Chapter 4: Application Domain

<<< Previous: Chapter 3   |   Next: Chapter 5 >>>

Chapter 4: Application DomainDownload:chapter-4-v6.pdf
Pages: 1 2 3
Page: 1Top Next
Section II Framework Componen ts Chapter 4 1 9 Ap pl icat io n D omain CHAPTER 4 A PPLICATION D OMAIN This chapter presents the application domain concept and introduces tasks[SAE1] . Tasks are at the core of the UTSQ model and are further developed in each of the following chapters where the concepts of user value, cost, return, and risk are addressed. The domain concept provides the mechanism for establishing the full set of tasks that users perceive as valuable and related. It is from this universal set that each brand draws the tasks it will address[SAE2] . [SAE1]Be consistent [SAE2]I with how you write this capital or not. don't see a domain as universal. Isn't it a defined set with related tasks? FRAMEWORK An application domain is the total set of related tasks that could deliver value to a user. A task is a bundle of actions, sometimes interfacing with technologies, which achieve a specific goal. For example, formatting a document, cutting checks for a pay period, and configuring IPv4 routing policies for an embedded switch are all tasks. A formal definition of a task is part of the Unifying Modeling Language (UML) defining "use cases" process [Kulak01,Schneider01]. Interested readers are referred to that literature for a more formal and complete explanations. The pseudo model for testing the inclusion of a task within a domain is: Task A belongs in an application domain if the users perceive task A as having value. And, see task A more related to "bench mark"task B the most defining task in the new domain rather than related to the benchmark
Page: 2Prev Top Next
Section II Framework Componen ts Chapter 4 2 0 Ap pl icat io n D omain task X from any other domain. That is to say, the users see tasks A and B more related the tasks A and X. ECONOMETRIC MODEL In the UTSQ model, a domain is defined as a cluster of tasks that have a high correlation of value for a set of users and are logically "close" to each other. One way a poor mans cluster analysis - to define this is, given that a user u places a value on a specific "benchmark[SAE1] " task T , than a second task T t belongs to the same Domain if the probability of the user finding value in T t is greater than some threshold probability . And, the benchmark task T is closer, in user perception, that a "center" benchmark task T for any other domain. [SAE1]If this is a unique usage of benchmark, than you need a footnote to define it not just quotes. Figure 4-1: Domain Assignment for a Task
Page: 3Prev Top
Section II Framework Componen ts Chapter 4 2 1 Ap pl icat io n D omain Domain Notation T t Is in the application domain if: (4-1) Pu [v u T t 0v u T 0] And, (4-2) Pu [d u T t , T d u T t , T ] Where, U The set of all users that find the "benchmark" task T of value. Pu [condition ] The set of all tasks addressed by brand b in time period . T The "benchmark" task for the new domain. T t The task being tested for inclusion into the new domain. T The "benchmark" task from the "center" of any other domain. d u T 1 ,T 2 A"distance" measure where the user u determines how close (similar) are tasks T 1 and T2 .

<<< Previous: Chapter 3   |   Next: Chapter 5 >>>

UTSQ/Chapter4 (last edited 2006-10-11 03:06:08 by jeff)