A High-level Topic Maps-based Language for the Semantic Web Semantic Web EU/NSF Workshop, Sofia- Antipolis, Oct 3-5, 2001 Michel Biezunski ([email protected]),

A A HHigh-level igh-level TTopic opic MMaps-based aps-based

LLanguage anguage for the Semantic for the Semantic

WebWeb

Semantic Web EU/NSF Workshop, Sofia-Antipolis, Oct 3-5, 2001

Michel Biezunski ([email protected]),with the complicity of Steven R. Newcomb

([email protected])

© 2001, Michel Biezunski. with the complicity

of Steven R. Newcomb.

All rights reserved.

HTML still needed HTML still needed

• HTML has made the success of the World Wide Web– Simple, straightforward.– Powerful enough to get the job done.– Implementable, because of its

limitations.

• The Semantic Web need something comparable.

• We need something simple that the world can adopt.




OutlineOutline

• What topic maps are.– How they help differentiate

what's useful for humans and what's useful for machines.

• Research project proposals– How they help building

something clear, simple, useful, and appealing.• Usefulness, Scalability, ...




What are Topic Maps ?What are Topic Maps ?

• A standardized (ISO/IEC 13250:2000) paradigm in which:– subjects become

objects ("topics").– Relationships

between subjects are asserted.

Web links

Topic Map




Making subjects Making subjects addressableaddressable

• To be processed, subjects should first be made addressable.

• The purpose of Topic Maps is to make subjects addressable.

• Subjects are defined by humans, and can be (in specific contexts) processed successfully by machines.

• A "topic" is a subject made addressable:– It is a computer surrogate for a subject.– It is characterized by a set of structural

locations (names, occurrences, associations).




Subject Indicators vs. Subject Indicators vs. Subject ConstitutersSubject Constituters

• A piece of information can indicate what a subject is.

• A piece of information can be the subject.

subject

subject




Two faces of Topic Two faces of Topic MapsMaps

• KOAN: Knowledge, Occurrences, Associations, Names

• The core: Graph made of nodes and arcs.




KOAN: High-level KOAN: High-level Topic Maps conceptsTopic Maps concepts

Association

Topic

Occurrences

(Scope)

-Subject-Name(s) -Roles

-Members-Template

-Type




The Core: heart of The Core: heart of Topic MapsTopic Maps

S Wtormy eather

Lena Horne sings Stormy Weather :"Don't know why ..."

TTLenaLena

TTSong: Song:

"Stormy "Stormy Weather"Weather"

AA

Lena Lena performperform

s s "Stormy "Stormy WeatherWeather

""

Topic Maps CoreTopic Maps Core

scopecomponent

association

scope

association

member

association

member

scopecomponent

TTLenaLena

TTSong: Song:


AA



""

TTMovie: "Stormy Movie: "Stormy

Weather"Weather"

SS TTSoundtrack of ...Soundtrack of ...

scopecomponent

association

scope

association

member

casting

association

casting

member

scopecomponent

CCperformerperformer

TTLenaLena

TTSong: Song:


AA



""


Weather"Weather"


scopecomponent

association

scope

association

member

castingcasting

association

casting casting

member

scopecomponent


TTLenaLena

TTSong: Song:


AA



""


Weather"Weather"


CCmusicmusic

scopecomponent

association

scope

association

member

castingcasting

association

casting casting

member

scopecomponent

casting

role


TTLenaLena

TTSong: Song:


AA



""


Weather"Weather"


CCmusicmusic

TTperformerperformer

scopecomponent

association

scope

association

member

casting

role

casting

association

casting casting

member

scopecomponent

casting

role role

casting


TTLenaLena

TTSong: Song:


AA



""


Weather"Weather"



TTmusicmusic

CCmusicmusic

scopecomponent

association

scope

association

member

casting

role

casting

association

casting casting

member

scopecomponent

casting

role role

casting


TTLenaLena

TTSong: Song:


AA



""


Weather"Weather"



TTmusicmusic

TTperformanceperformance

CCmusicmusic

scopecomponent

association

scope

association

member

casting

role

casting

association

casting casting

member

scopecomponent

casting

role role

casting


TTLenaLena

TTSong: Song:



Weather"Weather"



TTmusicmusic

CCmusicmusic

AA



""


AALena performing Stormy WeatherLena performing Stormy Weather

Is an instance of a performanceIs an instance of a performance

CCinstanceinstance

CCclassclass

scopecomponent

association

scope

association

member

association

template

casting

role

casting

association

casting casting

member

scopecomponent

casting

role role

casting


TTLenaLena

TTSong: Song:


AA



""


Weather"Weather"



TTmusicmusic


CCmusicmusic

scopecomponent

association

scope

association

member

association

template

casting

role

casting

association

casting casting

member

scopecomponent

casting

role role

casting

CCrolerole


TTLenaLena

TTSong: Song:


AA



""


Weather"Weather"



TTmusicmusic


AAperformances have a performances have a

performer roleperformer role

CCtemplatetemplate

CCmusicmusic

scopecomponent

association

scope

association

member

association

template

casting

role

casting

association

casting casting

member

scopecomponent

casting

role role

casting

CCrolerole


TTLenaLena

TTSong: Song:


AA



""


Weather"Weather"



TTmusicmusic



performer roleperformer role


music rolemusic role

CCrolerole

CCtemplatetemplate

CCtemplatetemplate

CCmusicmusic

scopecomponent

association

scope

association

member

association

template

casting

role

casting

association

casting casting

member

scopecomponent

casting

role role

casting

CCrolerole


TTLenaLena

TTSong: Song:


AA



""


Weather"Weather"



TTmusicmusic



performer role; performer role; performers must be performers must be

musiciansmusicians


music rolemusic role

TTmusiciamusicia

nn

CCRecognized Recognized

Player of RolePlayer of Role

CCrolerole

CCtemplatetemplate

CCtemplatetemplate

CCmusicmusic

scopecomponent

association

scope

association

member

association

template

casting

role

casting

association

casting casting

member

scopecomponent

casting

role role

casting

CCrolerole


TTLenaLena

TTSong: Song:


AA



""


Weather"Weather"



TTmusicmusic




musiciansmusicians


music role; music must music role; music must be a songbe a song

TTmusiciamusicia

nn

TTsongsong



CCrolerole

CCtemplatetemplate

CCtemplatetemplate

CCmusicmusic



scopecomponent

association

scope

association

member

association

template

casting

role

casting

association

casting casting

member

scopecomponent

casting

role role

casting

CCrolerole


TTLenaLena

TTSong: Song:


AA



""


Weather"Weather"



TTmusicmusic




musiciansmusicians



TTmusiciamusicia

nn

TTsongsong

AALena is a Lena is a musicianmusician

CCinstanceinstance



CCrolerole

CCtemplatetemplate

CCtemplatetemplate

CCmusicmusic



CCclassclass

scopecomponent

association

scope

association

member

association

template

casting

role

casting

association

casting casting

member

scopecomponent

casting

role role

casting CCinstanceinstance

CCrolerole


TTLenaLena

TTSong: Song:


AA



""


Weather"Weather"



TTmusicmusic




musiciansmusicians



TTmusiciamusicia

nn

TTsongsong

AA"Stormy "Stormy

Weather"Weather" is a song is a song


CCinstanceinstance

CCclassclass



CCrolerole

CCtemplatetemplate

CCtemplatetemplate

CCmusicmusic



CCclassclass

CCinstanceinstance

CCrolerole


TTLenaLena

TTSong: Song:


AA



""


Weather"Weather"


scopecomponent

association

scope

association

member


TTmusicmusic


association

template



musiciansmusicians



TTmusiciamusicia

nn

TTsongsong

AA"Stormy "Stormy

Weather"Weather" is a song is a song


casting

CCinstanceinstance

CCclassclass



CCrolerole

CCtemplatetemplate

CCtemplatetemplate

CCmusicmusic



CCclassclass

casting

association

casting casting

member

scopecomponent

casting

role role

casting




What the assertions What the assertions means to humans:means to humans:

• Lena Horne performs Stormy Weather.• Lena Horne is a performer.• The music she performs is Stormy Weather.• The performance is in the context of the

soundtrack of the movie "Stormy Weather".• The assertion that Lena Horne performs Stormy

Weather, is a valid instance of a performance assertion.– The performer is a musician.– The music is a song.– Lena Horne is a musician.– Stormy Weather is a song.

What the machine What the machine seessees

What the machine really What the machine really seessees

casting node

topic node

association node

scope node

casting-member arc

association-casting arc

scope-component arc

association-template arc

casting-role arc

association-scope arc

Topic Maps CoreTopic Maps Core




Two faces of Topic Two faces of Topic MapsMaps

• Human– Topic (Subject)

• Names (scoped)

• Occurrences (scoped)

– Associations• (scoped)

• Machine– Arcs

• association-template

• association-scope• scope-component• association-casting• casting-member• casting-role

– Nodes• a (association)• s (scope)• c (casting)• t (topic)

•Complex models•Multidimensional, multilingual•Multiple perspectives




Research still neededResearch still needed

• Core of Topic Maps and RDF– Reification: early vs. late– Mathematical foundations– Scalability

• Improved Knowledge/Content Mangement– Ontology integration– Separating description from processing

• Building popular standards




CoresCores

• Expressing the Topic Maps Core as an application of RDF– and/or expressing the RDF Core as an

application of Topic Maps Core.

• Mathematical foundations for the data models

• Graph theory:– What are the characteristics of

systems that support the assertions in the graph model just described for the Topic Maps Core?




Early/late reificationEarly/late reification

• Topic Maps requires "pre-emptive" (or "early") reification of assertions.

• RDF requires "lazy" (or "late") reification.

• Early: Reification is not an issue, never needs to be discussed. Taken for granted.

• Early: Things can be reliably and permanently addressed in terms of the arcs that they terminate.

• Late: They can't.




ScalabilityScalability

• Speed (quick response)• Scale of diversity• Sheer volume• Usefulness (control infoglut

enough and still make the result useful).




Scalability: Quick Scalability: Quick responseresponse

• Expert systems are designed to integrate everything (e.g., Cyc).

• We need ways to make knowledge useful and findable without fully understanding or integrating it into a single expert system.




Scalability: Scale of Scalability: Scale of DiversityDiversity

• Information resources not based on Platonic forms -> Difficult to understand.

• Likely to accumulate all kinds of inconsistency.

• Number of relationships grow exponentially.




Scalability: Sheer Scalability: Sheer volumevolume

• What problems appear at high scale?

• Test the theory: we can achieve very high scale integration with Topic Maps while preserving usefulness: demonstrate that, say what are the problems.




Scalability: Scalability: UsefulnessUsefulness

• Generic Platonic-form-oriented search engine technology needs to be developed. Such search engines will aggregate knowledge around a broad range of forms.

• The Platonic forms supported by existing search engines are mainly words and phrases in one or more specific natural languages. The "meanings" of these words and phrases is far more interesting than the words themselves.




Improved Improved knowledge/content knowledge/content

managementmanagement• How can the Web become a

friendly place for all kinds of content to be "owned and exploited" as opposed to merely "published" in one or more renditions?




Ontology integrationOntology integration

• Ontology integration and Topic maps paradigm

• Concepts need to become reliably addressable– regardless of their names,– regardless of the namespaces within which

they have those names.

• Leverage everybody's work in – identifying concepts, – describing them,– collating/organizing reference materials– codifying relationships between concepts




Intellectual property Intellectual property rightsrights

• How can intellectual property rights be verified?

• How can a piece of content be made to "know" who claims ownership?

• How can the owner's policies be enforced? What business models are supportable, and what enhancements to the infrastructure of the Web itself can or should be made in order to support the enforcement of content owner policies?

• What kind of cooperative ownership of intellectual property must be supported?




Build a consensusBuild a consensus

Within and between communities.– Build standards which are going

to be used, not only appealing models.

– Understand what everybody is doing.

– Open our own ears

• Remember HTML




Machine-driven processes(DAML+OIL)

Semantic Web

A proposal for a A proposal for a global architectureglobal architecture

Knowledge Management,Metadata models

RDF / Topic MapsOntologies

Documents(SGML, HTML)

Databases(Schemas)

XML

A High-level Topic Maps-based Language for the Semantic Web Semantic Web EU/NSF Workshop, Sofia- Antipolis, Oct 3-5, 2001 Michel Biezunski ([email protected]),

Documents

stormy weather t movie

stormy weather st soundtrack

subject slide

cmusic tperformer slide

type slide

complicity of steven

michel biezunski

tperformer tmusic cmusic