Lessons learned from Semantic Wiki
Download as PPT, PDF0 likes276 views
This document summarizes lessons learned from developing semantic wikis. It discusses how semantic wikis differ from traditional wikis by embedding structured metadata and propagating that metadata via semantic queries. It then outlines key features for different user groups, including improved data generation and propagation tools for end users, and light-weight data modeling and fast prototyping for developers. Remaining issues are also discussed, such as managing public and personal data, improving scalability, and data portability and protection across multiple wikis.
Lessons learned from Semantic Wiki
- 1. Lessons learned from Semantic Wiki Jie Bao and Li Ding June 19, 2008
- 2. Background Wiki => MediaWiki => Semantic MediaWiki (SMW) => SMW workbench How semantic wiki differs from wiki? Embed structured metadata to wiki pages Propagate metadata to other wiki pages via semantic query
- 3. Semantic Wiki @ RPI Social Webtop* Workbench on the Web – more data Social environment – more human interaction Semantic Web enabled – more structure Highlights Support Users at different levels Many end users Concentrate on data (easy data input/output/provenance) Some application developers Light-weight data modeling Fast prototyping Few SemWiki experts Application incubator * http://en.wikipedia.org/wiki/Web_desktop
- 4. For End Users Improved Semantic Data Generation Before Use RDF syntax Use Editor After From structured data Use SMW syntax Use Form <?xml version="1.0"?> <rdf:RDF xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#" xmlns:dc="http://purl.org/dc/elements/1.1/"> <rdf:Description rdf:about="http://www.w3.org/"> <dc:title>World Wide Web Consortium</dc:title> </rdf:Description> </rdf:RDF>
- 5. For End Users Smart Data Propagation Data can be easily propagated via semantic query, e.g. the project and publications are included from others’ input Data can be propagated smartly using wiki-script based inference, e.g. “James A. Hendler” has an alias “Jim Hendler”
- 6. For End Users Supports Provenance Tracking Provenance metadata natively captured and semantically encoded by PML Enable complex query on data and provenance
- 7. For Developers Light-weight Data Modeling A SMW page can be viewed as: One row in relational table - semantic template One object in frame systems – semantic template + category A node in RDF graph – SMW syntax
- 8. For Developers Enables Fast Prototyping Learning Curve (before) Web languages: HTML, CSS Semantic web languages: RDF, RDFS, OWL, SPARQL, Web tools: web server (apache, tomcat), database (mysql), languages (PHP, Javascript) Semantic web tools: editor (protégé, swoop), triple store (sesame), ontology reasoners (pellet, racer), Programming API (jena), Learning Curve (after) Wiki syntax Semantic wiki syntax and semantic form/template syntax Web tools hidden by Wiki Simplified SW tools via light-weight Wiki programming: Online authoring Triple store Semantic query with simple RDFS reasoning
- 9. SemWiki as Application Incubator Semantic Group Portal Group portal (tw.rpi.edu): Project, People, Event Multi-group portal (tw.rpi.edu/portal): protection Mash-ups RPI Map (map.rpi.edu) – import data Wine Wiki (onto.rpi.edu/wiki/wine) – export data More Prototypes on tw.rpi.edu Semantic Calendar – user experience Semantic Blog – protection Collaborative literature annotation – convergence Semantic Issue tracker – transactional data
- 10. RPI Map = SemWiki + Google Map http://map.rpi.edu/ Conventional mash-up Wiki data: location, name, alias, photo Data imported from GoogleMap: map, location service Technical highlights Import GoogleMap via MediaWiki’s parser function extensions Allow use-contributed data via semantic form/template
- 11. SemWiki + OWL = Wine Agent http://onto.rpi.edu/wiki/wine Semantic mash-up Collaboratively maintained taxonomy (class hierarchy), e.g. Food Can be integrated with external complex ontology, e.g. wine.owl Data exported to TW Wine agent ( http:// onto.rpi.edu/WineAgent/WineAgent.html ) Technical highlights Used as KB by TW Wine Agent Preserve provenance
- 12. Collaborative Literature Annotation Context A publication can be tagged with different topics in topic hierarchy Users may hold different topic hierarchy Goal: to converge and branch topic hierarchy among users Example http://tw.rpi.edu/wiki/index.php/Literature . we may build two topic hierarchies TW Topic: for TW related topics Topic: for general computer science
- 13. For SemWiki Experts A lot of work has been done 9 June 2008: SMW 1.2 preview: faster more powerful queries Semantic Form and Template extension allows users to have forms for adding and editing data on wiki, without any programming SMW halo extension improves UI for ease of use of Semantic Wikis for a large community of users Wiki parsing function extensions are heavily used to enrich SemWiki
- 14. For SemWiki Experts Many Remaining Issues! Convergence How to help users to manage public and personal data in information sharing context, e.g. a group website Scalability how to support more semantic data with acceptable response time Data Portability How to import/export semantic web data from/to external sources How to allow user registered across multiple wikis with preferences, e.g. data access privilege Data Protection How to protect data read/write access