RTCAL (Real-Time CALendaring) [] was developed as part of Sunil Sarin's thesis work at MIT under Irene Greif and is one of the first applications that showed the possibility of real-time sharing of artifacts. The application is a decision support system that allows users to collaboratively schedule meeting times such as a thesis defense. It displays both private and public data: for each time slot, it shows to each user, in a shared column, whether slot is free for all users, and in another, private column, the details of any appointment of that user at that time. It supports common scrolling of these so that as users scroll to different time slots, the displays of both the private and public appointments is updated. It does not allow users to privately scroll to private time ranges, a feature the authors suggest. It gathers and tallies votes for the users and allows a user to leave and then later join the conference. Users are given the option of ignoring meeting times committed by the group in their absence.
User commands are divided into application and conference-control commands. Application commands manipulate the calendar and only the controller can enter these commands at any one time. Conference-control commands are used to pass control and enter and leave the conference. These could be entered at any time by a participant. A special role, called the chairperson, has the authority to terminate the conference and determine who the current controller is. A special window displays conference status informations such as the topic of the meeting, who the current participants are, and who the chairperson and controller are.
How should this application be implemented? There are several approaches possible, which we shall study later. RTCAL creates a collaboration-aware replica on the workstation of each user and keeps these replicas consistent. This approach is used in almost all of the applications discussed in this section, except TeamWorkstation and ClearBoard, which use a collaboration-unaware centralized process, and CAIS, which uses a combination of the replicated and centralized approaches. Interestingly, when we see collaboration infrastructures, the opposite will be true, that is, most of them will be biased towards centralization. (What could be the reason for this?)