Standards, Standards Bodies and Web Services

Henry S. Thompson

14 Feb 2006

The wonderful thing about standards is

there are so many to choose from!
Seriously, why standards?
- Interoperability
- Monopoly sounds attractive
  - and periodically companies get sucked in to trying
  - But customers are (rightly) wary

ISO

The International Standards Organisation

IEEE

Institute of Electrical and Electronic Engineers

IETF

Internet Engineering Task Force

OASIS

Organization for the Advancement of Structured Information Standards

Vendor organisation, rebadged SGML-Open
Corporate membership
Sometimes seen as a bid by big guys (Microsoft and IBM in particular) to undermine W3C
Vertical 'standards', e.g. ebXML, WS-Reliability, WS-Security

Membership is for companies and non-profits
Membership costs real money
- Many discussions in public, more-or-less
- But only paying members participate in decision making
Work is done in Working Groups
- Open to all Members
  - Avoids exposure to anti-trust/Restraint of Trade problems
- Chartered with particular deliverables
- W3C staff provide basic support
- Mixture of face-to-face and phone conference work
W3C Process ensures
- Consensus-based decisions
- Regular review
- Interoperable implementations
- Transparent IPR situation

A huge problem for standards bodies
A way of getting an effective monopoly:
- Join a standards body
- Produce a standard which depends on a key bit of technology
- Have a patent for that technology!
The main alternatives

RAND
Reasonably and Non-Discriminatory

RF
Royalty Free
W3C has pioneered a requirement for a commitment to RF
- Took four years to negotiate agreement
- Still lost some members
- Commitment is, essentially, to give RF licence for any technology which is required to implement a Recommendation.
- Doesn't require a patent search, but many companies (believe/act as if) it does

OASIS began with no IPR policy, now requires a RAND/RF choice.
Early implementations of SOAP required licenses from four distinct companies, one of whom was very slow to respond to requests.
IETF has struggled with the issue, not really resolved yet
- It is generally accepted (but not officially agreed) that IETF dropped work on Sender ID because of Microsoft's RAND requirements for participation
In summary, a real minefield

Gather requirements and define an HTTP and/or SOAP-based protocol for selecting instances of subgraphs from an RDF graph
The goal of this work is to help make it as easy to 'join' data on the Web as it is to merge tables in a local relational database.
Publications:
- Use Cases and Requirements
- SPARQL Query Language, and SPARQL Protocol Working Drafts availiable

Provide guidance, in the form of documents and demonstrators, for developers of Semantic Web applications.
Develop consensus best practices on ontology engineering guidelines, vocabulary development, educational material and demo applications.
Support initiatives for transforming selected high-visibility ontologies and thesaurii to OWL and RDF.