The present thesis is an overview and survey of interaction with real-time, text-based, multi-user systems, especially MUD systems, by persons with disablements. It is argued that these types of systems could be excellent "social tools" for those who cannot participate in social and leisure activities as easily as users without disablements. Based on the exploratory studies of two groups of such users, documented in this thesis, a preliminary conclusion is that one of the major software interface problems is information overload. The second major problem is that information is not always presented in the most appropriate mode for a particular disablement, and thus interaction does not take place in the most appropriate modality. Above all, this is a "white paper" or vision document, proposing one philosophy for developing augmentative and alternative communication systems. This philosophy has three main foci: Firstly, let the user/consumer, not the producer, decide how information is presented and how interaction is to take place. Secondly, explore minimal solutions that use existing services and that do not require the user to acquire new equipment, both to minimise disruptions in how the user normally does things, to minimise the cost, and to avoid creating special purpose systems instead of integrating with existing services. Lastly, be visionary. Instead of correcting things that do not work, try to create solutions that take disablements into account from the very start ("universal design"). We believe this to be absolutely necessary if we are to avoid the current situation where users with disablements always are one step behind in terms of technological developments. This philosophy together with the results from the pilot studies has resulted in our vision of a solution: Intermodal systems. An intermodal system is our extension of multimodal systems, where the "intelligent presentation" of information is augmented with "intelligent analysis." This means that information that arrives to the system tied to a particular representation/mode is analysed into a form that lends itself to presentation and interaction in whatever modality the user chooses. This solution has been partially implemented in the form of prototypes for the core system and for information extraction from text.