|
With the advent of the World Wide Web, the use of hypertext as a structuring mechanism for richly interdependent information has become ubiquitous. With its rich mechanisms for distribution, the web also allows authors to collaboratively build work products by weaving together smaller sections into an interlinked whole. Unfortunately, relatively little is known about the process of such "collaborative hypertext construction", and current authoring support consists of little more than traditional configuration control mechanisms. Without an understanding of the appropriate way to collaboratively construct hypertext documents and without appropriate tool support, such jointly authored work products can suffer from quality problems ranging from inconsistency to incompleteness to even incoherency.
The goal of AEN research is to develop a better understanding of the
nature of collaborative hypertext authoring. AEN research is intended
to provide insight into the process and tool support required for the
development of high quality, jointly-authored hypertext documents.
AEN research is based upon the notion of a continuum of
collaboration, ranging from "weak" collaboration to "strong"
collaboration. In strong collaboration, each participant contributes to
the construction of every component of the document, and gains new
knowledge as a direct result of this construction. The constructed
document is not simply a patchwork of individual contributions, but
instead, an incremental, emergent synthesis that reflects the knowledge
created by the group as a whole. Strong collaboration is the antithesis
of a 'divide and conquer' approach, whereby the work product is simply
sliced up into components, each of which is then developed
independently and exclusively by a single author. We call this divide
and conquer approach "weak" collaboration.
The AEN system provides tools designed to support a strongly
collaborative approach to hypertext document construction. These tools
include an access control mechanism, a feedback tool, a dynamic table
of contents, a tool for representing virtual presence within the
document, and an on-line chat facility.
To evaluate AEN, we performed a case study involving 15 students
divided into three software teams who developed a software requirements
specification hypertext document over the course of a semester. We
hypothesized that use of AEN would lead to the presence of strongly
collaborative practice among some of the groups.
During the case study, the three groups spent over 180 hours using
AEN. They created 100 document nodes, 111 comment nodes, and 5 figure
nodes. You can see an HTML version of the AEN database as it was on
12/13/94, although this representation loses much of the richness of
the AEN environment.
To evaluate our hypothesis concerning strong collaboration, we devised five metrics for collaboration:
- Readers per node (RPN) Measures the average number of readers per node for a group.
- Editors per node (EPN) Measures the average number of editors per node for a group.
- Feedback node creation (FNC) Measures the percentage of document nodes containing a link to a comment node.
- Non-default
access control (NAC) Measures the percentage of nodes whose access
control has been changed from the default (which is no access for other
members.)
- Evolving Access Control (EAC) Measures the
percentage of document nodes that have had their access control changed
more than once.
High values for all of these metrics indicate the presence of strong
collaboration, while uniformly low values indicate a divide and conquer
approach, or weak collaboration. Our results indicate that strong
collaboration was present to some extent in all the groups.
|
