RELEASE-NOTES

====================
iServe Release Notes
====================

iServe 2.0 includes many new features, bug fixes and performance improvements.

iServe requires Java 1.6 or higher. The distribution should be runnable in
most J2EE servlet containers.

Requires SPARQL 1.1 compliant RDF store

Requires a Servlet 2.5 compliant Web application server, e.g., Tomcat 7

NOTE: Old configuration files will need to be updated due to changes in the configuration
mechanism.

General
=======
- The system has been ported to Jena as the base library for RDF and SPARQL
- Developed a new in-memory Java model and the corresponding parsers and writers for easier management of services data
- Updated MSM model (new namespace, new properties)
    - Added new WSDL grounding details
    - Use of dc:source rather than rdf:describedBy
- RDF Stores have been decoupled from the software by using SPARQL Updates and the Graph Management protocol instead
of direct access to concrete registry implementations. As a consequence a slight performance overhead is incurred but
the system is entirely decoupled from concrete RDF stores implementations.
- iServe now makes use of dependency injection and automated implementation loading to provide further modularity
and extensibility
- Data management has been adapted and makes further use of named graphs for better data management and performance
- Included a notification mechanism as part of the infrastructure to support updates when changes on the underlying data
occur. The implementation is based on Guava's event bus
- Better centralised configuration mechanism (old configuration files will need to be updated)
- Re-engineering of modules:
    - Consolidated modules for simplicitiy
    - More structured interfaces
    - Added notion of Knowledge Base manager (implementation with local indexing and/or remote SPARQL service)
    - Added notion of Non Functional Properties Manager
- Added new GUI based on ELDA for visualising and querying the registry
- Provisioning of Read/Write Linked API for service descriptions
- Added Ontology Crawling mechanism that automatically obtain required ontologies
- Significant performance improvements both for Services manipulation as well as Concept matching
- Added Indexing and Caching mechanisms for much faster discovery
- Added Notification support for discovery queries so that clients may register queries to be notified for changes
- We now do not require backends to support reasoning and compensate this using SPARQL 1.1 variable length paths
- Bug fixing


Import
======
- Import mechanisms now based on plugins that can be loaded at runtime
- Updated OWL-S import - provided also as standalone executable application
- Added mechanism added for the Web Services Contest
- Improved automated support for uploading referenced ontologies both synchronously and asynchronously
- Added basic roll-back features for dealing with partial imports
- External datasets have now been integrated and made available via maven automated downloads for testing and development
- Added Swagger 1.0 & 1.2 import support

Discovery
=========
- Discovery mechanisms are now pluggable and detected automatically by the system
- Default discovery implementation can now discover operations and services
- Discovery plugins are now accompanied by their own internal documentation exposed also via an interactive documentation GUI
- Advanced pluggable Scoring and Ranking mechanisms implemented for better results ordering
- Discovery results now include details on the inner matches for debugging and 
better scoring purposes.
- Homogenised discovery result types within the standard discovery plugin
- Scoring and ranking is only performed at the end to avoid recomputations
- Developed discovery implementation with in-memory or Redic backed local indexes for much faster discovery responses
- Added freetext discovery
- Added result caching and notification support for "continuous" discovery queries