27.03.2014 Views

SEKE 2012 Proceedings - Knowledge Systems Institute

SEKE 2012 Proceedings - Knowledge Systems Institute

SEKE 2012 Proceedings - Knowledge Systems Institute

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

With the GDUC approach, we can model functional and<br />

nonfunctional requirements by extending the use cases while<br />

analyzing the requirements as shown in Figure 2. A goal is<br />

associated to the original use case (U1), and U1 is extended to<br />

more detail use cases having their respective goals to achieve.<br />

These goals carry richer information that will be discussed in<br />

the next section. A goal can be classified with three facets:<br />

competency, view, and content. Looking into competency, we<br />

can model a goal as a rigid goal or a soft goal. A rigid goal is<br />

the goal that needs to be fulfilled fully while the requirement of<br />

a soft goal is more relaxed. The view of requirements can be<br />

actor-specific or system-specific, and the content can be<br />

functional or nonfunctional requirements. The content of a goal<br />

is derived based on functional or nonfunctional<br />

requirements.There are four steps used repeatedly in this<br />

process: identifying the actors, setting goals, constructing a use<br />

case model, and evaluating the goals.<br />

negotiation: a boss, a customer, and a salesperson. The boss<br />

expects to gain the largest benefit from a transaction and has<br />

the control over the pricing activity. He gives the price<br />

guideline to the salesperson while interacts with the customer<br />

for special needs. The customer has a certain habit for purchase<br />

and is looking for a lower price for merchandise. For ordinary<br />

purchase, he interacts with the salesperson, but for some<br />

special requests, he works out with the boss. The salesperson<br />

needs to meet the assigned sales quota by using his ability to<br />

persuade customers. He has some freedom in lowering price,<br />

but the boss has the final say. In a usual setting like in Figure 3,<br />

the customer first encounters with a salesperson, and the<br />

salesperson promotes some merchandise to the customer.<br />

When there is a special purchase request such as large volume<br />

of purchase, the customer may be directly dealing with the boss.<br />

Figure 3. Relationship between participants in selling items<br />

Figure 2. Relationship between goals and use cases<br />

After defining the categories and the relationship among<br />

goals, we now focus on identifying what issues are associated<br />

to the goals [7]. First, we can organize the existing information<br />

among goals, including goal categories, roles, cooperative<br />

goals, conflicting goals, and counterbalanced goals. Goal<br />

categories contain issues that will affect the result of analysis,<br />

and the actors and use cases provide attributes that describe the<br />

actors and their relationships. With such attributes discovered,<br />

we can analyze their relationship to see if they are cooperative,<br />

conflicting, or counterbalancing to each other.<br />

Identifying possible issues based on these attributes, we<br />

need to select which issues are best fit for the negotiation<br />

process. Comparing all issues at once causes heavy<br />

computation, so we use a hierarchical approach in evaluating<br />

issues. Issues are compared in pairs and the resulting scores are<br />

recorded in a matrix. The first step is to define the levels, and<br />

then the matrix with comparison results is constructed.<br />

Proposals containing different set of issues are derived, and the<br />

proposals are compared. Finally, the order of favorable<br />

proposals is determined.<br />

III.<br />

ILLUSTRATION AND EXPERIMENT<br />

A. Scenario and Assumption<br />

For presenting the detail steps of our approach, a scenario<br />

about a market activity is presented below to illustrate how our<br />

method can be used. There are three players in this purchasing<br />

B. Gaol-Driven Analysis<br />

We start from identifying actors as Boss, Customer, and<br />

Salesperson with their corresponding use cases as shown in<br />

Figure 4. For these actors, we define the corresponding goals,<br />

such as boss to get the highest benefit, customer to purchase<br />

necessary items, and salesperson to sell items. Using three<br />

facets (competency, view, and content) to analyze these goals,<br />

we observe that these goals are rigid since they must be<br />

fulfilled. They are also actor-specific and functional. In<br />

enriching the user model, we extend the original use case with<br />

more use cases to provide more information that is embedded<br />

in the requirements. These extended use cases are able to<br />

associate with goals that may be rigid or soft goals with actoror<br />

system-specific and functional or nonfunctional<br />

requirements.<br />

Figure 4. Use case model with associated goals<br />

Identifying the relationship between the use cases and<br />

corresponding goals based on our scenario, we can extend it to<br />

“Whole sales” and “Regular sales” from the original use case,<br />

“Selling Merchandise”. The use case “Purchase Necessity” is<br />

760

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!