The Case for Public Resource Identifiers A Call for Action

1
The Case for Public Resource IdentifiersA Call
for Action
Steve Pepper, Ontopia Identity, Reference, and
the Web Workshop, Edinburgh, May 2006
2
About this presentation

• Just a subset of my paper for this workshop
• 1. Introduction (on the need for identifiers)
• 2. Requirements on a Global Identifier Mechanism
• 3. Published Subjects
• 4. Alternative Proposals
• 5. Call to Action
• Focus here is on the Call for Action
• Expands on the ideas presented to the W3C AC
• http//www.w3.org/2006/05/pepper
• The Case for Published Subjects,http//www.ontopi
  a.net/topicmaps/materials/The_Case_for_Published_S
  ubjects.pdf

3
URIs and the Towers of Babel

• URIs constitute a universal way of naming things
• This has tremendous potential
• Globally unique, language-independent identifiers
  for any conceivable subject under the sun
• We could finally stop building Towers of Babel
• (Or at least make a real impression on info glut)
• Whenever I say URI, please hear IRI

4
The Semantic Superhighway

• At the very least, URIs give us the chance to
  create a semantic superhighway
• A foundation for solving the problem of infoglut
• A level of the Semantic Web below RDF
• But its not happening
• In fact, we are witnessing the creation of new
  Towers of Babel
• New, redundant vocabularies appear daily
• The potential is not being realised. Why?

5
Why the Potential is Not Being Realised

• Too few people are using URIs as identifiers
• Its partly our fault
• We dont explain the benefits loudly enough
• We dont make it sufficiently easy and worthwhile
• Weve been in turmoil on technical issues
• And those that do use URIs as identifiers tend to
  create new ones, rather than reuse existing ones
• This defeats the purpose of a universal naming
  scheme

6
Problems We Need to Address

• The concepts are too varied and confusing, and
  they are not being marketed well
• The world perceives a Heinz 59 variety of URIs
• IRI, URN, URL, httpURI, XRI, WPN, TDB, ...
• Reuse hard in practice
• No repositories to aid discovery
• No simple way to figure out what a given
  identifier is supposed to identify

7
What We Need To Do

• Make the concepts easy to understand and apply
• Agree on one flavour of URI
• Promote the hell out of it
• Deprecate everything else
• Encourage reuse in every possible way
• Make it easy to discover and interpret
  identifiers
• Encourage them to be made public

8
Make httpURIs the flavour of choice!

• Already widely used in Topic Maps and RDF
• And, increasingly, elsewhere
• No longer subject to paralysing controversy
• TAGs resolution of httpRange-14 issue allows any
  httpURI (including slash httpURIs) to identify
  anything
• Familiar to anyone who has used a web browser
• Having people type http//... will not be a
  problem
• Most importantly They resolve easily when you
  click on them
• So lets exploit that fact...

9
Have them resolve to something useful!

• The most useful thing to resolve to would be a
  descriptor
• definition, description, some other kind of
  indication of what the identifier is intended to
  identify
• (machine-processable information would be a
  useful optional extra)
• Allows users to know what the identifier means
• i.e., what it is intended to identify
• ... and decide whether it is appropriate for them
  to use
• Also easy to discover using web search engines

10
A Proposal for Public Resource Identifiers

• Three Key Principles
• A Public Resource Identifier (PRI pry) is
• an httpURI that has been minted for the explicit
  purpose of serving as an identifier
• It resolves to a Public Resource Descriptor (PRD)
• that describes which resource (subject) it
  identifies and states who minted it
• ANYONE should be able to mint a PRI
• A bottom-up mechanism (anarchic, like the Web)
• Survival of the fittest ( the most trusted)

11
Sounds familiar?

• Essentially identical to Published Subjects
• Proposed by SC34 in 1999 as Public Subjects
• Refined within OASIS as Published Subjects
• Only major difference
• The requirement to use httpURIs
• But also a restatement of basic Web Architecture
  principles
• Cool URIs dont change
• httpURIs can identify anything
• httpURIs should resolve to something useful
• Does it matter what we call it?
• (Does it even need a new name?)
• A rose by any other name would smell as sweet

12
Whats in a Name?

• Would this rose by any other name smell as sweet?
• Its a branding issue
• Does the name have useful / non-useful
  connotations?
• cf. resource and subject
• Can we create cool TLAs (if necessary)?
• pronouncable, depictable, available,
  distinguishable

13
A Call for Action Incubator Group

• Let the W3C and SC34 get together to push PRIs
• The details can be hammered out in a few months
• Then bring OASIS on board
• (before even more vocabularies appear...)
• Approach other standards bodies, including
• SC32 (databases, meta data), SC36 (e-learning),
  IFLA, and others (TBD)
• Lets build a Semantic Superhighway together
• I will be making a formal proposal shortly

14
The PRI Incubator Group

• A W3C activity with participation from ISO and
  OASIS
• Chartered for lt 12 months to
• refine and codify the Three Key Principles
• provide absolutely minimal recommendations for
• the form of the PRI
• the content of the PRD
• address issues of branding and outreach
• esp. naming and strategy for promulgating the
  paradigm

15
More on the PRI-XG

• Suggested participation
• 3 sponsors from W3C (TAG, SemWeb, ...)
• Invited international experts from ISO SC34
• Invited experts from OASIS (UBL, OpenOffice)
• Coordinator and team contact
• Suggested goals
• XG Report
• Possible fast-track as W3C Rec and ISO Technical
  Report

16
The Continuation of a Fine Tradition

• Diderots Encyclopédie (17511780)
• THE DESCRIPTION OF EVERYTHING
• Oxford English Dictionary (18571928)
• THE MEANING OF EVERYTHING
• Public Resource Identifiers (1999 )
• THE IDENTITY OF EVERYTHING
• X