![[spacer]](/schemes/blue/line.gif){kind=small}
|
Edit nilo.hw13 here. Assignment #13 Nilo Tsung 1. what do you consider the main message of the paper? The authors skillfully overcome the conceptual gap between situation model and system model by delivering task relevant component in the right time to the software developer. It is a very interesting example of human-computer collaboration. 2. how is this work applicable to your own 2.1. work? I am afraid that it may not be very helpful to my current dissertation research. 2.2. interests? My research interest is on construction project control. I define the conceptual framework of the next generation project control systems consisting of six aspects: (1) data collection via ubiquitous computing devices, (2) data analysis of project control ¡V a 4D visualization approach, (3) problem diagnosis based on a multi-agent system, (4) domain knowledge discovery, (5) corrective action critiquing, (6) corrective action predicting, (7) project outcome forecasting. At this moment, I have some basic ideas how to attack these problems except aspects (4) and (7). I always feel aspect (4) is the most difficult one. However, I think I can apply the some of the ideas in this article, such as ¡§latent semantic analysis¡¨ and ¡§interface agent¡¨, to understand how project managers analyze the project control data. 3. do you have ideas how the work in the paper could or should be further developed? I do not know. 4. have you used any reuse libraries Yes. 4.1. if not: why not? 4.2. if yes: which ones? did the libraries have any particular strengths and weaknesses? In the good old day, when I first studied numerical analysis, we used lots of library calls in FORTRAN. In 1994, I used some of the C++ templates in the UNIX environment when I took ¡§Data Structures, Advanced Programming, and Algorithms¡¨ class from Prof. John Canny at the UC Berkley. He made us to design and implement our class projects by using many C++ templates, but the experience was pretty bad. It is because there were very few documents or examples supporting those templates. We learned by exploring them and asked around our classmates. 5. How does Codebroker infer the "task-at-hand"? When the information such as a doc comment or a signature definition is entered by the software developer, the CodeBroker extracts the information as a query, and matches the query with components in the repository, and shows the relevant components in the RCIdisplay for the software developer. 6. what are the trade-offs between 6.1. to let the systems infer the task-at-hand versus that the users specifies the task at hand? Some of the information belongs to (L4-L3) or ¡§unanticipated information area¡¨, which means it is part of the information space but the software developer is not aware of it. It will be very difficult for the software developer to come up with a meaningful query to find components belonging to the unanticipated information area. If the system is not complex and the software developer has been using the system for a long time. All the components in the system are either well understood or vaguely known. Obviously, it is not need to have this so-called ¡§system inferring for task-in-hand¡¨ paradigm. Software developers can modify her queries by a trial-and-error way to find those vaguely known components. |
