Architecture de SDX-2

SDX-1 matched its primary goal: publish and search acollection of XML documents. For that purpose, various components have been installed:

A typical application was composed of four parts : presentation of documents, hierarchical browsing, simple and advanced search and management of the documents collection. The model corresponded to a common Web application, an interface relying upon an information system.

Cocoon and the XSP language have appeared very convenient to define server pages which make an application. Version 2, compatible with industrial exploitation, has led to books published y interest of large companies. Like other applications, but often before, SDX-2 has been built as a direct extension of the Cocoon servlet (SDX server). Thus, developping for SDX ensures to improve standard competences (XML, XSLT, XSP), aiming at spreading it.

The Lucene search engine is theoretically possible to replace (later), however in practice, few open-source components could reemplance it. The definition of a document repository is taking advantage of it. Another XSLT processor can be selected with no risk of uncompliance, with the condition that it is made in Java (eg : Xalan). Like Cocoon, SDX-2 can be installed on any other server using the Sun servlet architecture. A servlet engine is designed for being compatible with most HTTP servers. Finally, the primary objective of SDX-2, data sources can be of a great diversity, leading to the generic concept of repository (with no requirement of any external database for example).

This flexibility adds value to a document management application. Instead of adding an interface to a document collection, it is possible to start by building an editorial project, consider an audience, a prototype. Adding the necessary documents is a second stage, SDX must easily match technical questions.