HUMAN-COMPUTER INTERACTION
SECOND EDITION
HUMAN-COMPUTER INTERACTION
SECOND EDITION
We have discussed the necessity of rapid feedback to the user who performs an action, and we have seen how replication or partial replication can solve this. However, we must also worry about feedthrough, the reflection of one user's actions on the other users' screens. The requirements for feedthrough are not quite so stringent as for feedback, and this can be used to reduce the amount of network traffic. There is little difference in this respect between client--server and replicated architectures, so, for the sake of argument, we will assume a client--server architecture.
Imagine a user has just typed a character. The character appears on the user's screen, either through local feedback or after an exchange with the server. However, all the other clients need to be informed also. That is, with n participants, each user action causes a minimum of n -- 1 network messages. If this is repeated for each
The choice of this chunk size can be crucial to the success of the system. If the participants can talk to each other (either co-located or with audio/video channels) they will refer to the contents of their screens. In such cases, lags of more than a few seconds can be disastrous. However, even if the computer is the only communications medium, the chunk size has an effect as the gradual appearance of text is an indication of other people's activity.
More fundamentally, the functionality of toolkit widgets may be insufficient for groupware. This is particularly obvious for text areas. The toolkit often takes over a lot of the tedium of handling an editable text region: the user can type and delete, do cursor movement and even cut and paste, all without the application's intervention. However, the groupware developer may want to have multiple cursors, or to ensure that all the participants can see the same portion of a document. Unfortunately, even information such as what portion of the document is in view, or where a particular logical character is displayed on the screen is difficult to come by, as is control over a scrollbar if this is to be shared. One is often forced to design the application round the limited capabilities of the text widget, or to use bitmap operations to paint the text oneself.
Meeting and decision support systems are aimed at helping users to generate and record new ideas and reach decisions. Meeting rooms use large shared screens to support synchronous co-located collaboration. Participants' own terminals are often WYSIWIS and they may use group pointers to support deictic reference. When the participants are remote, various forms of work surface can be used instead. Finally, for asynchronous working, systems like gIBIS help designers to record their decisions and why they came by them.
A similar problem can occur in a video conference. Imagine Eustace and Bud have monitors with cameras mounted above, so that their offices are connected. The zoom on each camera is adjustable and Bud's camera is set with a wide focus, whilst Eustace's is set with a high level of zoom. So, if Bud and Eustace are the same distance from their cameras and monitors, then Bud sees Eustace's whole face filling the screen, whereas Eustace sees Bud sat on his chair in the middle of his office. Eustace moves closer to the monitor to see Bud more clearly, while Bud pushes his chair back to get away from Eustace's 2 foot (60 cm) high face - touché. Of course, the problem gets worse if the cameras are positioned in different places relative to the monitors, or if the monitors are different sizes. Ideally, Bud ought to be able to adjust the zoom on Eustace's camera and vice versa. In fact there is some evidence that the 'glass wall' afforded by the video screen makes the precise distance less important, which could have a positive effect during cross-cultural meetings. However, gross distortions, as described above, need to be avoided.
We have seen several groupware systems which attempt to compensate for these losses. In Section 13.4.2, we discussed the idea of a group pointer, a mouse-controlled icon which can be used to point to things on a shared screen. Somewhat more esoteric, but more immediate, are the shared work surfaces (Section 13.4.3) which mix an image of the participants' hands with the electronic screen. The participants can then simply point at the relevant item on the screen, as they would face to face.
Of course, the group pointer, although used in remote groupware, is also used in synchronous co-located groupware such as meeting rooms. That is, even though the participants can converse face to face, they still need deictic aids. One reason for this is that their electronic screens, although logically shared (they can all see the same thing), are not physically shared (they are in different places). So, if Jemima points to her own screen, her colleagues do not know what she is pointing at.
Even when the participants are in the same room, the existence of electronic equipment can interfere with the body language used in normal face-to-face communication. The fact that attention is focused on keyboard and screen can reduce the opportunities for eye contact. Also large monitors may block participants' views of one another's bodies, reducing their ability to interpret gestures and body position. Most computer-supported meeting rooms recess monitors into the desks to reduce these problems.
When collecting data from groupware systems, it is also very important to have synchronized records of the participants' conversation (whether audio, video or text based) and their electronic workspaces. We need to know what the participants can see on their screens in order to interpret their remarks to one another. In the case where participants may have different views at the same time, we are likely to see breakdowns in the conversation, where one participant makes an utterance depending on his screen, whereas his colleague sees something different on her screen.
processed in 0.005 seconds
|
| |
|
HCI Book 3rd Edition || old HCI 2e home page || search
|
|
| feedback to feedback@hcibook.com | hosted by hiraeth mixed media |
|
| |