Single-Media Session Management

К оглавлению
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 
17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 
34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 
51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 
68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 
85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 
102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 
119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 
136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 
153 154 155 156 

The research on session management ranges from empirical work studies (e.g.,

Kristoffersen & Ljungberg, 1999) to the implementation of new session

management mechanisms (e.g., Patterson et al., 1990). The majority of the

development-oriented research on the topic aims at reducing, and ultimately

removing, the work required by users to handle sessions as such. Edwards

(1994), for instance, notes that many collaborative systems are applicationcentered

rather than environment-centered [See Kristoffersen (1998) for a

review of session management in collaborative systems research]. This fact

typically leads to non-standardized and inflexible session management causing

a significant amount of overhead work for the user. While this overhead work

can pay off in situations where there is a certain degree of formality, Edwards

argues that it can be detrimental to more informal, spontaneous, or lightweight

collaboration. Addressing this problem, Edwards presents a session management

model, referred to as implicit session management, for convenient

establishment of new sessions in serendipitous collaboration. Using so-called

activity information, a session management service is designed to automatically

detect potential collaborative situations and take appropriate action.

As illustrated above, session management is typically viewed as a question of

automation. According to the empirical studies on the topic, however, this is not

necessarily applicable in practice, simply because there are so many situations

where users do not want to join together with another person accessing the

same document or entering the same place. Collaborative systems such as

Montage (Tang & Rua, 1994) and the Interaction Manager (Ljungberg, 1999)

provide session management mechanisms that seek to reproduce how people

establish interaction in the real world. In the case of Montage, for instance, the

user can set a door icon to indicate the desirable accessibility (Tang & Rua,

1994). If the door is shut, then the user is busy, and so on. The idea is simply

that the user, when busy, could close the door icon, thereby reducing the risk

of being interrupted by less important sessions. The door icon is, of course, a

simplification that may not entirely be based on the valid assumptions of how

people set up collaboration in real life (Kristoffersen & Ljungberg, 1999).

Nevertheless, it reflects the ambition to provide session management mechanisms

controlled by the users, as opposed to the systems.