This works presents an investigation regarding the computational support for net structured forum discussions. In forums, the discussion is usually structured on a tree format through hierarchically connected messages. The use of a more complex structure than a tree indicates some difficulties for the participants. It is more difficult to read, due to the lack of a linear structure, and to write, due to the possibility of more flexible associations among messages. The hypothesis being investigated is that, given a specific computational support, the participants of a forum will have an adequate participation in the net structured discussion; an adequate participation happens when a participant can understand the associations among messages (reading process) or when a new message is posted with associations in the discussion (writing process). The proposed solution was the usage of a multiple reference mechanism and the visualization and the way to build the forum net structure; these tools were implemented on the Conference Service of AulaNet environment. To investigate the proposed solution some forum sessions with Computer Science graduate and pos-graduate students were conducted. In these sessions, the forum discussion structure, its visualization and the way the discussion was built were varied with the aim at analyzing the influence of the student participation. From the case study conducted it was concluded that, despite the identified difficulties, the participants managed to follow the discussion and posted messages with associations, which indicates the viability of net structured foruns. Other conclusion was that the training with graphical tools to generate the net structured discussion is useful to eliminate some of the difficulties that is present on the more complex structured discussion.