Our first example is quite extreme. Suppose management decides to begin a project to develop a software and there is no potential market for this software. The primary decision to start the project is a bad decision. Even all lower level decisions and all other work is perfect, the project cannot be successful. The result of the project is useless. Theoretically the potential cost of this bad decision is infinite. In any real situation it is limited by project's resources. In our situation the potential cost of the bad decision is the upper limit of time and money the company can waste to develop this useless software. Of course we may add the money spent to marketing in empty market, the indirect losses and the cost of support too. These are results of other bad decisions. In our model we'll consider only loses within the budget of a software development project.
Actual cost of the bad decision about useless project is as rule lower than potential cost. Firstly management may save money by applying new progressive support tools and development methods. The staff experience is one other thing that may save budget. It is a positive effect of the project. Of course hardware and software that the company have bought for this project may be used for more reasonable purposes too.
As second example consider a decision to use a programming language that isn't appropriate for the design of the software product. The potential cost of this decision are all resources the company can spent for coding and testing. This extrapolation has the following reason. After the project reaches some definite size the problems with the language will rise the complexity very quickly and the development won't be done further. Of course the project's staff will adapt the language to its tasks. Many support tools and crafty tricks may be developed. The resources spent for them are the cost of the crooked way. This leads to the stabilisation of the bad decision about programming language. If a next task is to add some little function, it is not reasonable to change the language and write all the source code anew. Bigger tasks can be unreachable with the old programming language. If the cost of the crooked way exceeds the resources of the project, the project usually falls. Only the correction of the bad decision can help. Of course the situation is fatal, if there are no resources to do things in the right manner and the crooked way cannot reach the project's goals too.
I hope you have felt, what the bad decision is and I can close this theme.10 Didn't you forgot that we'll speak about Software Dragons? Now you are ready to accept the following definition.