Bloggo back to the blog
…. But, Does it Work on Paper?-->
Have you ever worked on a project that failed? Just after the decision to close the project was made, someone says with the wonderful gift of hindsight that they knew it would fail. I always want to know why these people did not say anything before-hand, (usually) remaining silent hitherto. With hindsight, all things are possible; I would surely have won the National Lottery each week for the last nine years!
I once worked on a project where Brett Gonzales kept asking “Does it work on Paper?” Rather than rely on the notoriously unreliable gift of hindsight, asking this important question clarifies minds. Until the proposed solution works in theory (“on paper”), there is little point in proceeding. Strangely, those on the project got used to Brett asking the question, and even voiced it for themselves. Together we moved towards a solution that eventually did work on paper, and it was only then that true testing could begin.
System components need to work both individually and together. The first place to begin integration testing is on paper. On one project, an interface between two systems had three different views of the file that was passed between them. The file specification of the sending process was different to that of the receiving application. Furthermore, the file that was actually produced was different to both specifications. In order for the interface to have a chance of performing correctly, at least two of the three parts needed to change. Part of testing in this instance involved assessing the file that is produced against the specifications for the sending and receiving processes.
Hindsight is available to all. Testing has as part of its role to ask in advance the ‘hindsight’ questions. We need to ask hard questions of the software before implementation, so the “hindsight” moments do not occur. How do we know what the hindsight questions are? That is the hard point, and why testing is an intellectually stimulating activity. We need to anticipate the hindsight questions and ask them.