Transcript Document
A Developer’s Viewpoint Prof Mark Baker School of Systems Engineering University of Reading Tel: +44 118 378 8615 E-mail: [email protected] Web: http://acet.rdg.ac.uk/~mab May 21, 07 [email protected] Outline • Who am I? • Sakai VRE Demonstrator: – Overview, – Enabling Technologies: • WSRP, • Bridges. • • • • VERA. Cross Database searching. Thoughts. Conclusions. May 21, 07 [email protected] Who Am I? • Research Professor at University of Reading: – Run a group involved in developing middleware for Parallel and Distributed Systems. • Worked on a number of Grid-based projects - test beds, performance, and tools. • Based on this experience we have produced a number of tools, including: – MPJ Express - Java version of MPI – Tycho - A wide area messaging framework with an integrated registry. – GridRM - A wide area monitoring system, – Slogger - Semantic Web-based system for distributed log analysis. • More Recently we have been involved in creating tools and services for Portals - using portlets - hide backend complexity. May 21, 07 [email protected] Sakai VRE Demonstrator • Partner in the JISC VRE I funded Sakai Demonstrator project (ends June 30 2007). • Other partners: – Lancaster University (lead), – Oxford University, – Daresbury Laboratory. • The aim of project was to create a working demonstrator of a VRE, based on the Sakai portal framework, that would enable and engage researchers and scientist to explore the benefits and capabilities of such a VRE for undertaking their work. May 21, 07 [email protected] Some Tools Developed • Standard tools based around the JSR-168 portlet API… • WSRP - used to access resources from Sakai. • Agora Communication Technology - being developed to support a suite of communication tools for Sakai. • Bridging Technologies have been investigated to link to legacy applications written in Perl, Python and PHP. • Security - Shibboleth Integration to allow VRE users to authenticate in the same way as they do to see other JISC-supported resources such as information repositories. May 21, 07 [email protected] Web Site - http://tyne.dl.ac.uk/Sakai/ May 21, 07 [email protected] Enabling Technologies WSRP/Portlet-Bridges May 21, 07 [email protected] WSRP Services Plug and Play Portlets Registry Find Portals Portals Clients Clients Web Web Web Web Web Web Clients Web Clients Web Clients Clients Web Web Clients Clients Clients Clients Clients Clients Portals Portals Portals Portals Portals WSRP Portals WSRP Portals Portals Services WSRP Portals Portals Services Services Bind e.g. May 21, 07 [email protected] Publish WSRP - Jaffer Imported into Sakai May 21, 07 [email protected] PHP-JavaBridge • Described as an XML-based network protocol. • Designed to communicate with native scripting engines, which have a Java or Common Language Infrastructure virtual machine. • Provides a suitable environment for PHP applications to run under Apache Tomcat. • Allows Java and PHP to share sessions, in a J2EE environment. • Has a number of configurations for installation and connections to existing Web servers. May 21, 07 [email protected] Examples: BibAdmin Portlet May 21, 07 [email protected] Examples: Guest Book Portlet May 21, 07 [email protected] Examples: Guest Book Admin Portlet May 21, 07 [email protected] Examples: Image Gallery Portlet May 21, 07 [email protected] Virtual Research Environment for Archaeology (VERA) • JISC VRE II funded project. • Addressing user needs and aiming to enhance the means of efficiently documenting archaeological excavation and its associated finds. • Creating a Web portal that provides tools for the user community, and developing utilities that help encapsulate the working practices of current researchers not so familiar with the research environment. May 21, 07 http://vera.rdg.ac.uk [email protected] Cross Database Searching • It is increasingly clear that researchers want to be able to search across multiple interdisciplinary databases • The VERA project wants to identify and research their finds and artefacts by potentially undertaking searches across multiple archaeological databases! – This could be aided by also being able to search other databases to identify manuscripts, graffiti or geographical features. • The infrastructure to create portal tools for searching across database is really very easy! – However, there are no standard ways of categorising or describing the data and information contained in the distributed databases, which means the task becomes tedious/hard… May 21, 07 [email protected] Cross Database Searching • Need to know about the database schema, which describes tables and records, to be able to undertake these searches coherently. • Lots of interest in Semantic Web technologies for this purpose - XML schema, RDF, RDFS and Ontologies (OWL): – But, whatever technology you use, it is still necessary to infer terms in one database against the other… • i.e. a search for “street” may return nothing if the database calls this information “road”. May 21, 07 [email protected] Thoughts • From a scan of the areas we are interested in (Archaeology and Manuscript databases) it appears they use catalogue numbers and have limited metadata about entities… • Whatever search API we decide to potentially use, e.g. OpenSearch or SQL92, some work needs to be undertaken to create/provide additional metadata to aid the search process. • Ideally, from a users point of view: – They login to a portal, – Hit the search page, – Then use pull downs and buttons to click on the databases, and terms they are interested in, – Finally adding some additional text to refine the search, – Click on go and lots of useful data and records are returned that they than refine or analyse, – Potentially may also want to save search and additional data. May 21, 07 [email protected] Thoughts • The VERA project are going to work with Oxford’s Manuscript project to create a prototype portal service that can be used to do search across Archaeological and Manuscript databases. • We are going to choose two databases and explore the best way of creating the additional metadata and cross database terms to allow efficient and effective searches: – Note - we do not own the databases, so will probably have to use a third party service for the annotation services. • This may involve marking up each database in XML, and mapping terms with technologies like XLST or OWL: – Unfortunately, there is no automatic way of mapping these terms, so this will involve experts working together with CS people in the first instance. May 21, 07 [email protected] Conclusions • Cross database searching is becoming increasingly important: – Not just in library systems. – Addition data can enhance all manner analysis undertaken by researchers. • Few standards in the area we are interested in. • Limited metadata associated with entities on database, • Makes the process of cross database searching tough, and probably a very manual process in the first instance to start. • We need to explore automatic ways of inferring terms - would be useful to get AI folk working in this area. May 21, 07 [email protected]