Experience of Software Process Improvement Applied to Education: The Personal Work Planning Technique

An Experience of Software Process Improvement Applied to Education: The Personal Work Planning Technique D. Antonio de Amescua Seco

IDEA GROUP PUBLISHING

An Experience of Software Process Improvement Applied to Education 1

IDEA GROUP PUBLISHING 1331 E. Chocolate Avenue, Hershey PA 17033-1117, USA Tel: 717/533-8845; Fax 717/533-8661; URL-http://www.idea-group.com

16#$ $

g n i h s i l b u P p u o r G a e An Experience of Software Process d I t h g i r Copy Improvement Applied to Education: The Personal Work Planning hing s Technique Publi up o r G a e d I t h g i r y EXECUTIVE Cop SUMMARY g n i h s i l b u P p u o r G BACKGROUND a e d I t h g i r y Cop g n i h lis b u P p u o r G a e d I t h g i r y p Co D. Antonio de Amescua Seco, Javier García Guzmán, María Isabel Sánchez Segura, Paloma Martínez Fernández and Juan Llorens Morillo Carlos III University of Madrid, Spain

This case describes the use of the Personal Work Planning (PWP) technique as a time management tool for student practicals on the Software Engineering II course at Carlos III University. It analyzes the results obtained and presents the methodology used to implement activities associated with the Personal Work Planning technique in an academic institution. In addition, an empirical study was carried out to determine the level of student satisfaction after using this technique. This case study concludes that many students realized the usefulness of PWP for their assignments. This technique is of invaluable help to lecturers who wish to improve the course curriculum and the practicals.

Carlos III University of Madrid, founded in 1989, is a public university with an enrollment of 13,342 students and a staff of 1,231 (910 lecturers and 321 administrative staff). It offers a total of 32 degree courses and has three faculties: Social Sciences and Law; Humanities, Documentation and Communication; and the Engineering School. The Engineering School offers studies in Industrial Engineering, Computer Sciences and Telecommunications among others. The university teaching staff is divided into 16 departments, which organize research and lectures in their respective areas. Carlos III University opened with modern, flexible and multidisciplinary curricula and, from the beginning, made measuring and controlling its basic processes an integral part of their policy. Every year, a University Quality Committee implements Improvement Programs and carries out a Teaching Quality Evaluation Program for all degree courses. Every semester, students complete questionnaires to evaluate both courses and lecturers for the Computer Science degree. The evaluation for the Computer Science degree has been positive. At present, the questionnaires consist of 13 questions: one assesses their interest in the subject matter, eight assess the lecturer, and the remaining four assess the practicals and time dedicated to the course.

Copyright © Idea Group Publishing. Copying without written permission of Idea Group Publishing is prohibited.

2 Seco, Guzmán, Segura, Fernández & Morillo

This experiment was based on an evaluation of the Computer Science course. The aim was to determine the effort students dedicated to developing their practicals. With the real data, it will be possible to design better courses where there is a correlation between the total number of hours and the time dedicated to practicals, thus a more realistic approach to achieving the objectives.

SETTING THE STAGE

g n i h s i l b u P p u o r G

The main objective of Software Process Improvement (SPI) is to increase the quality of products and services, which a software company provides, by improving the quality of their production processes. One of the first research centers of software process improvement was the Software Engineering Institute (SEI) at Carnegie Mellon University, in particular under Watts Humphrey. The SEI developed the Capability Maturity Model (CMM) (Paulk et al., 1993) based on the definition of Humphrey’s software process improvement (SPI) principles (Humphrey, 1989). SPI has already been edited in ISO 15504 standard (ISO, 1997) as an international model reference for software process improvement. However, experience related to SPI has shown that the lack of success in an improvement program is due to the bureaucratic nature of the improvement designs and the need for an amount of human and material resources that many organizations cannot provide. Important research has conclusively shown that one of the project’s success factors is the capability of the personal software process used by software engineers. Watts Humphrey defined the main purpose of the Personal Software Process (PSP) as continuous improvement of the individual activities related to a software project developed by a software engineer. The Software Engineering curriculum must, therefore, help software professionals to do their job well; that is, to design and develop high-quality software products at agreed cost and on schedule. This will increase the quality of the products and services provided by the software companies. In the PSP description, Humphrey (1997), introduces software process principles to teach students disciplined personal practice to produce high-quality work. In order to develop quality software systems, students must learn to plan and control their work. Previous case studies (Lisack, 2000) concluded that many students failed to recognize the benefits of such a process and felt that it only took time away from programming. The authors of this experiment firmly believe that PSP should be taught with each course but different aspects would correspond to different courses. For example, individual task management would be taught during a software project management course, and software error prevention and detection during a software development course. For this reason, the techniques proposed in PSP for planning individual work have been extracted and modified. This is how the definition of Personal Work Planning Technique (PWP) came about. PWP was taught during a software project management course in Information Systems, a specialization of Computer Science. PWP techniques help to improve the individual process used by software engineers since previously registered personal process performance experience can be retrieved to organize and estimate the size, duration and effort of the new task to be accomplished. Consequently, software engineers can reduce the time spent on re-work resulting from poor task organization, which means that software products can be delivered on time, thus satisfying all quality levels previously determined.

a e d I t h g i r y p Co

g n i h lis b u P p u o r G a e d I t h g i r y p Co g n i h s i l b u P p u o r G a e d I t h g i r y Cop g n i h lis b u P p Personal Work Planning Technique u o r G a e d I t h g i r y p Co

In this case, plan refers to the document that describes the way a specific project is to be developed: how, when and what time and effort are needed. The main contents of the plan were: tasks to be accomplished, starting and finishing dates and the time needed to execute them. Time could be spent on different work categories. A work category is a set of tasks of the same type; for instance, attending classes, studying, doing exercises. All work categories are types of tasks.

An Experience of Software Process Improvement Applied to Education 3

In this experiment, a personal work planning technique (PSP) was selected to teach students the importance of using a disciplined method to manage the time and effort each individual spent on a software project. The following steps were proposed in PSP to help students prepare a reliable and effective plan for their personal work on a software project: • Start with an explicit statement of the work to be done and check to ensure that it is what the customer requested. • Break up the projects that require more than a few days’ work into smaller tasks and estimate each task separately. The added detail will improve precision and will most likely improve accuracy as well. • Base estimates of this work on historical data of previous work done. • Record estimates and compare them with actual results. The tasks for Personal Work Planning can be seen in greater detail in the Diagram 1. We followed most of Humphrey’s recommendations, but with some modifications in the specification of this technique. The lecturers prepared an initial planning chart for students (see Table 3). This chart was used to estimate the time needed to develop their practicals. The main contents of the chart were: • A work breakdown (WBS) of the practicals, which included all the weekly tasks for the semester. • A set of task categories students could spend their time on was suggested in order to facilitate the process of planning and summarizing the weekly time. The task categories presented were: class attendance, studying, doing exercises, practicals and other. The daily registration table and the weekly activity balance (see Table 6 and Table 4) are the same

g n i h s i l b u P p u o r G a e d I t h g i r y p o C g n i h s i l b u P up o r G a e d I t h g i r y Cop g n i h s i l b u P p u o r G a e d I t h g i r y Cop g n i h lis b u P p u o r G a e d I t h g i r y p Co

Diagram 1: PWP Technique

D e f in e th e s c o p e o f th e w ork

D e te r m in e th e n e e d e d ta s k s t o d o th e jo b

E s tim a te s iz e , e ff o r t a n d d u r a t io n

R e p la n th e n e e d e d ta s k s

C o lle c t d a ta o n ta s k s p erfo rm a n ce

P e r f o r m t h e p la n n e d ta s k s

D o a w e e k ly b a la n c e o f ta s k s p e r f o r m e d

No

L a st w ee k?

Y es

4 Seco, Guzmán, Segura, Fernández & Morillo

as those proposed in the PSP. Lecturers introduced a new step at the end of the process. They modified the personal work planning process proposed in the PSP. This was to help the students learn how to plan and estimate time for their assignments. Every week, each group doing the experiment would use the current week’s experience to refine the plan for the following week’s estimations.

g n i h s Objectives i l b u P p u o r G a e d I t h g i r y Participants Cop g n i h lis b u P p u o r G a e d I t h g i r y p Co g n i h s i l b u CASE DESCRIPTION P p u o r G a e d I t h g i r y Cop g n i h lis b u P p u o r G a e d I t h g i r y p Co

The purpose of this experiment was to find out if the Personal Work Planning technique was a valid tool for lecturers to design and plan a good undergraduate degree course. The objectives were: • To observe the level of the student satisfaction using this technique. • To observe the effect of this technique on student grades. • To determine whether the findings of this technique provided useful data to improve courses.

The experiment was done by Software Engineering II students in their third year of Information Systems, a specialization of Computer Science. The curriculum comprises basic software project management concepts and techniques. These students already have knowledge of Structured Analysis and Database Design, which are essential for the practicals. The course is offered in both day and evening classes, but we decided that the day class would be the pilot group and the evening class would be the control group. All the students in both pilot and control classes were divided into groups of three for the practicals. The pilot group was made up of 23 groups, with a total of 69 students; the control class had 18 groups, with a total of 54 students. Both classes had to do the same practicals but only the pilot groups used the Personal Work Planning technique. Each pilot group was considered a unit for the purpose of this experiment. Two lecturers per class and the coordinator of Software Engineering II supervised the classes. These same lecturers taught the practicals for both day and evening classes.

Over the last few years, Information Systems students have been complaining about the workload. Lecturers have also observed a decrease in the quality of the students’ work. The reason for this poor quality was the somewhat chaotic system used for the practicals. We decided it would be convenient to use PWP so that: • students could learn to develop their practicals from a plan that was feasible. Experience in planning leads to progressive improvement of the same. • students could develop the practical aspect of the course methodically by using the plan that they themselves had prepared. • the proposed practicals would be easier for students to handle. Thus products could be submitted on time. • the quality of the work would improve. Many mistakes were a result of doing their work too quickly or making serious mistakes in the process, as well as lack of attention to details. The lecturers at Carlos III were, and are, driven by the culture of continuous improvement in the quality of the courses. For five weeks, they defined and planned the procedure to test the benefits of PWP in the field of education. They also designed the support materials that students had to use for the experiment. (See Appendix.) The norms and the procedures defined for this experiment are described below.

An Experience of Software Process Improvement Applied to Education 5

Norms Students involved in the pilot project had to faithfully record the data that indicated the real time taken to carry out each task. Therefore, they needed to plan their assignment thoroughly.

Experiment Development

g n i h s i l b u P p u o r G a e d I t h g i r y p o C g n i h s i l b u P up o r G a e d I t h g i r y Cop g n i h s i l b u P p u o r G a e d I t h g i r y Cop g n i h lis b u P p u Results o r G a e d I t h g i r y p Co

The management practices included in PWP are also used in software project management, so we decided to introduce them in Software Engineering II. Students worked in groups of three for the practicals, which involved software project analysis, estimation, organization and planning, in addition to basic tracking tasks. At the beginning of the experiment, lecturers provided students from both day and evening classes with the specifications for developing the software project, and deliberately divided the work into small and easily planned tasks spread out over nine weeks. They prepared a table (Appendix, Table 3) called “tasks distributed by weeks”. So, to begin with, students already had a tasks guide divided into weeks to estimate the hours necessary for each weekly task. In order to verify the objectives proposed, we decided that only one class should use PWP for their practicals and, although PWP techniques were not explicitly taught, they formed part of the procedure for the practicals. In the pilot class, groups were told how to fill in the table of tasks distributed by weeks Table 3), the table of daily registration (for students’ use only) (Table 3) and the table of weekly balance (Table 6), as well as how to send this information to lecturers. These tables were prepared on an Excel spreadsheet, and distributed to the pilot groups as follows: • Students had fill in Table 3 with estimated times before starting the practicals, as they had no previous experience. This table was then sent to the lecturers by e-mail. Every week for the next nine weeks, using the knowledge acquired, each pilot group revised the initial planning in order to reassign times to the tasks which had not been completed. An updated table was forwarded to the lecturers. • The pilot groups indicated the time spent on practicals (Table 3). This information was then used to complete the table of weekly balance (Table 6). It is important to emphasize that the table of weekly balance was filled in after completing the tasks assigned to each week. Therefore, this information represented the real time spent on each task. During the nine weeks, the control groups used Table 3 to follow what had to be done each week and developed the tasks which corresponded to the current week only. At the end of the semester, the pilot groups were given a questionnaire (see Appendix, Questionnaire 1) to find out their degree of satisfaction with the PWP technique used. This questionnaire evaluated eight statements ranging from disagree (1) to completely agree (4). They were also able to choose “Don’t Know”. Every student was expected to fill in the questionnaire, although it was not obligatory. To complete the experiment, lecturers evaluated: • data collected from the table of tasks distributed by weeks (Table 3) and the table of weekly balance (Table 6), and • data collected from the satisfaction questionnaires, Questionnaire 1. Diagram 2 shows the experiment activities diagram using UML notation (Booch et al., 1999).

Results Obtained From the Tracking Sheets After completing the practicals, the data (see Table 3 and Table 6) were analyzed to identify the evolution of the 23 pilot groups in their time estimations. The final number of groups evaluated was 19. Groups 8, 12, 13 and 23 were excluded because they submitted insufficient or unmanageable data. While 17.4% of the groups did not understand the PWP techniques, 82.6% finished the experiment

6 Seco, Guzmán, Segura, Fernández & Morillo

Diagram 2: Experiment Activities Diagram Pilot Class

Lecturers

Control Class

g n i h s i l b u P p u o r G

Design and Plan the Experiment

5 weeks

Select Pilot and Control Groups

a e d I t h g i r y p Co

1 week

Form groups

Form groups

Receive specific information to fill in in tracking sheets

g n i h lis b u P p u o r G a e d I t h g i r y p Co g n i h s i l b u P p u o r G a e d I t h g i r y Cop g n i h lis b u P p u o r G a e d I t h g i r y p Co 1 week

Fill in and send table of task distributed by weeks

Receive students data

Development of current week tasks

9 weeks

Development of current week’s tasks

Fill in and send table of weekly balance.

Receive students data

Revise and send table of task distributed by weeks

Receive students data No

Last week?

Yes

5 weeks

Fill in satisfaction questionnaries

Analyse data

Compare Results

successfully. The relationship between the estimated average time (EAT) for all the groups and the real average time (RAT) invested is shown in Table 1. The average EAT was 11.25 hours and the average RAT was 10.23 hours. Data from Table 1 is shown in Graph 1. The average EAT and RAT are also included. In Graph 2, there is a serious error in estimation (not unusual with inexperienced people). They

An Experience of Software Process Improvement Applied to Education 7

Table 1: Estimated and Real Average Time for the 19 Groups Evaluated TOTAL Week1 Week2 Week3 Week4 Week5 Week6 Week7 Week8 Week9

Estimated Average Time Real Average Time 12.17 8.66 9.16 11.65 8.42 14.10 6.26 8.05 9.91 13.03 21.32 7.91 7.97 10.41 10.28 8.21 16.18 9.65

g n i h s i l b u P p u o r G a e d I t h g i r y p o C g n i h s i l b u P up o r G a e d I t h g i r y Cop g n i h s i l b u P p u o r G a e d I t h g i r y Cop g n i h lis b u P p u o r G a e d I t h g i r y p Co

Graph 1: Relationship Between Estimated and Real Average Time W eekly average of the groups

23,00 22,00 21,00 20,00 19,00 18,00 17,00

Hours dedicated

16,00 15,00

E stim ated A verage Tim e (E A T)

14,00

R eal A v erage T im e (R A T )

13,00

A verage of E A T data

12,00

A verage of R A T data

11,00 10,00 9,00 8,00 7,00 6,00 5,00 4,00 3,00 2,00 1,00

k9

W

ee

k8

W

ee

k7

W

ee

k6

ee

k5

W

ee

k4

W

ee

k3

W

ee

W

ee

W

W

ee

k1

k2

0,00

W eeks

overestimated the time needed in 44.4% of the cases. In Graph 1, we can observe that the average EAT is greater than that of RAT. Also in Graph 1, the relation between EAT and RAT does not follow a pattern. This may be because the study of METRICA2 (Metrica, 1993), a full software development methodology which was assigned in week 6 (see Table 3), was a difficult task for the students. If we eliminate the data from week 6, we get Graph 4, where a clear tendency during the nine weeks can be observed. In Graph 3, the difference between EAT and RAT less data from week 6 is shown. In 37.5% of the cases, the estimated times were greater than the real times and in 62.5% of the cases, real times were greater than estimated times. Results Obtained from the Level of Student Satisfaction Once the students concluded the practicals using PWP technique, they were given a questionnaire with eight statements to find out their degree of satisfaction (see Appendix, Questionnaire 1). The results obtained appear in Table 3. Each one of the questions was dealt with independently and the data obtained appear in Graph

8 Seco, Guzmán, Segura, Fernández & Morillo

Graph 2: Difference Between EAT and RAT 15,00

g n i h s i l b u P p u o r G

14,00 13,00 12,00 11,00 10,00 9,00

7,00 6,00 5,00

a e d I t h g i r y p Co 4,00 3,00

Difference between EAT and RAT

2,00 1,00

k9 ee

k8

W

W

ee

k7

k6

W

ee

k5

W

ee

k4

W

ee

ee

k2

k3

W

-2,00

W

ee

W

W

ee

k1

0,00

-1,00

ee

Hours dedicated

8,00

-3,00

g n i h lis b u P p u o r G a e d I t h g i r y p Co g n i h s i l b u P p u o r G a e d I t h g i r y Cop g n i h lis b u P p u o r G a e d I t h g i r y p Co -4,00 -5,00 -6,00 -7,00 -8,00

W eeks

Graph 3: Relationship Between Estimated and Real Average Time Less Data From Week 6 Weekly average of the groups without week 6

17,00 16,00 15,00 14,00 13,00

11,00

Estimated Average Time (EAT)

10,00

9,00

Real Average Time (RAT)

8,00 7,00 6,00 5,00 4,00 3,00 2,00 1,00

Weeks

W ee k9

W ee k8

W ee k7

W ee k5

W ee k4

W ee k3

W ee k2

0,00

W ee k1

Hours dedicated

12,00

An Experience of Software Process Improvement Applied to Education 9

Graph 4: Difference Between EAT and RAT Less Data From Week 6 7,00 6,00 5,00 4,00

g n i h s i l b u P p u o r G a e d I t h g i r y p o C g n i h s i l b u P up o r G a e d I t h g i r y Cop g n i h s i l b u P p u o r G a e Problems Encountered During the Experiment d I t h g i r y p o C g n i h lis b u P p u o r G a e d I t h g i r y p Co 2,00

D iffe re n c e b e tw e e n E A T a n d RAT

1,00

k9

W

ee

k8

W

ee

k7

W

ee

k5

W

ee

k4

W

ee

k3

W

ee

k2

ee

W

-1,00

ee

k1

0,00

W

Hours dedicated

3,00

-2,00 -3,00 -4,00 -5,00 -6,00 -7,00 -8,00

W e e k s (w ith o u t w e e k 6 )

Table 3: Results Obtained From the Satisfaction Questionnaires Questionnaire of satisfaction about Personal Work Planning Value 1: Disagree

Statement Statement Statement Statement Statement Statement Statement Statement

1 2 3 4 5 6 7 8

1 1 2 1 5 6 6 3

Value 4: Value 2: Value 3: Completely Indifferent Agreement Agree Don't Know 14 36 10 0 19 26 15 0 22 23 12 2 19 32 9 0 15 28 13 0 17 24 14 0 25 21 7 2 9 26 19 4

5. Most students chose Value 3 for all eight statements.

In order to carry out this experiment correctly, we had to solve several problems. They were related to the misunderstanding of some aspects of the task and the technique to be used, as well as issues of planning. First, due to the difficulty students experienced with planning the first week, an entire two-hour class had to be dedicated to dealing with it. For their practicals, students found that the time they had estimated for class attendance and the relevant exercises did not correspond to the week in question but to the following. It was decided that students should write in the time they estimated even though there was no correlation between the two, since it was time they were going to use and they had to cater for. It should also be borne in mind that this information was not included in the conclusion. In addition, under class and exercise times, students included some hours that were not directly related to Software Engineering II, but to other courses taught at the same time. This was considered acceptable because they could have used that time for the practicals. Finally, students were able to modify the initial planning of the tasks (estimated planning) while

10 Seco, Guzmán, Segura, Fernández & Morillo

Graph 5: Satisfaction Questionnaire Results Don't Know

Value 4: Completely Agree

Value 3: Agree

a e d I t h g i r y p Co

g n i h s i l b u P p u o r G Statement 8 Statement 7 Statement 6 Statement 5 Statement 4 Statement 3 Statement 2 Statement 1

Value 2: Indifferent

g n i h lis b u P p u o r G a e d I t h g i r y p o C g CURRENT CHALLENGES/PROBLEMS FACING n i h s i l ORGANIZATIONS b u P p Conclusions from the Levels of u Student Satisfaction o r G a e d I t h g i r y Cop Conclusions from the Tracking Sheets g n i h lis b u P p u o r G a e d I t h g i r y p Co Value 1: Disagree

0

2

4 6

8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38

Number of students

the project was in progress. However, this was only feasible if the changes were made before they had completed the tasks. It was thought that, based on similar tasks already completed, the students themselves would provide the information needed to re-estimate the times. It would seem that this was not done properly because tasks differed radically from week to week.

• • •

After analyzing the questionnaires in detail, we came to the following conclusions: Students were able to track and control the state of the assignments of their practicals. The efficiency of team members doing the practicals improved in many cases. The PWP technique was well accepted for use in the academic world, as well as its application in the real world.

From a detailed analysis of the tracking sheets, we can conclude that the PWP technique was well accepted among the students. However, it must be added that the students had difficulties in learning how to use this technique because the activities were completely different from week to week and they did not know how to use the experience gained from the previous activities to estimate the time to develop the ones which followed. At the beginning of the semester, the different times estimated for the activity were far greater than the real times spent on carrying them out. As the semester progressed, the difference between estimated and real time narrowed, until estimated times dropped to just below real times needed for the tasks. This may occur because, as their knowledge in a particular area increases, students tend to

An Experience of Software Process Improvement Applied to Education 11

overestimate their capacity to work and thereby make shorter time estimations for their tasks. However, this tendency was distorted in the sixth week of the semester as students had to select and customize an unknown software development method. This meant that the workload for that week became heavy and the students overestimated the time they needed. In the future, lecturers will reduce the workload for that week, allowing students to use an already tried and tested methodology. Graph 6 shows the grades awarded to the students upon completion of their practicals. The results show that the average grades of both day and evening classes were practically identical. The control class obtained an average of 7.05, while the pilot class had an average of 6.98. On the other hand, the number of students who exceeded 8.0 was greater in the pilot class. This means that the effort invested in the PWP technique was reverted. In many cases, the same products were obtained but of better quality. Lastly, it must be stated that the students delivered their work on time and with acceptable quality levels using the PWP technique. We can therefore conclude that the application of the PWP technique does not imply a substantial increase in each student’s workload. There was a general sense of satisfaction from achieving the proposed tasks and a clear understanding of the benefits of these techniques while the work was in progress. The final conclusion is that this technique is suitable for both students and lecturers. It also offers a valuable tool for analyzing what happens during the semester, and for designing and planning better courses. In order to apply successfully PWP technique, as it is described in this case, it is important that lecturers consider the following issues: • The practicals must be developed in a period of time corresponding to a semester without many difficulties. • The practicals must be decomposed in small weekly tasks. Some of these tasks must be similar in order to let students use their accumulated experience for estimating future works. Also the students must consider that the gathering of time dedicated to develop their practicals is not intended to control their work. The goal of this gathering activity is to acquire the needed experience for correctly estimating the effort for future practical tasks.

g n i h s i l b u P p u o r G a e d I t h g i r y p o C g n i h s i l b u P up o r G a e d I t h g i r y Cop g n i FUTURE TRENDS h s i l b u P p u o r G a e d I t h g i r y Cop g n i h lis b u P p u o r G a e d I t h g i r y p Co Graph 6: Distribution of Grades in Both Classes

Pilot Class

Control Class

35,0

32,9

31,5

30,0

27,4

27,4

%Frecuency

25,0

20,0

16,4

15,0

13,7

10,0

8,2

6,8

5,5

5,0

2,7

1,4

0,0

0,0 0,0

0,0 0,0

0,0 0,0

0,0 0,0

1

2

3

4

0,0

0,0

0,0 0,0

0,0

0

5

6

Grades

7

8

9

10

bigger...

12 Seco, Guzmán, Segura, Fernández & Morillo

We plan to repeat this experiment on courses in which problems with practicals have been detected. That is why we are going to present the results of this experiment to academics in charge of degree programs, such as deans and heads of departments. We are preparing training material on the Personal Work Planning technique and how to implement it for lecturers. A program will be designed to institutionalize this practice in our university if further results confirm the success of this experiment.

ACKNOWLEDGMENTS

g n i h s i l b u P p u o r G

The authors would like to thank the students of Software Engineering II (99/00) for their invaluable contribution and interest in carrying out this research.

a e d I t h g i r y p Co

FURTHER READING

Disney, A.M., Johnson, P.M. (1999).A critical analysis of PSP data quality: results from a case study. Empirical Software Engineering, 4(4). Hilburn, T.B. (1999). PSP metrics in support of software engineering education Proceedings 12th Conference on Software Engineering Education and Training (Cat. No.PR00131). IEEE Comput. Soc, Los Alamitos, CA, USA, 135-6. Humphrey, W.S. (1995). Introducing the personal software process. Software Eng. Inst., Carnegie Mellon Univ., Pittsburgh, PA, USA. Annals-of-Software-Engineering, 1, 311-25 Linda, L and Brennan, M. (2000). Learning Technology Management While Teaching Technology Management: A Trial of Distance Learning in Higher Education – Annals of Cases on Information Technology, Vol 2. Miller, J and Mingins, C (1998). Putting the practice into software engineering education. Proceedings. 1998 International Conference Software Engineering: Education and Practice (Cat. No.98EX220). IEEE Comput. Soc, Los Alamitos, CA, USA, 200-8. Mingins, G., Miller, J., Dick, M., Postema, M. (1999). How we teach software engineering. JOOP11(9), 64-6, 74 Sauer, L.D., Lindquist, T.E. and Cairney, J. (1999). Tracking personal processes in group projects. Proceedings. Twenty-Third Annual International Computer Software and Applications Conference (Cat. No.99CB37032). IEEE Comput. Soc, Los Alamitos, CA, USA, p.364-9. The Personal Software Process(PSP) has been developed by the Software Engineering Institute (SEI). Canadian site: http:/ / www.cs.usask.ca/ grads/ vsk719/ academic/ 856/ project/ node3.html Software Engineering Institute (SEI) : http://www.sei.cmu.edu/tsp/ Building High Performance Teams Using Team Software Process SM (TSP sm) and Personal Software Process SM (PSP sm) East Tennessee State University’s Personal Software Process Studio Home page. http://wwwcs.etsu.edu/psp/. Personal Software Process (PSP SM) for Engineers. Spanish site: http:/ / esi.es/ Training/ Catalog/ pspeng_desc.html . Personal Software Process. Educational site: http:/ / psp.distance.cmu.edu/

g n i h lis b u P p u o r G a e d I t h g i r y p Co g n i h s i l b u P p u o r G a e d I t h g i r y Cop g n i h lis b REFERENCES u P p u o r G a e d I t h g i r y p Co

Batini, C. (1992). Conceptual Database Design. Booch, G., Rumbaugh, J. and Jacobson, I. (1999). The Unified Modeling Language. Addison Wesley. Humphrey, W. S. (1999). Introduction to the Personal Software Process. SEI Series in Software Engineering, Addison Wesley. Humphrey, W. S. (1989). Managing the Software Process, Massachusetts: Addison Wesley Publishing Company ISO (1997). ISO/IEC Std.15504 Software Process Improvement and Capability Determination, ISO/IEC/

An Experience of Software Process Improvement Applied to Education 13

JTC 1 Lisack, S.K. (2000). The Personal Software Process in the classroom: student reactions (an experience report). Thirteenth Conference on Software Engineering Education and Training. IEEE Comput. Soc, Los Alamitos, CA, USA, 322. METRICA versión 2; Metodología de planificación y desarrollo de sistemas de información. (1993) Instituto Nacional de las Administraciones Públicas. Paulk, M.C., Garcia, S.M., Chrissis, M.B., and Bush, M. (1993a). Capability Maturity Model for Software, Version 1.1, CMU/SEI-93-TR-25. Technical Report. Software Engineering Institute. Carnegie Mellon University. United Kingdom Software Metrics Association (UKSMA). (1998). MK II Function Point Analysis. Counting Practices Manual Version 1.3.1. Yourdon, E.- (1989). Modern structured analysis. Prentice Hall.

g n i h s i l b u P p u o r G a e d I t h g i r y p o C BIOGRAPHICAL SKETCHES g n i h s i l b u P up o r G a e d I t h g i r y Cop g n i h s i l b u P p u o r G a e d I t h g i r y Cop g n i h lis b u P p u o r G a e d I t h g i r y p Co

D. Antonio de Amescua Seco. holds a Ph. D. in Computer Science and is a Full Professor in the Computer Science Department of the Carlos III University of Madrid. He worked as a researcher at the Polytechnic University of Madrid from 1983 to 1991 and from 1991 to date at Carlos III University of Madrid. Also, I have been working both in a public company (Iberia Airlines) as software engineer and in a private company (Novotec Consultores) as a software engineering consultant.His research interests include new software engineering methods, with results edited in papers and congresses. He was the research project leader for the development of the Information System Development Methodology for the Spanish Administration and participated in others project sponsored by the European Union.

Javier García Guzmán has a Degree in Computer Science Engineering. He is Assistant Lecturer at the Carlos III University of Madrid and has 6 years of experience as a software engineer and software engineer consultant, especially in software projects and processes management and software engineering methods. He was a researcher in the project for the development of the Information System Development and Maintenance Methodology for the Spanish Administration. He also has participated in agreements between Carlos III University and Spanish companies for developing information systems and for evaluating software development methods. He has participated as a teacher in different training programs related to software engineering. Moreover, he is author of different books and conferences in software engineering congresses.

María-Isabel Sánchez-Segura has a degree in Computer Science Engineering and Master on Software Engineering and Knowledge Engineering. She is assistant lecturer at Carlos III University of Madrid. She is member of the “Virtual Environments Group” of the UPM, where she has been researching in several projects related to definition of Software Technology Transition Packages for Spanish Enterprises and development of Virtual Spaces for Individual and Collective Presence and Interaction (ESPRIT Program). Her research interests include software engineering, technology transfer and multi-user virtual environments. She has some publications and communications in International Congresses. Paloma Martínez Fernández received her degree in Computer Science from Universidad Politécnica of Madrid in 1992. Since 1992, she has been working as an assistant lecturer in the Department of Computer Science at Universidad Carlos III of Madrid. In 1998 she obtained the Ph.D. degree in Computer Science from Universidad Politécnica of Madrid. She is currently teaching Advanced Databases and Natural Language Processing. She has been working in several European and national research projects about Natural Language Processing, Advanced Database Technologies, knowledge-based systems and Software Engineering.

14 Seco, Guzmán, Segura, Fernández & Morillo

APPENDIX Table 3: Table of Tasks Distributed by Weeks Name of course: Group Nº: Components:

Date:

Work to be done: DATE

g n i h s i l b u P p u o r G

Practical Nº:

Estimated Time

TASK

a e d I t h g i r y p Co

Class

from 23/03/00 to 29/03/00

Specification Requirements document detailed reading Intermediate Level Data Flow Diagrams first approach (Yourdon, 1989) Intermediate Level Data Flow Diagrams beta version Data model construction Intermediate Level Data Flow Diagrams final version Rest of Data Flow Diagrams Entity/Relationship model (Batini, 1992) Check consistency between Process Model and Logical Model

Study

Exercises

practical

Total time 0

0

g n i h lis b u P p u o r G a e d I t h g i r y p Co g n i h s i l b u P p u o r G a e d I t h g i r y Cop g n i h lis b u P p u o r G a e d I t h g i r y p Co

from 30/03/00 to 05/04/00

from 06/04/00 to 12/04/00

from 13/04/00 to 19/04/00

from 27/04/00 to 03/05/00

from 04/05/00 to 10/05/00

from 11/05/00 to 17/05/00 from 18/05/00 to 24/05/00 from 25/05/00 to 01/06/00

0 0 0 0 0

0

Refine the Excel sheet which automate the estimation method MK-II (USKMA, 1998), using the features of the current project.

0

Count Inputs, Outputs and Entities for each one of the logical transactions identified. Assign values to each one of the 19 features in the MKII estimation method, for each logical transaction. Write the practice work memory including: process model, data model, estimation with the description of the features values. Organization, planning and tracking practice work section, reading. Reestimation if needed Metrica Versión 2.0 ,detailed reading.(Metrica, 1993) Work Breakdown Structure, Product Breakdown Structure and Resource Breakdown Structure building. Gantt diagram building, using the last estimation done.

0

0

0

0 0 0

0 0

Document the Project Plan.

0

Build Project Tracking cases.

Total

0 0

An Experience of Software Process Improvement Applied to Education 15

Table 4: Daily Registration Table

g n i h s i l b u P p u o r G a e d I t h g i r y p o C g n i h s i l b u P up o r G a e d I t h g i r y Cop g n i h s i l b u P p u o r G a e d I t h g i r y Cop g n i h lis b u P p u o r G a e d I t h g i r y p Co

Table 5: Table of Weekly Balance

16 Seco, Guzmán, Segura, Fernández & Morillo

Questionnaire 1: Questionnaire on Personal Work Planning Techniques

g n i h s i l b u P p u o r G Value

1

2

3

4

Don´t Know

Using the technique learnt, I was able to develop a clearer vision of the task to be completed. Therefore the time and resources that had to be assigned to each were foreseen.

Comments:

With this technique, the level of control to estimate time and changes for each task was greater.

a e d I t h g i r y p Co

Comments:

With this technique, the work team could do the tasks more realistically.

Comments:

g n i h lis b u P p u o r G a e d I t h g i r y p Co g n i h s i l b u P p u o r G a e d I t h g i r y Cop g n i h lis b u P p u o r G a e d I t h g i r y p Co

With this technique, the students succeeded in improving their own time estimations by applying previous experience acquired in similar activities.

Comments:

With this technique, the efficiency of the team members has improved.

Comments:

With this technique, the behavior of the team and the performance of each member improved the tracking of the practicals.

Comments:

I would use these techniques on different courses of my studies.

Comments:

I would use these techniques in different projects during my professional career.

Comments: