustworthy and therefore that the link is worth following.

Whilst HTML link elements do make POWDER documents discoverable, the preferred method is to configure the server to include an HTTP Response Header that does the same job.

Link: <powder.xml>; rel="describedby"; type="text/powder+xml";

This has several distinct advantages:

  1. It allows all resources, not just HTML documents, to point to a POWDER document.
  2. It allows the POWDER document to be discovered and parsed before the described content is parsed.
  3. In some content production environments the process of configuring servers can be easier to achieve and lead to longer-term stability than changing templates.

Where POWDER documents are discoverable via HTTP Link: headers, a user agent or Web crawler can discover them by doing an HTTP HEAD request which may affect how, or indeed whether, a subsequent GET request is made.

The describedby relationship type is made available as an RDF property, where a more formal relationship between a POWDER document and a resource that it describes is required. This may be used, for example, in XHTML documents annotated with RDFa as described in section 4.1.3 of the Description Resources document [DR]. One important benefit of using RDFa to link to POWDER documents is that it allows you to point to a description of the destination of a hyperlink.

As an example of this, imagine an XHTML document about a rock band that includes hyperlinks to ring tones, images and videos.

An XHTML/RDFa Fragment with Hyperlinks described using POWDER

  <li><a href="/clips/low_res_clip.mpg" 
         about="/powder.xml">30" clip</a></li>

  <li><a href="/videos/full_video.mpg"
         about="/powder.xml">Full 10 minute video</a></li>

  <li><a href="/tones/ring_tone1.mp3" 
         about="/powder.xml">Ring Tone</a></li>




      <tag>Large download</tag>
      <label>Only suitable for download via high-bandwidth connections</label>

      <typeof src="http://www.w3.org/2007/07/mobileOK#conformant" />

This ordered list of DRs states that resources on the example.com host where the IRI path includes /videos/ are large downloads. All other resources on example.com are conformant with mobileOK [MOK]. By taking note of the available descriptions, the user agent may display the links differently on different devices (or chose not to display them at all).

Notice a couple of points here:

  1. Use of the "rev" attribute which reverses the relationship so that "powder.xml" is "about" the hyperlink.
  2. The same POWDER document contains both the DRs and can therefore be used from cache.

How do I make my DR more trustworthy?

Trust is a critical aspect of Description Resources; however, trust is very much a matter of opinion. The level of trust demanded of a given DR will depend on what the description says and to what use it will be put. For example, an individual user finding a DR that declares a Web site to offer children's birthday party ideas can make his/her own assessment of its quality and usefulness. In contrast, a multi-million dollar business will need very strong assurance that a DR declaring a Web site to be medically accurate and freely available is trustworthy before including it in a portal of high quality, license-free health care materials. For this reason, we do not define a single trust mechanism that must be used. Rather, there are a variety of different methods of adding trust to DRs, some of which may be used in combination.

Where applicable, we define vocabulary terms designed to aid the building of trust.

The methods cited here do not comprise an exhaustive list. Other techniques, such as XML Signature [XMLSIG] and Web of Trust [WOT], may be equally applicable. Trust is a human judgment that can only be made by weighing the likelihood that the data is true against the consequences of it being false. This judgment is highly dependent on the circumstances under which the need to extend trust arises. It is clear, however, that Description Resources are unlikely to be trusted in isolation and that both their publishers and consumers will only benefit from their existence if one or more techniques for enhancing trust are employed.

These methods would also serve to increase trust placed in the meta information provided for Web documents in general. Today's meta data, such as the keywords provided in META elements, can have little value as this mechanism has been abused to such a degree that renders it practically useless from an informational point of view. Search engines do not place much stock on keywords to give any indication about the relevance of a given Web page to a given topic and use different means, such as incoming links, as a basis for page ranking. False claims made in meta information, intended to lure the user into clicking on a link to a resource, also serve to lower the value of meta information.

Also, a less deliberate yet still disturbing, fact about meta information is aging. As meta information is not kept up to date, it loses its relevance to the content it describes. POWDER elevates meta information once again into the spotlight by allowing a third-party to certify the veracity of the meta information given, declare the date of the verification, and define a date after which the certificate will no longer be valid.

How do I find DRs for a document?

DRs should describe reflect Web resources in their current state. This requires that DRs are kept up-to-date and will thus change frequently. Therefore it is important to provide a link to the "latest" version of DR document, which has the added benefit of providing a DR history. If a request is then made for this document, an HTTP 302 redirect can be used to point the requesting client to the actual, current POWDER document.

A further option would be querying a known repository of DRs. Since the creation of DRs is not limited to the author or provider of a Web resource, repositories of DRs may be created by companies or special interest groups, for example those specializing in standards compliance certification or child protection. These may then be queried to obtain descriptions and scope information. For example, a large content provider, upon planning to switch from regular meta data to POWDER, could, in a first step, create one or several DR documents. The scope would be set such as to cover all areas of the content provider's content. Meta elements and various other information may be copied into the DR documents. Insertion of the appropriate link in the web resource, pointing to the correct DR, would be the last step prior to publication of the DR documents.

Personal collections of DRs may be traded or passed between users of social networks or other groupings of similar interests. Search engines' indexes may also contain references to publically available DRs.

In short, DRs are normal documents that may be discovered in all the usual ways that other documents are discovered on the Web.

How do I process POWDER documents?

A POWDER Processor accepts queries for descriptions of specific resources which it generates by reading POWDER documents and returning RDF triples. There is a minimal set of requirements for a conformant POWDER processor given in the Description Resources document [DR] where implementers are encouraged to go further than the minimum. A POWDER Processor may act as a gateway to a repository of DRs and use that as its only source of data, or it may be more generalized, accepting both the IRI to be described and one or more POWDER documents as the source from which to generate the description. The application environment will determine which factors are important but it is of course useful to think about usability, error trapping and potential extensions of its functionality.

How do I process POWDER-S?

The operational semantics, meaning the XML file, are underpinned by formal semantics. A GRDDL transform is associated with the POWDER namespace that allows the XML data to be rendered and processed as RDF/OWL.

However, RDF/OWL cannot currently interpret string values of RDF properties as International Resource Identifiers (IRI). In other words the string "http://example.org" is not necessarily recognized as the web address (IRI) http://example.org.

To allow this to happen a semantic extension has to be created that makes this definition and relates to the matchesregex property. The extension is defined in POWDER: Formal Semantics

5. What do I need to use POWDER?

Basic requirements

No special knowledge is needed to understand plain POWDER.

The examples provided in the documents should be sufficient to begin using POWDER with only rudimentary knowledge of markup languages.

You will need to be able to publish (upload) the POWDER web descriptions to some location (server) where people can reach (download) them.

This may be the same location as the web resources themselves.

Creation of a DR

The creation of a DR could be accomplished via online tools provided by companies who offer repositories of DRs or created by hand with a desktop editor in the simple form outlined in this document. In that case all that is needed, if so desired, is to link from the web resource to the DR via a link element or other mechanism as outlined here.


POWDER-S is not be dealt with in great detail in this document since it is essentially an extension of RDF and OWL which are fully documented elsewhere. The interested reader is encouraged to consult the [FORMAL] specification, which outlines POWDER-S exhaustively.

The FORMAL specification provides the necessary underpinning for POWDER and POWDER-S such that processing done on these types of documents can be conformant and consistent with existing Semantic Web technologies. Furthermore, the Grouping specification [GROUP] also outlines the support for non-url environments, e.g. ISBN codes or similar resources.

It should be noted, however, that the creation of POWDER-S documents requires the implementation of an RDF extension in the reasoning engine.

See Section 4.3 POWDER-S IRI Set Semantics in the Formal document [FORMAL] for details.

Generic example of a POWDER-S Document Containing a Single Description Resource

This is a repeat of Example 2-3 from the Description resources document [DR]

1  <?xml version="1.0"?>
2  <rdf:RDF
3     xmlns:wdrs="http://www.w3.org/2007/05/powder-s#"
4     xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"
5     xmlns:rdfs="http://www.w3.org/2000/01/rdf-schema#"
6     xmlns:owl="http://www.w3.org/2002/07/owl#"
7     xmlns:ex="http://example.org/vocab#">
9    <owl:Ontology rdf:about="">
10     <wdrs:issuedby rdf:resource="http://authority.example.org/company.rdf#me" />
11     <wdrs:issued>2007-12-14T00:00:00</wdrs:issued>
12   </owl:Ontology>
14   <owl:Class rdf:nodeID="iriset_1">
15     <owl:equivalentClass>
16       <owl:Class>
17         <owl:intersectionOf rdf:parseType="Collection">
18           <owl:Restriction>
19             <owl:onProperty rdf:resource="http://www.w3.org/2007/05/powder-s#matchesregex" />
20             <owl:hasValue rdf:datatype="http://www.w3.org/2001/XMLSchema#string">\:\/\/(([^\/\?\#]*)\@)?([^\:\/\?\#\@]+\.)?(example\.com)(:([0-9]+))?\/</owl:hasValue>
21           </owl:Restriction>
22         </owl:intersectionOf>
23       </owl:Class>
24     </owl:equivalentClass>
25   </owl:Class>
27   <owl:Class rdf:nodeID="descriptorset_1">
28     <rdfs:subClassOf rdf:resource="http://www.w3.org/2002/07/owl#Thing"/>
29     <rdfs:subClassOf>
30       <owl:Class>
31         <owl:intersectionOf rdf:parseType="Collection">
32           <owl:Restriction>
33             <owl:onProperty rdf:resource="http://example.org/vocab#color" />
34             <owl:hasValue>red</owl:hasValue>
35           </owl:Restriction>
36           <owl:Restriction>
37             <owl:onProperty rdf:resource="http://example.org/vocab#shape" />
38             <owl:hasValue>square</owl:hasValue>
39           </owl:Restriction>
40         </owl:intersectionOf>
41       </owl:Class>
42     </rdfs:subClassOf>
43     <wdrs:text>Everything on example.com is red and square</wdrs:text>
44     <wdrs:logo rdf:resource="http://example.org/icon.png" />
45   </owl:Class>
47   <rdf:Description rdf:nodeID="iriset_1">
48     <rdfs:subClassOf rdf:nodeID="#descriptorset_1"/>
49   </rdf:Description>
51 </rdf:RDF>

Line-by-line explanation:

Lines 1-7
Document header: with namespace declarations
Lines 9-12
Attribution: Each POWDER-S document is an OWL ontology and the attribution section is an ontology header.
Lines 14-25
Scope: An OWL class acts as the IRI set. It is this class that makes use of the matchesregex property which is only understood by software that implements the semantic extension which confers membership of the class on resources whose IRIs match the given regular expression(s).
Lines 27-45
Description: These are the actual descriptors.
Lines 47-49
Assertion: The encoding of the DR is completed by asserting that the IRI set is a subclass of the descriptor set. This is the crucial step that allows the inference to be drawn about members of the IRI set.

7. Examples

ICRA Labelling

The ICRA vocabulary facilitates what is intended to be a culturally neutral description of online content in terms that may reflect parental concerns around the world. Such descriptions, especially where backed up by third-party checks, can be useful in delivering appropriate content to different target groups, particularly children.

In the following scenario, we imagine that example.com publishes content across its portal that does not include any sex, nudity, violence or other potentially offensive or harmful material, except in its night life section, where references to alcohol, tobacco and potentially offensive language are to be found, especially as it invites users to post reviews of bars, pubs and clubs they have visited. Since all pages relevant to the night life section have a URL of the form http://www.example.com/nightlife... it is able to describe its own content in the following POWDER document.

Example: ICRA labelling [XML]

1  <?xml version="1.0"?>
2  <powder xmlns="http://www.w3.org/2007/05/powder#" 
3          xmlns:foaf="http://xmlns.com/foaf/0.1/"
4          xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"
5          xmlns:icra="http://www.icra.org/rdfs/vocabulary2008#">
7    <attribution>
8      <issuedby src="http://www.example.com/company.rdf#me" />
15     <certifiedby src="http://independent.example.org?verify=http%3A%2F%2Fwww.example.com%2Fpowder.xml" />
16   </attribution>
18   <ol>
19     <dr>
20       <iriset>
21         <includehosts>example.com</includehosts>
22         <includepathstartswith>/nightlife</includepathstartswith>
23       </iriset>
25       <descriptorset>
26         <icra:nz>1</icra:nz>
27         <icra:sz>1</icra:sz>
28         <icra:vz>1</icra:vz>
29         <icra:lb>1</icra:lb>
30         <icra:lc>1</icra:lc>
31         <icra:ha>1</icra:ha>
32         <icra:hb>1</icra:hb>
33         <icra:dz>1</icra:dz>
34         <icra:ua>1</icra:ua>
35         <icra:pa>1</icra:pa>
36         <displaytext>
               No nudity; No sexual material; No violence; 
               Profanity or swearing; Mild expletives; 
               Depiction of tobacco or its use; Depiction of alcohol or its use; 
               No potentially disturbing material; 
               User-generated content such as chat rooms and message boards (moderated); 
               Contains advertising
37       </descriptorset>
38     </dr>
40     <dr>
41       <iriset>
42         <includehosts>example.com</includehosts>
43         <includepathstartswith>/nightlife</includepathstartswith>
44       </iriset>
46       <descriptorset>
47         <icra:nz>1</icra:nz>
48         <icra:sz>1</icra:sz>
49         <icra:vz>1</icra:vz>
50         <icra:lz>1</icra:lz>
51         <icra:hz>1</icra:hz>
52         <icra:dz>1</icra:dz>
53         <icra:uz>1</icra:uz>
54         <icra:pa>1</icra:pa>
55         <displaytext>
               No nudity; No sexual material; No violence; 
               No potentially offensive language; 
               No potentially harmful activities; 
               No potentially disturbing material; 
               No user-generated content; Contains advertising
56       </descriptorset>
57     </dr>
58   </ol>
60 </powder>

The document contains two Description Resources that reflect the two different types of content on (the fictional) example.com. This document makes use of POWDER's ordered list feature such that every page on example.com, irrespective of its content, can be linked to the same file. This can be done by including a link element in each page's HTML thus:

<link rel="describedby" href="/powder.xml" type="text/powder+xml" title="ICRA labels" />

but is more efficiently done by configuring the example.com server(s) to include the equivalent HTTP Link header thus:

Link: </powder.xml>; rel="describedby"; type="text/powder+xml";

A user agent, such as a content aggregation service, can now support different policies with respect to the resources available from example.com. It may, for example, choose not to include the night life section for all its customers, or conversely to promote that section as it is a better fit for its own target market.

Notice also that example.com has referred to an independent certification body in Line 15. A user agent might be satisfied that example.com's description of its own content is sufficiently trustworthy to be of use or it may choose to query the service operated by http://independent.example.org before conferring trust on the data. In the specific case of ICRA descriptions, it is often the case that a description of content as being potentially harmful or offensive is trusted more readily than descriptions of content as being suitable for children. Thus external verification mechanisms can have varying degrees of importance even within a single domain of interest.


mobileOK is a standard that deals with content that is rendered within a mobile context and adheres to a set of guidelines called the mobileOK Basic Tests. Claiming mobileOK has been one of the major use cases of POWDER and the example below outlines how this claim can be made using a POWDER document.

Example: mobileOK (from Example 5-3: A DR Claiming Conformance to mobileOK Basic, Supported By the mobileOK Basic Checker [XML]

1  <?xml version="1.0"?>
2  <powder xmlns="http://www.w3.org/2007/05/powder#">

3    <attribution>
4      <issuedby src="http://www.example.com/company.rdf#me" />
5      <issued>2008-06-25T00:00:00</issued>
6      <supportedby src="http://validator.w3.org/mobile/" />
7    </attribution>

8    <dr>
9      <iriset>
10       <includehosts>example.com</includehosts>
11     </iriset>

12     <descriptorset>
13       <typeof src="http://www.w3.org/2008/06/mobileOK#conformant" />
14       <displaytext>The example.com website conforms to mobileOK</displaytext>
15       <displayicon src="http://www.w3.org/2005/11/MWI-Icons/mobileOK.png" />
16     </descriptorset>
17   </dr>

18 </powder>

The above example shows a POWDER document which contains information about who has made the claim, when the claim was made and, in this particular case, also lists a supporting link to the W3C mobileOK Checker. The Checker is used to test conformance.

Next the DR contains the resources to which this claim applies and the descriptors. In Semantic Web terms, the typeof element is a shorthand for rdf:type and asserts that all resources on example.com are instances of the mobileOK Class. Descriptive text and an icon are also provided.

8. Acknowledgements

The editor gratefully acknowledges the substantial contributions made by:

Phil Archer - Institute of Informatics & Telecommunications (IIT), NCSR "Demokritos" (formerly with FOSI)
Alan Chuter - Technosite
Diana Pentecost, AOL LLC

9. References

Dublin Core Metadata Initiative This information is at http://dublincore.org/
Protocol for Web Description Resources (POWDER): Description Resources , P. Archer, K. Smith, A. Perego. This document is at http://www.w3.org/TR/powder-dr
Evaluation and Report Language (EARL) 1.0 Schema Schema Document 23 March 2007, Shadi Abou-Zahra. This document is at http://www.w3.org/TR/EARL10/
FOAF Vocabulary Specification Namespace Document 24 May 2007, D. Brickley, L. Miller. This document is at http://xmlns.com/foaf/spec/
Protocol for Web Description Resources (POWDER): Formal Semantics 2008, S. Konstantopoulos, P. Archer. This document is at http://www.w3.org/TR/powder-formal/
Gleaning Resource Descriptions from Dialects of Languages (GRDDL) W3C Recommendation 11 September 2007. D. Connolly. This document is at http://www.w3.org/TR/grddl/
Protocol for Web Description Resources (POWDER): Grouping of Resources, A. Perego, P. Archer. This document is at http://www.w3.org/TR/powder-grouping/
Internationalized Resource Identifiers (IRIs), M. Drst, M. Suignard. IETF, January 2005. This document is at http://www.ietf.org/rfc/rfc3987.txt
W3C mobileOK Scheme 1.0, J. Rabin. This document is at http://www.w3.org/TR/mobileOK/
Protocol for Web Description Resources (POWDER): Test Suite 2008, A. Kukurikos. This document is at http://www.w3.org/2007/powder/Group/powder-test/
POWDER: Use Cases and Requirements W3C Working Group Note 31 October 2007, P. Archer. This document is at http://www.w3.org/TR/powder-use-cases/
Web Content Accessibility Guidelines 2.0, B. Caldwell, M. Cooper, L. Guarino Reid, G. Vanderheiden. This document is at http://www.w3.org/TR/WCAG20/
XML-Signature Syntax and Processing, Donald. Eastlake, J. Reagle, D. Solo (Eds). W3C Recommendation 12 February 2002. this document is at http://www.w3.org/TR/xmldsig-core/
Web Of Trust RDF Ontology, D. Brickley. This document is at http://xmlns.com/wot/0.1/