next up previous contents
Up: Perspective-based Reading Previous: Description of the

Perspective-based Scenarios

The efficiency of this reading technique is determined by the scenarios. A scenario describes certain activities that should be performed by the reader while reading the documents. In addition to the activities the scenario contains some questions dealing with the activity.
To be much more precise, a scenario is a collection of procedures that operationalize strategies for detecting defects.
It could be argued that scenarios are another kind of checklist. A checklist is a set of questions or statements without any hint or procedural description of how to read a document. As there is no description in checklists, scenarios have to be separated from them.
The question is how to generate ``good'' scenarios?
The model we use is an abstract description of the different roles in the software development process. The different roles together with a set of questions for each of this roles will determine the scenarios we use for perspective-based reading.
Now the interesting point is:

There are several reasons why it is difficult to generate scenarios. We have to find on the one hand a set of questions and activities that can be used for a broad range of requirements documents and that has on the other hand the detail that the reader of the requirements knows what to do.
We have used the following process in order to develop the scenarios:

The starting point was an abstract model of the software development process. Each role within this process was characterized by a set of high level activities, e.g. ``Make up the testcases'' for the tester. Then we looked at the NASA process to see how they perform the activities (if at all!). In order to ensure that we can find all the defects, we have used a defect classification scheme. The classification scheme provides some question for each defect class. The next step was that we assigned question to each activity of a role about the results of the activity or the activity itself.
Throughout the development of the scenarios we there were several questions left open:

  1. What is the level of detail which is necessary to support the reader? If the scenario is too detailed it will become much more of a checklist than it was intended. If the scenario in not detailed enough it won't provide any help.
  2. What are the really important questions that can expose classes of defects while performing the activities of each role?
  3. To what extent should the scenario be tailored to a specific environment?
At the moment we have decided that the scenarios include questions that are not too detailed for each activity. The questions don't describe how to perform the activity. Each scenario can be used in a broad context. There are only one or two questions which are more interesting in the NASA environment than in others. The scenarios we have developed and used can be found in appendix A

In order to validate the effectiveness of perspective-based reading we have conducted the controlled experiment that is the focus of this lab package.

next up previous contents
Up: Perspective-based Reading Previous: Description of the

Generated by latex2html-95.1
Mon Jun 24 13:58:35 EDT 1996

Web Accessibility