Everyone is different, and software systems should take this into account. By “putting the user in the centre”, we mean that the KiWi system allows to tailor the presentation and functionality of the platform to the information needs and experience of the user, and to make the use of the platform as easy as possible.

“Information needs” materialise in the explicit or implicit preferences and interests of the user regarding e.g. user interface or topic, in the context of the currently presented content, and in the current context of the user, e.g. “at home”, “at work”, “travelling”.

In KiWi, preferences will be represented by appropriate user models that may even be generated automatically by observing user behaviour. The context of the currently presented content is made explicit to the system by appropriate annotations and tagging, (and the current context of the user by allowing users to switch between different roles).

For instance, the system could observe that a user is frequently accessing content that is concerned with “portrait photography” while she is in the role “at home” and derive from this that this is a topic of interest for her. For portrait photos, the system could offer a specific presentation that is different from other photos, e.g. by displaying a user profile box of the depicted model. While “at home”, the user could then get a list of recent “portrait photos”, whereas she would get a list of recent posts to the company blog while “at work”.

User interface adaptation will be supported by allowing advanced users to expand the platform by custom widgets, and by defining custom layouts that allow arranging widgets in context dependent ways in the browser window.

For example, there could be a custom layout for portrait photos as described above, and the user could have a personalised start page displaying widgets with information that is relevant for him or her only.

In addition to personalisation and adaptation, the KiWi system as a platform will provide generic support for common tasks to make using the platform as easy as possible for the user. Widgets are not only suitable for the presentation of information, but also for interaction with the system.

Developers and advanced users could thus easily provide so-called “Semantic Forms” (as e.g. in Semantic MediaWiki ) that make it easy to enter content in a structured way. For example, an editor widget for “portrait photos” could offer a field to either provide a link to an existing profile or to immediately enter information about the depicted person.

Information extraction will support the user in efficient searching and browsing as well as in tag recommendation. For example, the information extraction component in KiWi could automatically detect that the user is entering “Paris” and ask whether he means “Paris, capital of France” or “Paris, first name”, and tag the content appropriately.

Furthermore, a rule-based reasoning system will allow deriving implicit information from what is given explicitly. Such reasoning could also be used to make recommendations, e.g. of the form “content tagged with jaguar, the animal, should also be tagged with animal”.

Go back to the beginning of this section