Vol.1 No.2 2008

29/85

Research paper : Realization of a collaboration system for everyone to develop and manage (K. Eto et al.)−98 Synthesiology - English edition Vol.1 No.2 (2008) and the average was 7.7 people. The majority were MLs with 10 or less people, and this matched the design intent of a ML to be used by a small number of people. Wiki was used in about 70 % of the MLs, and many users selected qwikWeb since Wiki can be also used. While the average active period of MLs using Wiki (2,235 MLs) was 145.1 days, it was 42.4 days for MLs without Wiki (875 MLs). It can be seen that Wiki was employed for long-term use. The number of closed MLs is currently about 10 %, and this is low compared to 50 % closed ML for QuickML, and a longer use is expected to promote more accumulation of knowledge.The cumulative values of users are shown in Figure 4. Number of users, number of Wiki users, and number of users who created MLs all increased. About 70 % were MLs with Wiki, and about 40 % were users of Wiki. This is probably because many users started using Wiki by induction through the ML. About 10 % of users created MLs (1,909 people). Maximum number of MLs created by one person was 40 MLs, and the average was 1.6. As seen from the fact that there was a single user who created 40 MLs alone, the cost to the manager is relatively low. About 25 % or 505 people created 2 or more MLs. On the other hand, looking at the number of MLs in which the users participated, of the 18,519 people, the maximum number of participation was in 48 MLs, and the average was 1.3. About 20 % or 3,515 people belonged to 2 or more MLs. Participating in several MLs may be due to that fact that the users considered qwikWeb to be useful to some degree. The difference between Wiki and ML was studied from the difference in use by time zone. Figure 5 shows the number of submissions, number of Wiki accesses, and number of edits to Wiki by time. It can be seen that peaks of submission to a ML occurred during 9:00~11:00, 12:00~13:00, 17:00~19:00, and 23:00~24:00. The time zones corresponded to start of work, after lunch break, end of work, and before retiring for the evening. In contrast, the number of Wiki accesses and edits increased from start of work in the afternoon to end of work, and reached a peak at 17:00~18:00. The reason for this is, since submission to a ML is communication, it was done most often at start and end of segments in daily routine, while access to and editing Wiki are part work so it was done during work in the afternoon, and as summary at end of work and before retiring for the evening.Next, we looked at the use by groups that use both Wiki and ML. Here, we focused on MLs with over 50 submissions, 50 times or more editing of Wiki page, and durations of 1 year or more. Of the total 3,110 MLs, 64 MLs met this condition. Comparing the number of submissions and number of edits by month, about 80 % or 54 MLs had more months with higher number of submissions than edits. Figure 6 shows the cumulative values of number of submissions and edits in some MLs. Figure 6(a) is mainly ML and Figure 6(b) is mainly Wiki. Overall, the first few months were most active (with many submissions and edits) and the activities tended to decrease afterwards. Particularly Wiki showed that trend, and it tended to subside after initial heavy editing. In some MLs, the number of edits to Wiki suddenly increased. Figure 6(c1) and 6(c2) show steadily growing number of submissions, while the number of Wiki edits increased suddenly at some point and then subsided. It is thought that both ML and Wiki are regularly used if they are used for communication in daily work and for collaborative knowledge structuring, but when they are used for a particular event only, use increased suddenly before and after the event.The use of qwikWeb and the difference in how ML and Wiki were used was discussed based on operation data. Although this was a research operation, the efficacy of qwikWeb design and implementation was demonstrated.6.2 Example of qwikWeb in practiceArticles on Wiki appeared over 12 issues of Software Design of Gijutsu Hyoronsha, from August 2005 to July 2006. One article of the series was “Complete Guide to qwikWeb”[16]. At the same time, qwikWeb was used as a collaboration system to create the article series, and the actual process of Fig. 5 Difference in use status of ML and Wiki by time zone.The solid line of the graph shows the percentage of submissions by time, when the number of submissions throughout the day is set as 1.0. The same is applied to numbers of Wiki access and Wiki edit.Fig. 4 Time transition of number of users.Feb-04Apr-04Jun-04Aug-04Oct-04Dec-04Feb-05Apr-05Jun-05Aug-05Oct-05Dec-05Feb-06Apr-06Jun-06Aug-06Oct-06Dec-06Feb-07Apr-07Jun-0720004000600080001000012000140001600018000200000Number of usersCreated MLsWiki use0.010.020.030.040.050.060.070.080123456789101112131415161718192021222324Number of submissionsNumber of Wiki editsNumber of Wiki accesses(26)−

※このページを正しく表示するにはFlashPlayer9以上が必要です