Present:

• Les Car
• Gary Wills
• Arouna Woukeu

1. Discussion

Following on from the previous discussion (03 October 2003) we need to make clear what we mean by a Document Ontology.

A document is to do with protocol and information exchange. For example the EPSRC have a specification for a project proposal, you in turn fulfil this specification by returning a document.

A document is not just words with style/formatting, a document has words that carry knowledge (meaning). So a document can be viewed as a publication object, which is a carrier of knowledge.

From a WICK perspective, documents carry knowledge and makes reference to knowledge we have previously formally modelled and is held in other computer systems. A document contains more that data it has prose.

By contrast, in the document management production world, a document consists of structure and form, data/structure which can subsequently be interpreted with implicit stylistic attributes (you can check features via spreadsheet).

Let us look at a research proposal for a hypertext project. We require a:

• Document ontology.
• Hypertext Ontology
• Proposal Ontology, that will include the concepts of Principle Investigators, Work plans, Funding, dissemination.
• Research Ontology similar to the AKT ontology, which tells you about people, project, publications, etc.

For example let us examine a research proposal for Qusi Interactive Document System (QuIDS). We require:

1. Document ontology:- XML-FO and Dublin Core, etc
2. Proposal Ontology. This is a general proposal ontology can cover EU projects, NSF, local government grants, membership of societies, etc. It covers the purpose/genre/ raison d’etre/ why/question for which the document. For the EPSRC consists of a specification of what a project proposal should look like and a form.
3. Research Ontology, similar to AKT
4. Subject Ontology can use ACM classification. Hypermedia which has special, mix reality hypermedia, etc.
5. Narrative, we need to understand this better.
6. Project Ontology

The first thing the EPSRC form requires is form the Institution’s Name, and later the PI. The constraints here are that the person must be o that Institution, and of n appropriate grade (informally an academic).

EPSRC Form deconstructed should give us a rough proposal ontology. The guidelines should then supply the constraints.

Hypothesis: Can we use the form as the biases of the constructs/requirements expresses in the guidelines for the Case for Support.

For Example:

Part 1

• Bullet point 1:Literature review, the knowledge can come from ontologies c and d. We would need agent technology to workout significance of the work. We will need to be able to write queries to express the domain: Individual, lab/institution /country/collaboration/global all in the constraints of funding and collaboration. Is this a new knowledge service for AKT – next COP?)
• Bullet 2: another Bloom agent, evaluating outcomes of products
• Bullet 3; CV trawl- relevant papers and programs
• Bullet 4: Dynamic CV.

Part 2.

• Bullet point 1: Overall aims and individual measurable objectives. Requires a comparison of other projects.
• The program Objectives requires new created knowledge; expand the guidelines plus example from (non) funded projects. Requires a synthesis agent.

To Do

1. Ask Nick G about developing the ontology in OWL.
2. Talk to EPSRC reference the development of the new form (why was it changed)
3. Papers!!!!

• WWW- Good tool some background (14 November)
• JODI: Special Issue on Information Design Models and Processes (15 December)
• HT – Semantic Web and Tool (4 February, early bird)
• Web Eng Conference – Focus on Engineering (15 February)
• Web Eng Journal
• PAKM – Rational and tool