The first prototype is meant to give a preview of what could be the IntelliTun web interface. This gives an immediate feeling of what is the interaction with the system, while leaving a good level of separation between the interface and the back-end logic. The functionalities of the first prototype are: user roles and management (authentication, subscription); document work flow; document indexing and tagging (handmade, basic input fields); flat document storage; document search (according to basic indexing)

The IntelliTun web interface is based solely on Open Source components and software. Zope is the Web Application Server of choice, along with Python as the main programming language. Open standards such as XHTML and, generally, XML is preferred in any possible case. Since Zope will not hold the whole system and there are many external processes and components to interact with it, mainly for user management, users are stored in a standard SQL database, such as MySQL. This is accomplished with commonly available Python-MySQL and Zope-MySQL products. Document work flow is implemented with the help of OpenFlow, an Italian Open Source Zope product for work flow management. User interface is designed using XHTML and CSS, producing W3C compliant pages. Where required, pages are rendered via an XSL transformation. This is useful in the following prototypes, e.g. where interaction with OntoTunnel produces XML documents with the fields which the indexing engine has not been able to fill automatically. Documents are stored inside the ZODB, Zope's object database. ZODB's advanced features, such as dynamic object type definition and tagging, will not be used by this prototype. The document archive is initially flat, meaning that all documents are stored inside a single folder and indexed with Zope's basic indexing components (Dictionary).

With the second prototype, user and work flow management are improved. There is also a first interaction with OntoTunnel, the ontological layer of the Intellitun application, deriving forms for manual document indexing and tagging directly from the OntoTunnel specification, via XSL transformations. Users are prompted by the system about the upcoming deadlines, both via e-mail and via messages on the web interface. Documents are stored with versions, enabling the tracking of the various releases. Using Zope's content management capabilities, users are able to share documents between members of a group (previously defined by an administrator). Again, in below a list of the functionalities of the second prototype: User invitation via e-mail, OntoTunnel interaction (form generation), e-mail reminders, to-do lists, document sharing, document versioning, run-time document work flow modification, document search results sorted by relevance.

User invitation is managed via an external component which interacts with both Zope and MySQL, reading mail templates from the ZODB, filling them with data coming from MySQL, where the users' personal data are stored. This component is implemented from scratch as a Python script is invoked as a cron job, which parses a Mailbox file searching for accepted invitations and sends new invitations, assignments and reminders. Since only the ZODB-connected part (where mail templates are read to fill) is strictly related with the project architecture, this is a simple "just-do-the-job" component, which can be easily replaced with better Open Source solutions (provided they can connect to the ZODB and find there what they need). Both e-mail reminders and to-do lists are automatically generated by the system at user-interaction time, they can also be issued manually by the system administrator to a user or a group. System administrators are also able to move documents between work flows, if required. Search results are sorted by relevance, according to their tags.

The third prototype's functionalities are: automatic document indexing and tagging; document history logging; run-time new work flow creation; address book import; document ranking; and ZEO scalability.

The following functionalities will be provided in further implementations: automated reasoning; and peer to peer knowledge management.

purchase the full-text of this paper (price £20)

go to the previous paper
go to the next paper
return to the table of contents
return to the book description
purchase this book (price £80 +P&P)