semantic frameworks: taking stock

1
semantic frameworks taking stock

• Steve Harris
• Oxford University
• Computing Laboratory

2
reusing data is hard

• representation
• identification
• atomic meaning
• context
• user interfaces

3
let me tell you what I want

• LAMP
• integrated stack of
• Linux
• Apache
• MySQL
• PHP
• out-of-the-box web development framework

4
what I really, really want

• a LAMP like stack for semantic software
  development
• a set of tools for developing software that is
  parameterised by knowledge
• all tested and integrated that I can install and
  configure for a data community
• what components are available?

5
requirements

• early adopters
• identification and resolution
• community ontology
• metadata
• models
• workflow
• generative framework
• user tools

6
data community

• need a community with low activation energy wrt
  adoption
• hard problems
• clear benefits
• deep pockets
• candidates
• cancer research
• The War On Terror
• e-Government

7
identification and resolution

• resource identifiers must never be deleted
• retire or supersede
• must contain
• issuing authority
• resource
• resource release
• element identifier
• element version
• consistent policy for resolution
• supporting resource distribution, local theories

Who is Number 1? You are Number 6.
8
standards

• URI
• XRI

9
open source implementations

• XRI
• questions over XRI seem to be resolved
  http//en.wikipedia.org/wiki/XRILicensing
• resolver?
• best practices for implementation of REST
  interfaces for ontologies and metadata
• tools for maintaining mapping between physical
  locations and URIs
• http//www.w3.org/Provider/Style/URI

10
community ontology

• makes the meaning of concepts clearer
• definitions
• synonyms
• consistent identifiers
• basis for annotation
• conceptual mapping
• human discovery
• model checking
• supports local logical theories

11
types of resource

• terminology/dictionary
• words and definitions
• thesaurus
• including synonyms
• taxonomy
• subsumption
• conceptual model
• more complex knowledge
• logical theory
• rich, complex, consistent, meaningful

generality
utility
12
standards

• OWL
• SKOS
• Topic Maps (ISO13250)
• ISO704
• ISO1951/12200
• common logic (ISO24707)

13
open source implementations

• representation OWL W3C
• authoring Protégé Stanford University/Mancheste
  r University
• serving LexGrid Mayo Clinic
• user interface terminology open portal NIH,
  Mayo Clinic, Stanford, cancergrid (under
  development)
• web services cancergrid, common terminology
  services (CTS) Mayo Clinic

14
metadata

• lists attributes appropriate to the data
  community
• organises terms into atoms of reuse
  pre-coordinate ontology mark-up
• ensure consistent presentation of annotations
• compose coordinated models
• capture complex measurement protocols

15
standards

• ISO11179

16
open source implementations

• cancergrid
• caDSR

17
static models

• describe interfaces, (minimum) data sets,
  documents, messages
• offer reusable model fragments
• semantic lego
• address model, form model
• standard forms and sections
• ontology -gt class names, annotation
• metadata -gt class attributes

18
standards

• RDBMS schema
• XML schema
• SOAP
• UML class diagrams in XMI

19
open source implementations

• are there any products that output XMI?
• cheap alternative Sparx Enterprise Architect

20
process models

• dynamic operations
• processing and flow between models
• framework for branching
• uses models for start and goal
• uses ontology for names

21
standards

• BPMN
• BPEL

22
open source implementations

• Active BPEL engine and designer
• Taverna

23
generative framework

• essential
• intended semantics are faithfully preserved by
  the delivered system
• specification
• interoperability
• language
• governance
• flexibility for behaviour to change as the
  community requires

24
standards? implementations?

• rolling our own service generation has been
  trouble free
• OS portal software NOT a good target for
  generation
• JBoss
• GridSphere
• excellent results with SharePoint 2007
• Veterans Health Administration

25
user tools

• semantic plugins
• word processor
• terminology/thesaurus for authoring,
  disambiguation
• document metadata for annotation
• spreadsheet
• metadata for column typing
• UML designer
• terminology for class names, process descriptions
• metadata class attributes
• ontology for model checking
• desktop XForms
• as UML designer

26
open source implementation

• MS Office
• for Office 2007
• (nearing completion) cancergrid
• for other versions
• MIMIx for excel from the EBI
• Open Office
• ?

27
extraction tools

• transformation XML into RDF/Topic Maps
• unpack spreadsheet
• strip cruft
• retrieve element and term definitions
• transform into triples
• serialise to XML
• early tools in cancergrid for schema/XMI

28
conclusion

• nearly complete coverage
• tools already display significant
  interoperability
• compelling user and developer interfaces required
• could this be the next AJAX?
• we just need a catchy name

29
user tools
30
(No Transcript)
31
enumerated conceptual domain document
non-enumerated conceptual domain document
non-enumerated value domain document
data element document
enumerated value domain document
32
(No Transcript)
33
(No Transcript)
34
(No Transcript)
35
(No Transcript)
36
(No Transcript)
37
(No Transcript)
38
(No Transcript)
39
(No Transcript)
40
(No Transcript)
41
(No Transcript)
42
(No Transcript)
43
(No Transcript)
44
(No Transcript)
45
(No Transcript)
46
(No Transcript)
47
(No Transcript)
48
(No Transcript)
49
(No Transcript)
50
(No Transcript)
51
(No Transcript)
52
(No Transcript)
53
(No Transcript)
54
(No Transcript)
55
(No Transcript)
56
(No Transcript)