|
|
FHTW Berlin |
Use-cases are a szenario-based technique for requirements elicitation. They were first proposed by Ivar Jacobsen as part of his Objectory methodology. There are many different variations on use cases, but the basic idea is to record the situation in which something happens, explain what should happen, and describe any exceptional circumstances which might arise.
In a scenario you describe a concrete use of the proposed system, using the names of a proposed user. In the use case we abstract away from the concrete situations in the scenarios and try and describe the general case. We have user roles and system functionality, which are depicted in diagrams as little stick figures and circles, respectively:
| An actor is something or someone that interacts with the target system to produce an observable result. There are active and passive actors. They are usually drawn with a stick figure, although I sometimes enjoy using a stick-woman figure just to be ornery. | ![]() |
A use case is an instance of how an actor would use a software system or a service offered by the system. More than one actor can participate in a use case. We draw use case diagrams to visualize the use cases and write out the use cases using templates, such as the one given in Wikipedia. Use case diagrams are decorated with the name given to the use case.
A scenario will typically describe the sequence of concrete activities that happen when a user is interacting with the proposed system. It is a detailed, step-by-step description of observable behaviors - that is, it does not describe the internal workings of the machine, but what a user can observe from the outside. We are not worried about how it will be done on the inside, but instead just want to see if the use of the system makes sense.
We can branch inside a scenario using an if statement and can repeat steps using a while statement. A scenario will not have exceptional conditions. If it is describing one such exceptional interaction, then it explains each step of what happens. A scenario also does not extend another scenario - it just describes what happens in this concrete instance.
We give names to scenarios, so that we can talk about them. The name should somehow be connected to the use case it is describing. Then we describe any precondition that must hold before the scenario can commence. For example: The user has a valid EC card. Then the sequence of steps is listed, followed by a postcondition describing what must hold at the end of the scenario. For example, the account linked to the EC card has been debited by 100 Euros and the user now has 100 Euros cash in her hands and the card is now marked as having been used on the current day.
Name |
ATM Cash Withdrawal |
| Precondition | Emmy has an EC card with a readable magnetic stripe & the EC card is valid & the EC card is linked to an account with this bank |
| Sequence of steps |
|
| Postcondition | Emmy is in possession of her card and 150 Euros & |
There are also scenarios sometimes referred to as "secondary scenarios" which describe alternate paths. They have their own name, the name of the scenario to which they are referring, and a description of the problem or alternative faced.
| Name | Invalid PIN |
| Scenario based on | ATM Cash Withdrawal |
| Exception | Repeat steps 7 and 8 twice if the user enters in an invalid PIN. If the third time still does not give the correct PIN, then retain the card. |
A number of scenarios are usually collected and abstrated into a Use Case and expressed graphically by a Use Case Diagram. The elements of a Use Case often include:
You should realize that every author and every company have their own ways of describing Use Cases, but usually each of these aspects are addressed.We will be using the structure given here.
You may have heard that many countries are spending a lot of money purchasing election machines. Many of the machines have very serious software engineering errors in them. Ed Felten's blog "Freedom to Tinker" has an article in which he states that his sophmore Software Engineering students can do better, so I think you can design a system better than any on the market, too.
There is one problem: in Germany we have something called the "Wahlgrundsätze".
Die Grundsätze für die Wahl zum Deutschen Bundestag, zu den Landtagen und zu den Gemeindevertretungen sind im Grundgesetz in Art. 38 und Art. 28 festgelegt. Die Wahlen sind:
Darüber hinaus muss eine Wahl nachträglich überprüfbar bzw. nachzählbar sind. |
Your job over the course of the semester will be to design a system for holding elections according to the German rules for elections. We will not be implementing it, just designing it. Your first task is to form a team. There will be five groups of three and four group of two (one group of three uses the table and the teacher's computer, the rest sit together in the lab the way the computers are set up). You may choose a team name if you wish. Use one of the UML tools we have available in the lab for drawing the use case diagram: Innovator 9.1, Together 2007, Visual Paradgm, ObjectiF. I personally like ObjectiF, but you can use one of these - or any you find on the net that you can install on your laptops.
There is a forum for submitting exercises in Moodle - please have your exercises online here by midnight the day before the next exercise, that is, usually 2 weeks later. Please include all materials - scenarios, use cases and use case diagrams - and don't forget to put the names of each person involved on the report. A report should include information on the class, the lab, the date, and all explainations should be in complete sentences, either in German or in English. Include a short (2-3 sentence) reflection on how difficult you found this exercise and how long you spent in and out of class. All material in this exercise should reflect your own work - submitting work of others is not acceptable. If your materials are not in on time, the exercise will be marked failed, 0 points.