Open sTeam is an infrastructure for computer-based cooperative learning. The project was started in 2000 and was supported by Germany's National Research and Education Network (DFN) until 2003. Today it is mainly run by the department "Computers and Society" at the Heinz Nixdorf Institute of the University of Paderborn in Germany. The initial idea was to develop an innovative system architecture for structuring information through an open source development community. Today open sTeam is part of the Debian Linux distribution. It is used to support eLearning, content management and knowledge management initiatives.

ARCHITECTURE

Open sTeam was developed in Pike, a dynamic programming language with a syntax similar to Java and C aiming to be "€œthe first scripting language for the Semantic Web"€ .

The open sTeam server is based on MUD (Multi User Dimension) technology, similar to what can be found in multi-player computer games. But the event-driven environment is based on a subscriber/publisher paradigm rather that on callbacks like in most gaming-oriented MUDs.

The sTeam infrastructure is open to integration into heterogeneous environments. The Pike server supports numerous protocols like http(s), ftp, irc, smtp, imap and jabber. APIs for Java, C++ and PHP ensure easy integration of external services as well as the integration of the sTeam system into existing infrastructures.

sTeam objects are stored in a MySQL database accessed through a persistence layer. Using MySQL to hold the search index may pose a slight performance problem for very large repositories as a relational database is not optimized for this use. An object-relational database like PostgreSQL could help support the systems architecture without have to do a relational mapping.

CREATE AND STORE

Uploading of documents is handled through the web client, through FTP or WebDAV or by sending an email to the server. The uploaded documents are organized in knowledge areas and virtual rooms in which the collaboration and content advancement is organized. sTeam allows for the creation of links to external object. E.g. you can add a PDF to the sTeam repository which is located on a server somewhere on the internet.

Metadata is not extracted from documents but so called keywords can be defined. Additionally one can add annotations to the uploaded assets. Annotations can be plain text as well as HTML and file attachments. The user is free to choose an icon to better distinguish different documents within rooms as well as a publication path and a style sheet used for inline display of the document.

ORGANIZING AND COLLABORATION

The virtual rooms represent central meeting points within areas. Rooms can be connected through gates. Folders help to further structure content below this level. Within those entities users can collaborate by maintaining a document base as well as chatting through an integrated Java chat tool or communicate through freely definable message boards. A calendar module supports the organization of schedules; a shared whiteboard is available for visual online communication.

Check in/out is not available but a function called "rucksack" allows users to take documents or copies of documents with them to other rooms or areas. Changes to documents result in a new version. Versions can be compared and restored.

Areas and containers can be monitored. Users will receive a message in case of certain modifications. This is particularly useful for areas and containers containing a message board. As soon as a new contribution is entered, users get notified through their sTeam mailbox or if desired by email.

SEARCH & ACCESS

Access rights can be defined per user and group for read, write, execute, move, insert, annotate and sanction. An object may acquire its access rights from its environment. If acquisition is activated, the access rights for a particular user or group are composed from those explicitly set for this object and from those of the environment. By default, acquisition is activated in order to simplify the rights management. The new version 2.0 (still beta) will allow for LDAP integration.

Workspaces can be accessed via ftp which allows for easy browsing and bulk download. Objects may be published in order to be freely accessible. A virtual path for publication may be specified at which the object will be published. Anybody may then access the object by specifying the server name, port, and virtual path in their browser. By exporting, the area or container including all or selected contents may be downloaded. The server creates a file with that content in compressed form (in tar format) and presents it for download.

Search allows filtering and full-text search covering the documents inline content, metadata and annotations. The search in the demo version was rather slow.

USABILITY

The metaphors used for the sTeam processes as well as the user interface are easy to understand and use. Some of them are not common for document management environments, features like the rucksack or gates have commoditized pendants in the document management and collaboration space but novices to this topic in general will still benefit from easy to understand functions.

The context-sensitive help has a well elaborated content.

COMMUNITY AND RESSOURCES

The usage and community around sTeam is mainly in German universities. Around 20 diploma theses have been written around this project dealing with various topics around organization of distributed learning environments. The results of this scientific work have all contributed to the feature set of sTeam in one or the other way. The website offers a lot of documentation and a well-maintained demo of the software.