The linked data rules (from Tim Berners-Lee)

Like the web of hypertext, the web of data is constructed with documents on the web. However,  unlike the web of hypertext,  where links are relationships anchors in hypertext documents written in HTML, for data they links  between arbitrary things described by RDF,.  The URIs identify any kind of object or  concept.   But for HTML or RDF, the same expectations apply to make the web grow:

1. Use URIs as names for things

2. Use HTTP URIs so that people can look up those names.

3. When someone looks up a URI, provide useful information, using the standards (RDF, SPARQL)

4. Include links to other URIs. so that they can discover more things.

Simple.  In fact, though, a surprising amount of data isn't linked in 2006, because of problems with one or more of the steps.  This article discusses solutions to these problems, details of implementation, and factors affecting choices about how you publish your data.

http://www.w3.org/DesignIssues/LinkedData.html

Architecture - “Ingect” (working title)

What does a Linked Data Approach mean for a metadata stores project?

1. No more typing name-strings into web forms.

2. Agreed names/URIs for things like resource-types.

3. Sorting out URIs for things so that (unlike with Institutional Repositories) we can:

   1. Agree on terms before we start.

   2. Match-up terms later if we don't reach agreement.

Goal: Establish URIs for person-parties at UoX before we start

So, a simple matter of:

• Extracting name data from the existing institutional repository
  (the UoX model is to build their metadata store so collections are described in the IR)

• Disambiguating name-strings, for the usual reasons; there is typically more than one string used to refer to the same person and often the same string used to refer to more than one person. Sorry, party.

• Establishing new URIs for each party that the UoX cares about (they don't care about parties from UoY).

• Injecting the URIs back in to the IR.

Establishing IDs? How?

We considered these options.

1. Use People Australia / ARDC-PIP services. (It was a bit early)

2. Use Nicnames (plus).

3. Use a combination of the above.

We chose option 4

Build a name-authority server inspired by and informed by the NicNames work.

Why?

1. Can ship as an integrated part of the broader application we are building, in the same language (Java).

   1. Easy-install under the same web container as Fedora and Solr.

   2. Same configuration files as the “Ingect” application.

2. Having a local service means we can use:

   1. Private data such as staff-IDs internal to the application.

   2. Local URI-schemes for local things, such as internal projects.

3. Our service will deal with JSON formatted , not RDF to make it easy for web-interface designers.

The process – disambiguating names

1. Import all the name-string/publication pairs from the IR into a new repository (in EAC format) .

   That's one for each name on a paper. Records will be presented as citations keyed by a name.

   name:<name-string> id:<pID> title:<dc:title> (subject:<dc:subject>)*

2. Import a set of canonical names we care about into the new repository.

3. Turn all the canonical names from a local directory into master-records/name-packages.

4. Allow a data librarian to drag all the name-strings into the name-packages.

   [Jane Hunter CONTAINER RECORD

   Hunter, J: Gerber, A. and Hunter, J. (2008). A compound object authoring and publishing tool for literary scholars based on the IFLA-FRBR. In: C. Rusbridge, A. Trefethen and D. Berry, Proceedings of: 4th International Digital Curation Conference "Radical Sharing: Transforming Science?". ``4th International Digital Curation Conference (IDCC 2009)``, Edinburgh, Scotland, (1-10). 1-3 December 2008.

   Hunter, Jane: Hunter, Jane (2001). Adding Multimedia to the Semantic Web: Building an MPEG-7 Ontology. In:``International Semantic Web Working Symposium (SWWS)``, Stanford University, California, (). July, 2001.

   Hunter, J.: Hunter J. (1999). An "Improved" Proposal for an MPEG-7 DDL. ISO/IEC JTC1/SC29/WG11, 47th MPEG Meeting M4518, .

   ...

   ]

``

Process

But wait! There's more!

This new module, which is called, um … will:

• 'Mint' a new URI whenever someone types a string (in desperation).

• Provide locally available, fast services for accessing any ontology. Eg:

  • FOR codes.

  • Geonames.

One final thing

(And this bit is not ANDS funded)

We want to make it possible to encode and decode web-semantics in a URI. Eg:

http://ontologize.me/meta/?r=http://purl.org/dc/terms/creator&o=http://nla.gov.au/nla.party-541658

I can use this to assert that “ptsefton” is the author of this resource.

Would ANDS and/or the NLA consider supporting services like this?