Shoulder to shoulder collaboration

Most computer applications written today are single user applications. They are intended to run on a single workstation controlled by a single user. Systems that support group collaboration are traditionally run on multiple workstations communicating across a computer network. In addition to supporting collaboration between remote participants through peer-to-peer communication the Pond also supports some of the ideas in the Single Display Groupware (SDG) model . SDG focuses on co-present collaboration with an input channel for each user and a single shared output channel. The shared output channel in the Pond consists of the visual and aural rendering of the Pond based on actions performed by all participants. The input channels are the pressure pads, the RF tags, the touch screen and the video tracking.

In the available version of the Pond, one application could be in a frequent party-situation, as a tool for a small group to discuss or debate music or to argue about their musical preferences.
Behavior and interactions


A user is given a token (i.e. RF tag) preloaded with links to some music. The user moves up to the pond table and puts down the token on the tag reader. After a few seconds, a music icon rendered with information from the album cover will appear in the pond and the preloaded music will start to play. Soon thereafter related music selections will begin to surface and swim around. There is a continuous supply of new items introduced as if coming from a stream upwelling in the center of the pond. Items can also be introduced by a user with a set of music style tags. As each item reaches the surface a few seconds of a song from the CD that item represents is played. The item then slowly starts to float round the pond. If a user selects an item when it floats by him the item will rise up towards the surface and again start to play one of the available audio clips. At the same time other related items in the pond will start to move towards the selected item. If the user decides that the item wasn’t really appealing and deselects the item it will sink down into the pond and disappear. Furthermore, the flocking behavior of related items will stop and the circling around the pond of the other items will continue. If, on the other hand, the user continues the selection the item will move to a spot in front of the user and position itself there and the related flocking items will stay in the vicinity and

wait to be selected as well. If a selection does not occur within a specified timeframe those items will sink into the depths of the pond and disappear. The items having been collected in the above way will constitute a school of aquatic creatures flocking around the user. This group or flock can then in a very simple manner be "offloaded" to a tag of the users choice. If the user at a later point in time puts the tag back on the appropriate tag reader the members of the group contained in the tag will be reintroduced one by one into the pond and the whole process starts all over.
Search refinement is done by the system detecting user actions that convey interest in an item such as playing a large set of the available clips. With the help of the related item information on the Amazon.com database, the system can insert new items that have a high chance of being interesting to the user as well. Combined with the fact that items considered uninteresting will sink and disappear, this has the effect of gradually refining the ecosystem to contain more and more interesting items. This procedure also allows for moving through the set of available items in that data source without burdening the user with endless dialogue windows and manual cleanup operations. So even when very specific selections are being used as "seed" material, the pond will at best have represented in it several general tastes, specific favorites and complete wildcards
The Sonic Interface


In the present version of the Pond, sound is used at two levels. The first level is the ambient environment. Since the metaphor is a "pond" or "pool" of information, the soundscape should consist of a family of wet, bubbling, splashing and whirling sounds, giving the impression of the information-units (the music records) ascending from an abyss of ooze or mud. The ambient sounds fades out when the chosen music-sample is played, and fades smoothly in when the music stops. The second level is the interface-sounds indicating user-presence, selecting/unselecting, clicking, dragging etc. These sounds do also originate from the above-mentioned family of pond sounds. The music samples associated with the Amazon.com database of CDs is a further level in terms of audio. These music clips are played "outside" (i.e. not spatially rendered) of the 3D pond environment using an adapted Real Audio player.
Contact: Lennart E Fahlén