Information Technology Reference
In-Depth Information
Fig. 5.8
A printout including
annotations of two users
Of each set of overlapping annotations, only one annotation is printed at its correct
context position, while all other annotations are moved to the margins. Their context
position is marked by a connecting line (see Fig. 5.8).
An important issue of views that integrate annotations from all users is their scal-
ability to a large number of annotations and users. An evaluation with annotations
made by students in real lectures (which is presented in more detail in Section 5.4.1)
supports the assumption that in a lecture scenario, the views scale well to a larger
user community. The reason is that even in very large audiences, the average number
of shared annotations per document page remains rather small. This will be demon-
strated by the following example calculation: Scalability clearly does not depend on
one's own annotations. Similarly, annotations which are made by members of the
user's group do not have an influence, since the average number of members in a
group is not affected by the size of the entire audience. Hence, scalability only de-
pends on public annotations. Our experiences show that in a lecture setting, only a
very small fraction of annotations is published to the whole community. In our case
this were 1.6 % of all annotations. In our lecture evaluation, each student made an
average of 0.59 annotations per slide. Assuming four members in a learning group,
an average of 1.7 additional annotations are shared by these members. Public an-
notations of the entire audience average out at 0.9 annotations for 100 participants
and at 4.7 annotations for 500 participants. Hence, the total number of annotations
keeps quite small.
There are however some slides which are heavily annotated. In our evaluation,
the most frequently annotated slide contains an average of 5.2 annotations per user.
