There are almost as many definitions out there as books available about Requirements Engineering. But do these definitions really help in our daily work?
A couple of weeks ago I had the pleasure to participate in a workshop about requirements methodology. At the beginning we did a small exercise. We got the following text frament – taken from a real-life example (!) - and we were asked to count the requirements regardless whether they are “good” or “bad”:
Sounds like a simple question. But: although most of the participants had several years of experience working with requirements we got an interesting result (the constructor wrote down the answers of the individual participants):
Why that?
Of course, each participant has its own experiences and knowledge about the system to be build which highly influences the way they read the document. So, you cannot expect to get a single answer…
Then we started to analyze the given text sentence by sentence trying to formulate requirements with better quality based on two of the most important characteristics for a “good” requirement:
- You must be able to implement the requirement
- You must be able to verify that the implementation fulfils the requirement
Using these quite simple rules we came to the following improvement (which still may not be perfect):
From this example you can see one of the major drawbacks when using word processing tools for requirements: Authors tend to provide justifications for certain requirements in subordinate clauses (e.g. 3 is the justification for 2). Then it may happen that they introduce redundancy (e.g. requirements 2 and 5 in the example above) which makes it difficult to maintain these documents. Apart from the implicit relations the author has “created” using subordinate clauses there are of course more relations between the different requirements we identified above. If you try to represent these relations in a more structured way you may get the following [please do not use a word processing tool for this… = ) ]
So, finally we identified requirements from four different abstraction levels within the first three sentences of the original document!
Think about it…
No comments:
Post a Comment