Harmony and Synthesis

1
Harmony and Synthesis
Dan Brickley, ILRT. University of
Bristol danbri_at_bris.ac.uk
2
Harmony project overview

• Who?
• Dan Brickley, Libby Miller (ILRT)
• Carl Lagoze (Cornell)
• Jane Hunter (DSTC)
• What were trying to do
• simple vs complex multimedia metadata ABC
• Progress milestones and successes
• Challenges and lessons learned

3
Project Partners

• Instititute for Learning and Research Technology
  (ILRT), University of Bristol
• Cornell University, USA (NSF funded)
• DSTC (self-funding partner)
• common background metadata, digital images,
  Dublin Core, XML, RDF, extensibility architectures

4
Harmony original goals

• Extend Dublin Core for multimedia
• Model problematic cases
• versioning (document workflow)
• alternate manifestations (formats, conversions)
• content of composite media objects
• cross domain education media medicine
• Propose cheap common approach

5
What have we been doing?

• The Harmony collaboration has led to a metadata
  representational model (aka ontology, vocabulary,
  schema) and XML/RDF query implementation (query
  language and database search system) that
  supports resource discovery applications over
  multi-media, multi-vocabulary metadata
  descriptions.(!)

6
In a nutshell...
Metadata headaches, 1995-1999 works /
manifestations objects, their depictions and
creators agents, actions, change and
versioning... (everyone knew metadata could get
complicated)
7
We implemented it...
8
ABC at a glance
9
Query syntax
SELECT ?event, ?type, ?time, ?place FROM
http//ilrt.org/discovery/harmony/oai.rdf
WHERE (webtype ?event abcEvent)
(abccontext ?event ?context) (dctype
?event ?type) (abctime ?context ?time)
(abcplace ?context ?place) USING web FOR
http//www.w3.org/1999/02/22-rdf-syntax-ns
abc FOR http//ilrt.org/discovery/harmony/abc-0.1
dc for http//purl.org/dc/elements/1.1/
10
Implementation

• Ongoing development of ABC (NSF and DELOS funded
  workshops)
• Testbed collaboration with CIMI
• Prototype query language (Squish)
• Software query engine and database
• Implementation feeds into ABC model
• collaboration focus on metadata model

11
Implementation(2)

• Acronyms XML, RDF, W3C, DARPA/DAML, MPEG, CIMI,
  RSS, OAI, DELOS (dig-lib and beyond...)
• Testbed data CIMI (museums, images...), OAI
  archives, RSS feeds
• Schema language convergence (XML/RDF Schema,
  DAML)
• opensourced software Java, Perl

12
Example RDF/XML Queries

• Find resources whose dcformat is audio/mp3,
  whose dcrights is http//opencontent.org/opl.shtm
  l and which were the output of an
  abcCreationEvent involving an agent who works
  for the department whose home page is
  http//www.cs.cornell.edu

13
Was that a multimedia query?

• Multi-media content focusses our attention on
  mixing content from diverse contexts
• multi-media content requires combination of
  descriptive techniques
• multi-media data isnt a world apart
• avoid mm-specific technology if we can

14
Recap solving which problem?

• Q What has this got to do with multimedia?
  (scope creep?)
• A The scope creep problem!
• ABC and Harmony are a response to the metadata
  wars (DC/Education/INDECS/multimedia)
• ...and the too many committees problem

15
ABC in context (1999-2000)

• The Metadata wars
• domain overlaps
• Warwick Framework (myth...)
• Extensibility, qualifiers...
• Multimedia () metadata

16
Harmonisation using ABC

• Define some abstract, basic concepts
• A common representational style
• Supports vocabulary mixing
• XML/RDF encoding for exchange
• Focus on events, agents, relationships
• MPEG-7 RDF encoding by DSTC
• putting the multi in multimedia

17
Current Activity

• Ontology Harmonisation w/shop findings
• Preparing a 2nd DELOS w/shop
• OAI and RDF/RSS interoperability
• Refining the model representing events and
  states (accurate versioning audit trail)
• Track model with implementations

18
Project Management issues

• initial (too) tightly-coupled project plan
• Staffing (p/t PI as uk project manager)
• UK/US/Australia timezones, travel
• identifying common ground (ABC)
• Evaluation difficult for metadata formats (usage
  measures? Encourage adoption)
• dissemination (groundwork in place)

19
Lessons for DNER

• Metadata modularity is hard
• Technology in this area immature
• Standards expensive to invest in expensive to
  ignore
• Semantic Web and DNER wider computing industry
  landscape important
• The scope problem may have technical as well as
  social/political aspects

20
Lessons(2)...

• Data formats not protocols schemas not RPCs
• OAI and RSS 1.0, the two largest initiatives in
  this tradition, both developed with Harmony
  involvement, are maturing fast. Cheap technology.
• Tech complexity relates to staffing issues (file
  formats vs protocols)

21
Questions...
22
data merging / namespace mixing