Making identifiers concrete

Here is a link to LukeW's good summary of Stuart Weibel's plenary that closed VALA 2008. I liked what Stuart had to say and my notes probably differ a tad from Luke's notes, but maybe I just misunderstood what Stuart had to say?

Stuart comes from OCLC and presented really well, leaving most of us with a new perspective on what could have been a dull a dry topic. I found his message easy to follow and quite inspiring.

Branding & Web 2.0
OCLC have released a report on the perceptions of libraries and information resources.
Libraries and search engines are trusted about the same.
People care about the quantity and quality of information.
They do not view paid information or free information differently.
Branding can be achieved by building on trust by making things look free. Scale is represented by libraries and their presence everywhere. They have global scope and reach (via networks?). BUT people need more awareness. We must be part of the new online environments that dominate our lives.
Social networking software. Only technical manifestation is new (we've always networked). Motivate people to tag, participate. Wired said 40% of those they interviewed contributed in one way or another. (Higher than Yahoo's figures.)
Re social consumer environments. Facebook, etc. are not just for games. But they are probably not the right models. There are lessons to learn though (OCLC has just put an application into Facebook). They are flawed – closed gardens, rudimentary features, but offer an experience as well as a service.
Libraries must compete and compare favourably with popular models (in Seattle books with coffee is the law!). But can we compete and should we? What can we do to fit in and how to distinguish between the trends and the trendy?
Catalogues – how can they change morph/grow? Networking. Collections linked to people, organisations, concepts, context, metadata, etc. (So I've just started an account on Library Thing to learn how this works.)
Do we need a web or scaffolding – do we want more – coherence, durability, etc.
Mentioned FRBR – works, expression, manifestation and item. But with other dimensions.
He said that for discovery on the web, a book review is more useful than a MARC record (I agree - how many people truly understand MARC records?). They are a social bibliography. He also cited: lists, services, commentary, etc.

Infusing bibliographic ideas into the web & vice versa?
First class objects need: persistent identifiers; access to all; stand alone status (identification & clear IP); and they are curated (not left lying around unintended - bit-rot!). Allow users to enter and traverse the catalog from any point!
Establishing a canonical identity on the web is very important.
See WorldCat identities. This should have been done ages ago. Tag cloud into popular Ids. Has stuff by/on author, works, links, encourages serendipitous discovery, associated subjects. All from bibliographic data.

Identities on the web
What characteristics are best in identifiers? There are no hard/fast rules – just suggestions. He thinks URLs need to reflect something about what they are. Make them meaningful.

Design criteria for identifiers
Characteristics:

• persistence (function of organisational commitment);
  
• universal accessibility & global scoping (work everywhere, open to all, WorldCat provides architecture for library assets mapping global surrogate to the local);
  
• optimised for search engines and canonical (raises search engine ranking);
  
• branding via URIs (mini-billboards);
  
• usability by people and machines – speakable, short, predictable (hackable).

This amounts to gluing the pieces together with identifiers.

WorldCat identifiers – are they good enough?
Unique, free, citable, resolvable, linked, canonical (no, not really). Some functional duplicates (more records pointing to same thing).

A glimir of the future?
A global manifestation identifier. Global, business neutral, canonical, provides URL equity, fits with FRBR model.
There are other identifier schemes. So, OCLC is cautiously exploring this territory.

Summary:
IDs are the key; they are needed for mission, to compete, brand, to bring bibliographic values to web, to provide services and access to digital tribe. Books not done yet.

See particularly his blog posts on related subjects (persistent identifiers).

Andy Powell asked whether he was talking about the semantic web. Stu supports it, but is skeptical about the technologies involved. He spoke of middleware as the plywood of the internet. It needs to become the plywood of our arena. He said the abstract model has fundamental importance on the web.