Harmony - final summary report

JISC Information Environment Development Programmes

International Digital Libraries Programme

This form must accompany a detailed final report and a final expenditure statement for the project. The form and the expenditure statement will be for internal JISC use only.

1.

Project details

  • Project Director:

  • Project Manager:

  • Start Date:

  • End Date:

  • Reference number:

  • Names of any collaborating organisation:

  • Project URL:

Lesly Huxley

Pete Walker

July 1999

July 2002

CEI: JISC/NSF Digital Libraries Initiative

Institute for Learning and Research Technology (ILRT)

Cornell Digital Library Research Group USA

Distributed System Technology Centre AUS (DSTC)

http://metadata.net/harmony

2.

Title of Project

HARMONY

3.

Objectives

(list main objectives in order of priority agreed at the award stage)

A research initiative that investigates tools and mechanisms for resource discovery of multimedia digital objects. Aimed at releasing the potential within digital libraries to represent new types of content and deliver them in innovative way s

The research aimed to build on a number of high-profile digital library and metadata efforts with which the team (ILRT, Cornell and DSTC) were closely associated:

  • The extensibility of the Dublin Core metadata set for describing and enabling resource discovery of multimedia objects. Investigation of a qualification framework for the basic Dublin Core elements that extends its descriptive semantics to the specific as pects of multimedia objects

  • The utility of the Resource Description Framework (RDF) and the developing Dublin Core data model for expressing this qualification framework. Investigate and develop an RDF schema for the multimedia qualifiers and validate our work against this schema.

  • The utility of co-existing metadata standards within the multimedia digital object to express finer granularity resource attributes and facilitate object-specific resource discovery. Investigate and refine developing multimedia metadata standards such as the Multimedia Content Description Interface (MPEG-7) and the Synchronized Multimedia Integration Language (SMIL).

  • The suitability of an evolving digital object architecture - FEDORA - as the container for the multiple content and metadata streams of multimedia content and the mechanism for automatically generating metadata packages

  • The tools and mechanisms for indexing RDF metadata descriptions and processing queries against those indexes.

4

Highlights of project

(describe main achievements and key findings in bullet points)

  • Engaged with Dublin Core, XML, RDF, W3C, DARPA/DAML, MPEG, CIMI, RSS, OAI, DELOS and various other initiatives, with Harmony PIs often taking a leading role with specifications and Working Group efforts. (T1.1, T1.3).

  • Created a metadata proposal, ABC, that drew upon our experiences in the "metadata wars", ie. the periodic turf battles that we've seen amongst various metadata initiatives. Most notably, the INDECS critique of Dublin Core and the ongoing talks b etween Educational Metadata efforts and the Dublin Core group. (T1.1, T1.3, T1.4, T1.5, T2.1, T2.4, T2.6)

  • Organised the 1st Harmony ABC workshop (Jan 2000, Washington, hosted by CNI), enabling the project team to gain input from a number of experts from the field. (T.1, T1.3, T2.1, T2.4)

  • Established a high-profile collaboration with CIMI that has ensured a rich and diverse testbed collection from participating museums. CIMI Harmony Testbed data was became available within the project at the end of 2000, and is forming the basis for ongoin g refinement both of the ABC model, and the ILRT metadata query work which connects ABC-based metadata back to our original resource discovery concerns (T1.2, T2.3*, T3.4).

  • Organised (through the DELOS network) a 2nd workshop on ABC and "Ontology Harmonisation" (Rome 26-272001). Now that we have an implementation environment (ILRT metadata query engine) to evaluate real queries against real (CIMI testbed) data, the task of refining the ABC model can be revisited in more detail. The Rome meeting will compare/contrast the Harmony/ABC approach with the CIDOC ontology and related efforts. (T2.1, T2.4, T2.5, T5.1)

  • Continued involvement with XML schema, RDF schema, DARPA Agent Markup Language (DAML) and DC Architecture, tracking and participating in work on the integration of XML Schema Datatypes with the RDF metadata system. The addition of datatyping facilities is critical to ABC, since many multimedia related properties require representation as numeric of date-type values. Since Harmony's metadata model is based on Events and the representation of change and workflow structures, the ABC model requires a

  • representation of dates and times. Recent (Feb 2001) releases of the DAML language articulate an integration of XML and RDF that may support this. (T1.3, T1.4, T2.4)

  • Implemented, packaged and distributed (currently at release 0.7) one of the first RDF/XML query engines capable of providing mixed-namespace, RDF-compatible query support for metadata applications. (note: this work is similar to, but distinct from, our or iginal plan to build upon the Cornell Fedora tools). (T1.5, T2.2, T2.3, T2.6, T3.3, T3.4).

Also see details in final report at:

http://www.ilrt.bris.ac.uk/discovery/2002/07/harmony-final/

5.

Project Progress

5a) Did the project proceed as expected and on time ?

Research findings in the first year resulted in a redefinition of project deliverables. The relevant finding are fully documented within the mid-term report at:

http://ilrt.org/discovery/2001/03/multimeta/

The redefinition of deliverables is described in the following section

5b) Was there any significant change to the project compared with the agreed project plan?

During June/July 2001 a revised set of aims were agreed with JISC. While a measure of international collaboration was expected to continue it could not be relied on, so the revised aims focussed on what it was possible for the ILRT to produce independentl y from Cornell and DSTC (see Section 5c).

The changed aims and plan were approved by Joanna Harrison in a letter dated 14th August to ILRT.

A summary of the changes agreed are as follows:

  • Database and query of CIMI data - A demonstrator of querying CIMI data in various formats is specified in both the project proposal and in the original milestones plan (a prototype). Both ILRT and DSTC have written demonstrators which allow the querying o f CIMI data.

  • ABC formal schema (f2) - The ABC schema was the main focus of the original milestones plan and work has continued on developing it within the Harmony team.

  • SquishQL releases version 6 and version 7 - The SquishQL implementation is an essential component of the ILRT's Harmony deliverables, before and after the deliverables were revised. A query language and implementation appears in the project proposals, and in the original milestones plan.

  • Squish - a formal specification

  • Calendar Schema -The aim of this deliverable was to examine the possibility of modeling complex event data in RDF. Events are an essential part of the ABC model. Events are used in the model to link works, and their manifestations or versions at different times, and actions that occur to them, and the agents involved with them. Date and time constructs also appear in the model for dating manifestations of works and the changes that occur to them. Technical issues that occur in calendaring and scheduling o ccur in the storage and query of any event, in particular datatyping and querying datatypes.

  • Article for XML.com or similar on SquishQL - To promote the query language to relevant communities.

5c) Were there any circumstances that aided or impeded the progress of the project?

Impeded

  • The funding structure under which JISC funded ILRT, NSF funded Cornell and DSTC were self-funded caused difficulties. This was primarily because of the different funding bodies' reporting requirements and the notion of the flexibility of the project plan. There was no means by which overall coordination and direction could be imposed on partners.

  • Timezone issues were a particular difficulty with one partner in Australia.

Aided:

  • Contact with CIMI organisation who provided data for testing.

  • Contact with the RSS, OAI and Dublin Core communities.

6.

Development

6a) Has the project led to, or is it likely to lead to, further development work (please describe both US & UK developments)

The SquishQL/Inkling work will be continued within the EC funded SWAD-Europe

project. See

http://www.w3.org/2001/sw/Europe/

6b) Has the project led to improvements to the range of services the JISC offers? (identify any potential improvements too)

No, this is primarily research work. Potentially it can help with digital libraries modelling complex metadata.

7.

Project Staff

Original team

Dan Brickley, Libby Miller (ILRT, Bristol)

Jan Hunter (DSTC, Australia)

Carl Lagoze (Cornell USA - NSF funded

From July 2001

Libby Miller

Peter Walker (project management)

Plus ongoing assistance from Dan Brickley

8.

Collaboration

8 a) did any other external body, organisation or person collaborate in the project?

Hewlett-Packard Laboratories Bristol on Squish (Andy Seaborne andy_seaborne@hp.com)

Web Weaving Internet Engineering, Arnhem, The Netherlands

(Alberto Reggiori areggiori@webweaving.org)

8 b) please describe the nature of the UK/US collaboration?

See section 5 above.

9.

Expenditure

9 a) Has expenditure exceeded or fallen short of the total sum awarded under any particular heading?

£2,500 was not spent on an evaluation report. Sum returned to JISC

9 b) Were funds from other sources sought or used to support the project in any way

Please attach a separate final expenditure statement for the project

No

10

Dissemination of project outcomes

10 a) List, describe & give location of the following types of output arising from the project:

  • papers (refereed)

  • reports

  • articles (non-refereed)

  • conference proceedings

  • other (e.g. software/content)

Papers (refereed)

XMLEurope 2002

http://ilrt.org/discovery/2002/03/skical-daml/

ISCW 2002

http://www.ilrt.bris.ac.uk/discovery/2002/05/squish-iscw/

see here: http://metadata.net/harmony/Publications.htm

Reports

A list of publications is available at

http://metadata.net/harmony/Publications.htm

Also see

An RDFWeb Aggregation Service for the ILRT

http://ilrt.org/discovery/2001/11/ilrt-rdf-paper/

RDF Calendar taskforce

http://ilrt.org/discovery/2001/04/calendar/

Hybrid schema to to implement calendar/scheduling data

http://ilrt.org/discovery/2001/06/schemas/ical-full/hybrid.rdf

Articles (non-refereed)

RDF Query by example

http://ilrt.org/discovery/2002/04/query/

Software

Inkling: RDF query using SquishQL

http://sw1.ilrt.org/rdfquery/

Software demonstrations

http://sw1.ilrt.org/discovery/2002/07/abc/

http://sw1.ilrt.org/discovery/2002/04/rsscal/

http://sw1.ilrt.org/discovery/2001/12/xmlcal/

http://sw1.ilrt.org/discovery/2001/07/swws/

http://sw1.ilrt.org/discovery/2001/03/amol/

Squish defined in Backus Naur Form (BNF)

http://sw1.ilrt.org/rdfquery/squish-bnf.html

For more details on BNF see

http://foldoc.doc.ic.ac.uk/foldoc/foldoc.cgi?Backus-Naur+Form

10

10 b) Detail any other means whereby project outcomes have already been disseminated

Only as described in 10 a) above

10 c) Detail what publicity has been received for the project

None other than those described in 10 a) above

10 d) Did you refer to JISC as the provider of support?

Not applicable

11.

Records Management

11 a) Will the project website be available for three years or more after completion of the project?

The site will certainly be available for 2 years and it is highly likely that it will be available after that as well.

11 b) Can the website and project outputs be archived by JISC or other institutions and made accessible

It is likely that the RDF query software will be available for 2-3 years, and because it is open source it is possible to take a copy for archival purposes.

12.

Preservation Issues

12 a) Are there any preservation issues for materials arising from the project?

No

12 b) Have any arrangements been made with your host institution for preservation or maintenance and updating of the project outputs with long-term value?

Our organisation is dependent on short-term funding. If something is not explicitly funded within a project, then it cannot be supported. However, because some of the software is useful enough to be used in another project, it is likely that it will conti nue to be supported and hosted for the next 2-3 years.

13.

Beneficiaries of the project

Who are the likely beneficiaries of the project outcomes and how will they benefit?

Many members of the digital library community could benefit from Harmony. ABC allows very careful modelling of objects, including their change over time. However it is extremely complex to model and can be difficult to understand. For many organisations a part of the model can be useful but in its entirety it is too detailed. For this reason, in parallel to the detailed modelling of ABC, Harmony members have been working with the Dublin Core, RSS and OAI communities to increase the use of structured, mach ine-readable data. By creating machine-processable data in an interoperable fashion, they will benefit from reuse of that data, within and between organisations.

As previously mentioned SqusihQL/Inkling is already being used by other researchers as part of the SWAD-Europe project

14.

Detailed Report

The final Harmony report is available from:

http://www.ilrt.bris.ac.uk/discovery/2002/07/harmony-final/

15.

Signatures

Project Director

Project Manager

………………………………………………..

………………………………………………..

P.Walker Page 3 ILRT