Database Reference
In-Depth Information
3.1.2.5 Continuous Development
Only a few rules are present from the beginning. Additional rules are subsequently
created, often after a wide consensus among users. The functionality of technologi-
cal background is improved in small steps based on specific demand. A slow
response to users' needs often creates confusion. Butler et al. [
7
] emphasize the
need for sequential evolution of policies in collaborative systems, rather stating
them in advance.
3.1.3 Particular Types of Data
In this section, we focus on seeing the CBD in terms of social networks analysis.
We consider the CBD data as being one of the following types (see Fig.
3.2
):
l
User versus document (or other unit of knowledge)
l
User versus set of documents (e.g., topic/genre/etc.)
l
User versus user
All types can be grouped as multiple objects, thus forming subtypes like group of
users versus group of documents.
The first two types can be seen as so-called
two-mode networks
. These networks
consist of two disjoint sets of nodes (e.g., users and documents), where an edge can
connect only nodes from the first set with nodes from the second set (a user created
a document). A good overview can be found in [
4
].
Data considered under the third type is very often not present in the database
directly (with some exceptions like systems strongly based on friendship) but can
be quickly inferred, for example, from discussions or mutual collaboration on one
document or topic. From the view of SNA, this data can be considered as so-called
one-mode networks
.
The illustrative images (Fig.
3.2
) provide good insight into the structure of social
networks and CBD. But with an increasing number of objects, the clarity of images
decreases quickly. In many cases, the computational complexity of the methods
being used increases rapidly. There are several possible ways to address this issue
(illustrated in Fig.
3.3
):
Fig. 3.2 Particular types of data in CBD with respect to the social point of view
Search WWH ::
Custom Search