Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
ORIENT ACADEMIC FORUM Analysis and Applications of Personal Task Executive Model in Software Development LIU Yu1, XU Aifeng2 1. Road and municipal information center of Tianjing, China, 300050 2. Surveying and Mapping Institute of Tianjing, China, 300381 [email protected] Abstract: To deal with the problems in the management of personal development software engineering, especially the lack of standardization in personal task execution, we proposed an integrated personal task executive model based on our experiments. Two project teams were tracked and recorded based on comparison method through the personal task tracking table, which was generated using our proposed model. We further carried out continuous recording and statistical analysis of 9 software engineers with 20 tasks, the results showed that the efficiency and succuss rate were significantly improved under the guidance of the our model. Keywords: personal, task executive model, tracking table, processing management 1 Introduction Since the 80th of last century, with the appearance of high failure rate within software project caused by the scale expansion of software engineering, software experts, represented by Watts S. Humphrey in the United States, started researches on the processing management of software project. They studied regularly upon the software development procedure so as to improve the software products development steadily[1][2][3][4]. Based on his experiments, Watts S. Humphrey put forward an improvement methodology of procedure management using personal, team as well as organization levels, in according with the PSP, TSP, and CMMI theories, which was valued and further studied domestically. However, based on our knowledge of many domestic software companies, the processing management are still at low level[2]. Especially for the basic level of personal procedure management, which is based on personal experiment yet instead of systematically training and improvement. As a team member, a software developer faced a next task everyday under Work Breakdown Structure (WBS), and the accomplishment of these tasks in time and satisfactorily would directly affect the progress of the entire project, even the further implement of the project [2]. Therefore, inspired by the pattern and application of the task execution, we found its particular discipline and applied it into practical applications through the continuous tracking of 2 project team in half a year. 2 Executive Model Analysis Internal and external interference source Task acceptance Task implementation Task accomplishment Intermediate feedback Task source Final feedback Figure 1. Personal task executive model 291 ORIENT ACADEMIC FORUM ① ② 2.1 Model restriction Only aim at software development Only deal with personal task ① ② ③ ④ ⑤ 2.2 Model node Task source: could be project manager, client, or even the developer himself, generally a person who is required to finish a specified task in accordance with certain progress, quality and expense Task acceptance: indicate personal ratification of the task, and start it based on the planned date. Task implementation: put into effect of the project task formally, and in the working process Task accomplishment: present the task is done, and could put in the outcome. Internal and external interference source: from the standpoint of communication. Internal: with positive personal communication, there should be promotion effects; vice versa. External: with positive encourage and remind, there will be promotion effects. Intermediate feedback: certain intermediate communication and feedback are required in the software development process. Final feedback: stands for submit outcome to “Task source”, and get positive feedback, then task is finally done. ⑥ ⑦ 2.3 Model procedure The regular task implementation workflow is as: obtain task information from “task sources”, “task acceptance”, “task implementation”, “task accomplishment”, and feedback to “task sources”. Normally, it should also include problem feedback, reconfirmation (intermediate feedback), as well as “internal and external inference” to tasks, which is a loop. In the software project, personal tasks not only include the commencement (task acceptance), implementation (task process), accomplishment (task fulfilled), but also the inference factors and feedbacks, which is determined by the complexity of the software engineering [4]. Persons without training would has less comprehending of the task steps and further lead to the unsatisfactory of the “task sources” as well as project procedure management out of control. 3 Model Testing To verify the efficiency and rationality of our personal task executive model, we tracked and recorded the task implementation of two project teams in half a year. The tracking table is given as: time span project team 20100302 -04 04 A: Jie Gao ( ) Table1. Task tacking table loop or communicati not on or not Write research yes not programs about emergency plan content result evaluation Short of the requiremen ts with the task Need communication among at least once … ① 3.1 Tracking table design principal Key points representing the personal task executive model. “time span” recorded the dates of task commencement and accomplishment; “content” indicates the “task implementation”; “loop or not” represents that whether the task is done; “communication or not” reflects the “feedback” elements; “result” indicates whether the obtained “conclusion” has reflected the “task sources”; “evaluation” is the problem sources (internal and external inference sources). The integration of the above fields indicates a 292 ORIENT ACADEMIC FORUM whole record of a task procedure whether it is a loop or not. “project team” is the track object, which is utilized to propose solutions on account of the problems. Tasks should be comparable, statistic and easy to find out problems. Find out whether the task has been accomplished timely, if not,what the problems are, and whether positive communication has made contribution to the fulfilling of the task. Keep tracking objects unknown. Ensure the data collection be real and reliable, and the whole process be confidential to tracking object, without interference to its normal work. ② ③ 3.2 Methodology We utilized the tracking method here to testify the “communication effects”. As the principal for two project, we did not inference the first team, while provide early warning to the second group. It means the 1st group will generate results based on the personal executive without inferences; while the 2nd would be reminded or early warned when a task is due or there is any other problems. The effects of inferences manually can be discovered by comparing the above two datasets. 3.3 Tracking results Number of people of group A Number of people of group B Time span of tracking Without prejudice to group A Task accomplishment statistics of group A : Table 2. Tracking results analysis Number of tracked tasks of group A 5 Number of tracked tasks of group B March 2 to August 16,2010 4 : With prejudice to group B Recorder Times of on-time completion :5 Task accomplishment statistics of group B Times of qualitative completion :4 Number of communication interfered tasks: 5 On-time completion rate: 62.6% Qualitative completion rate:50% Communication rate:62.6% 8 12 Yu Liu Times of on-time completion :8 Times of qualitative completion :12 Number of communication interfered tasks:12 On-time completion rate: 66.7% Qualitative completion rate:100% Communication rate:100% The upper half table represents the quantitative accumulation of the record contents in the tracking table; while the bottom half mainly reflects the comparative analysis of the accumulative data in percentage. The results illustrate that: On-time completion rate: the difference between groups A and B is not apparent. It illustrates that there is also communication within group A, as the communication inference factors has also proved this. Besides, other factors influence the completion rate significantly, such as the inaccurate estimation of the planned work load. The improvement of this percentage is large relied on the reinforcement of personal PSP training and practice. Qualitative completion rate: the gap is large between A and B groups. It indicates that the completion rate of group A is tremendous low without reliable confirmation of “task sources”; while the quality of group B is guaranteed due to the active alert and communication from the “task sources”. ① ② 293 ORIENT ACADEMIC FORUM ③Communication rate: group A differs significantly from B, and the variance rate is 100%-62.6%=37.4%, which is largely because there is less positive inference from the “task sources”. It can be summarized from and that their common affection factor is the positive inference and communication from the “task sources”, which means there is great affection from communication to the qualitative accomplishment of the tasks. Positive communication would improve the success rate of the project, while the rate would be greatly reduced without positive communications. Besides, we can see from Table 2 that personal executive model structure is suitable for the personal task process management and quality tracking requirement. Field data can also be acquired so as to improve the personal and team process management, or even can be utilized as reference data for the evaluation of personal process management by the organization level. 4 ② ③ Applications in Procedure Management 4.1 Personal procedure management training Before the setting up of the team, the project manager must train the team member first, where the personal task executive model also should be included. Since the deficiency of execution would usually happen in the task implementation when a developer is not trained with the related knowledge. After received the requirement specification, it is risky not to communicate with relevant persons and reconfirm in the software development process. It is also ridiculous not to hand over the outcome to relevant person when the task is accomplished as well as alter the plan discretionarily depending on personal assumption. This kind of software developing problems caused by personal issues can be ideally solved using training and reinforcement. The training testing on group B has validated our point. 4.2 Team management A skilled project manager should monitor the personal task development processing in the team development management. Besides, he also is required to find the problems within the development and solve them in time, where the communication is extraordinarily important. Table 1 is employed to record the development stage of team member in order to manage each member’s personal task efficiently. It is necessary to communicate timely in the problem finding procedure. As the principal manager of the project, we found that the positive affection and management of B group has validated our proposed method. 4.3 Organization management In fact, the tracking data under this model is also beneficial for the organizational level, as the tracking record referred to every single developer. First of all, comprehensive evaluation can be carried out on the development quality of each member via quality management department. And Quantitative assessment can also be implemented according to the technique ability and contribution via performance evaluation department. And finally, the psychological characteristics of each developer can be discovered via human resource department so as to prevent personnel crisis and evaluate the quality of talent. 5 Conclusion In this paper, a personal executive model is proposed by summarizing the disciplines within the personal executive process of software engineering comprehensively, and further tested in the practical project using local data through recording and tracking. Without large number of samples in our experiment, we only tracked 9 software engineers with 20 tasks in 2 team systematically. However, as a supplementary method for the improvement of personal task process management, our approach is easier to use and much more efficient. If the model knowledge is integrated through the process management tools, the efficiency and effect 294 ORIENT ACADEMIC FORUM will be improved significantly. References [1]. Watts S. Humphrey. A Discipline for Software Engineering, Addison-Wesley Professional, 1995:1 19 [2]. Watts S. Humphrey. Introduction to the Personal Software Process, Addison-Wesley Professional,1996 :1 6 [3]. Watts S. Humphrey. TSP-Leading a Development Team, Addison-Wesley Professional, 2005:1 2,34 [4]. Zhi Jin, Lin Liu, and Ying Jin. Software requirement project: principals and methods, Beijing: Science Press, 2008: 2~6. (in Chinese) ~ ~ ~ 295