Solutions in global software engineering: A systematic literature review. Computer Science, Engineering. View 2 excerpts, cites methods and background. Project managers in global software development teams: a study of the effects on productivity and performance. Software Quality Journal.
View 3 excerpts, cites background. Highly Influenced. View 5 excerpts, cites background and methods. Offshore insourcing in software development: Structuring the decision-making process.
Engineering, Computer Science. View 4 excerpts, cites methods. View 4 excerpts, cites background and methods. View 1 excerpt, cites methods. View 1 excerpt, cites background. ICGSE Globalization has significantly changed the way the market operates today. In particular, the accessible population for this family of experiments was the group of B. The sampling approach was convenience sampling. On the one hand, this sampling approach is easy and readily available. On the other hand, the sample produced by convenience sampling might not represent the entire population i.
To increase the external validity of the results, we recruited a mix of B. SE students from four universities to take a part in a family of experiments. Previously in Section 3. Group G : participants in this group had to discuss a software design as represented by a graphical description UML class diagram. Group T : participants in this group had to discuss the same software design, but as represented by a textual description.
Explainer : this role consisted in: i understanding the design representation, and ii explaining it to a Receiver. Receiver : this role consisted in understanding the software design based on the discussion with an Explainer. Having the roles assigned, we randomly formed Explainer - Receiver pairs. These pairs were involved in discussing a design case which we detail in the next Section 3.
We created a design case for our family of experiments. The design case describes a structural view of a mobile application of a fitness center, the Fitness Paradise. This fitness center gives its clients the opportunity to book facilities and activities. The featured application enables clients to consult the schedule of activities, manage bookings, keep track of payments, and visualize performance data when available.
We believe that the selected design case relies on a familiar domain, Sport and Gym , from everyday life which is quite popular and easy to understand without prior knowledge. To introduce the Explainers with the design case, we created a design case specification Footnote 5 document which describes the fitness center and lists the features of the mobile application in natural language. The two design descriptions only differ in the way they represent the design i.
We created the UML class diagram Footnote 6 of the design case. The diagram includes 28 classes 21 model entities, 3 controllers, and 4 views and 30 relationships. We chose to use the Model View Controller MVC design pattern for structuring the design, as this pattern is well known by the participants of the experiments. The entities of each part of the MVC were given a specific color.
The model entities have a yellow color, the controllers are in blue, and the views are in green. The colors were added to the entities in the GSD in order to mimic the characteristic of structured textual document which we describe next in facilitating a visual distinction between different sections i. In other words, we were thoroughly keen to make both the graphical and textual designs present the exact same amount of information or design knowledge in order to control eventual bias due to a different amount of information.
The textual description Footnote 7 was arranged into two main structured sections. In the first section, we orderly described the entities of each module of the MVC: First the entities of the model part, then the entities of the controller part, and last the entities of the view part.
In the second section, we described the relationships between the entities following the same appearance order of the entities. Altered TSD. In particular, the description of the relationships of each entity was moved and placed right after the description of the entity. The main task of this family of experiments was inspired by the Chinese Whispers or The Telephone game.
In this game, players form a line, and the first player comes up with a message and whispers it to the ear of the second person in the line. The second player repeats the message to the third player, and so on.
When the last player is reached, they announce the message they heard to the entire group. In contrast, we created a message i. After that, the players i. Finally, the original message i. The main task of the experiments reflects common scenarios in software engineering industry where developers collaborate, communicate, and exchange knowledge in order to create software.
For example, the main task reflects a common knowledge-transfer scenario between a software architect i. Moreover, knowledge communication is especially important when new employees enter a company and struggle to learn the existing tacit knowledge. In this direction, our task reflects the scenario of onboarding of novice developers by experienced developers e.
Answer the pre-task questionnaire. All participants have to answer the pre-task questionnaire based on the group they are assigned to. No time-limit is imposed for this task. We noted the required time for this step during the experiments and found that it takes 15 minutes on average. The questions in the pre-task questionnaire varyaccording to the role of the participant Explainer vs. Discuss the Design i. The Explainers are allowed to individually ask questions to experiment supervisors to clarify issues related to the design, if required.
After 20 minutes, the Explainers give the design case specification back to the supervisors, but keep the design description GSD or TSD. Each Explainer is randomly paired with a Receiver from the same group. Then, each Explainer - Receiver pair is given 12 minutes defined based on to the pilot studies, see Section 3. The Receivers can unhesitatingly ask questions. Moreover to help the understanding process, Receivers are allowed to take notes during the discussion, but all notes are collected by the supervisors before the next task.
This is because two of the communication aspects, Understanding and Recall , that we measure require the participants to, respectively, apply and remember the design knowledge without using the design descriptions or the notes that they took during the discussions.
Answer the post-task questionnaire. All participants have to answer the post-task questionnaire based on their groups. We also noted the required time for this step and found that it takes 15 minutes on average.
The questions vary according to the role of the participant Explainer vs. G-Explainer : participants belonging to this subset have to answer questions on i how good they are in remembering UML models, ii how well they did understand and explain the design, and iii how much diagrams did help them in understanding and explaining the design.
G-Receiver : participants belonging to this subset have to answer questions on i how good they are in remembering UML diagrams, ii how well they did understand the design from the discussion with the Explainer , and iii how much diagrams did help them in understanding the design.
Two of these questions are open requiring free-text answers, six questions are multiple-choice questions which require the participants to choose only one choice, and two questions are check-boxes questions which require to select one or more answers from the available. To measure the Understanding , we formulated three questions Footnote 11 focusing on MVC design maintenance using maintenance questions to measure understanding is motivated in Section 4.
In each question we introduce a design maintenance i. The three questions are multiple-choice questions which require the participants to choose only one choice from 4 provided choices.
To evaluate the answers of the participants on recall and understanding questions, we defined grading rules that can be consulted online Footnote During the translation process, each word was carefully chosen to match the semantics of the original English textual description as close as possible. The only independent variable and manipulated factor is the design description. In this study, we consider six dependent variables See Table 3.
These variables correspond to the six communication aspects which we described in the introduction. The original experiment and replications were conducted under the same environment conditions and by following a well-defined protocol to ensure that the impact of any other variable on the results is relatively negligible. REP3 varies one variable intrinsic to the object of study i.
Before presenting the experiment procedure, we would like to highlight that we conducted several pilot studies, 2 in the university of Gothenburg, 1 in Aachen university, 1 in Lille university, and 1 in the Slovak university. To cover the treatments of our study, each pilot study involved 2 Explainer - Receiver pairs B.
One pair was assigned to the G group using a graphical design description, and the second pair was assigned to the T group using a textual design description.
These pilot studies helped us in:. Designing a research protocol and assessing whether or not it is realistic and workable, especially in estimating the time that is required by: i the Explainer to understand the design 20 minutes , and ii the Explainer and Receiver to discuss the design 12 minutes. Identifying logistical problems and determining what resources e.
The experiment procedure was created to define the process of the experiment and to ensure strict replications of the original experiment. Figure 1 presents the four main steps of the experimentation procedure:. Step 1 : To anonymize their identity and thus their answers, all participants were randomly assigned an identification number ID. We asked the participants to bring their PCs to be able to answer the online pre- and post-task questionnaires.
Eduroam Internet connection was available in the rooms where the experiments were running. Also, we asked the participants to bring a device to record the discussions either by downloading audio-recording software on their PCs or by using a smart-phone with a recording application. We booked large university lecture-rooms which can host all Explainer - Receiver pairs with a sufficient distance between each pair.
This helps to reduce voice interference to a minimum and produce better-quality audio recordings. We randomly assigned the participants to two groups G and T. Furthermore, we randomly assigned each participant one role, Explainer or Receiver. After that, we asked the participants to answer the pre-task questionnaire. Step 2 : Once all participants filled the pre-task questionnaire, the Explainers were taken to a second room where they received the design case specification and the design description GSD or TSD.
The Explainers were asked to understand the design that they received as good as they can in 20 minutes. We also informed the pairs that Receivers can ask clarification questions to the Explainers. This allowed us later to match the discussion records of the participants with their corresponding answers to the questionnaires. Step 4 : After 12 minutes, the participants were informed that they should stop the audio recording. Then, we asked all the participants to answer the post-task questionnaire individually.
Lastly, we asked the participants to rename the audio recordings with their ID numbers and put the recordings in a USB flash drive that we provided. The data of this study was collected via questionnaires and by audio-recording discussions between Explainers and Receivers. In this section, we describe three types of analysis procedures that we used:. Data Set Preparation : To check and organize data collected from different sources and prepare it for analysis.
Descriptive Statistics : To describe the basic features of the data by summarizing and showing measures in a meaningful way such that patterns might emerge from the data.
Hypothesis Testing : To make statistical decisions by evaluating two mutually exclusive statements about a population and determining which statement is best supported by the sample data. Meta-Analysis : To obtain a global effect of a factor on a dependent variable by combining the effect size of different experiments, as well as assessing the consistency of the effect across the individual experiments Borenstein et al.
In particular, 5 pairs discussed the design assignment for too short time less than 2 minutes and decided to discuss other topics of their interest for the rest of the time.
Moreover, the audio quality of the recorded discussion of 2 pairs was bad and the corresponding data from these pairs was eliminated. The final number of participants in each experiment is provided in Table 4.
The discussions between Explainers and Receivers were recorded by using either mobile phones or Audacity , an easy-to-use audio editor and recorder that works on multiple operative systems Footnote We transcribed approximately 23 hours of audio recordings and performed a manual coding of more than discussion records between Explainers and Receivers. For coding the discussions, we used the collaborative interpersonal problem-solving skill taxonomy of McManus and Aiken , as presented in Figure 2.
This taxonomy captures the collaborative interpersonal communication aspects; Active Discussion, Creative Conflict, and Conversation Management, which we described previously in Section 1. More examples are provided online Footnote Collaborative interpersonal problem-solving conversation skills McManus and Aiken NVivo Footnote 15 was used for coding the transcriptions. Based on this result, the raters collaboratively continued to code the rest of the data i. In particular, we measured: means, medians, standard deviations, and ranges.
These descriptive statistics help to analyze central tendencies and dispersion. So, we assigned our participants to these two groups by following the between-subjects design. In this setting, different people test each condition to reduce learning- and transfer-across-conditions effects. The collected data during the experiments include both interval and ordinal measures.
Moreover, they are not normally distributed. Thus, we used non-parametric tests. In particular, the hypotheses that we formulated in Section 3. Therefore, these hypotheses were tested by performing the non-parametric independent-samples Mann-Whitney test. We perform a fixed-effect meta-analysis, as all factors that could influence the effect size are the same in all the experiments Borenstein et al.
We use different scales to measure the communications aspects. The assigned weight to each experiment is:. We report the result of the meta-analysis by using forest plots Borenstein et al.
After that, we present the results of the individual experiments and the performed meta-analysis. The perceived based on self-evaluations design experience and communication skills are detailed here Footnote In summary, we find that:.
There are no statistically significant differences in the perceived design experience and communication skills between groups G and T in the different experiments. Accordingly, we assume that the design experience and communication skills of participants are not influencing the results of this study. Table 5 shows the descriptive statistics of the studied communication aspects sorted by two subgroups of studies:. Regarding Conversation Management , the results show that the participants of all the experiments spent more effort on conversation management when using TSD.
Considering Subgroup B , we observe that the unbiased estimate of the effect size i. Moreover, the participants spent more effort on conversation management when using Altered-TSD. We tested whether or not the distribution of the communication aspects i. Table 6 shows the results of the test. The p-value is the probability of obtaining the observed results of a test, assuming that the null hypothesis is correct.
We set the probability of type I error i. The statistical power is the probability that a test will reject a null hypothesis when it is in fact false. A power value of 0,80 is considered as a standard for adequacy Ellis In this section, we report and discuss the meta-analysis by means of forest plots. The squares in each forest plot indicate the effect size of each experiment.
The diamond shows the global effect size the location of the diamond represents the effect size , while its width reflects the precision of the estimate i. The plot also shows the values of the effect size, weight, and p-value relative to each experiment and to the global measure. Figure 3 shows the forest plot for perceived quality of Explaining in the two subgroups of studies, A and B. We observe that the effect size values are positive in all the experiments.
This implies that using a GSD has a positive effect on perceived Explaining quality. Figure 4 shows the hierarchy of the six cognitive learning levels.
According to Anderson, remember is the recalling of the previously learned topic. Apply instead, comes on top of understand. It is the ability to use the acquired and comprehended knowledge in a new and concrete context or situation. The participants in the two groups G and T answered ten recall questions. We formulated the recall questions see Section 3.
Figure 5 shows the forest plot for quality of a Understanding and b Recall ability of design details. Regarding the quality of Understanding , the effect size value is negative for OExp, which means that TSD is the improving condition.
For the other experiments in subgroups A and B the values of the effect size are positive. This implies that using a GSD in these experiments has a positive effect on the understanding quality. Despite these tendencies, the global effect size of subgroups A and B is not statistically significant p-values are 0, and 0,, respectively.
Considering the Recall ability, we observe that the effect size values in subgroup A are positive. This implies that using a GSD has a positive effect on the Recall ability. This effect is statistically significant and has a medium effect size for REP1.
In contrast, the effect size value in subgroup B is negative. This implies that using a Altered-TSD is the improving condition. Considering AD, we observe that the effect size values of Subgroup A studies are positive.
The effect size value of Subgroup B study is negative. Considering CC, we observe that the effect size values of all the studies are positive.
Considering CM, we observe that the effect size values of all the studies are negative. Falessi et al. Tang et al. They found that practitioners recognize the importance of documenting design rationale for reasoning about their design choices and supporting the subsequent implementation and maintenance of systems.
To achieve the goal of REP3, we use a fixed-effect subgroup analysis Borenstein et al. In particular, we compare the mean effect for two subgroups of studies:. For each subgroup of studies, we report in Table 7 the mean effect size and variance of the studied communication aspects.
We also observe that the effect size of CM is higher in Subgroup B. The results of the test are presented in Table 8. In addition to reporting the test of significance, we report the clinical significance. The perceived experience based on self-evaluations in working with different design representations are detailed here Footnote Our experiments investigate whether design communication between software engineers can become more effective when using GSD instead of TSD to exchange design information.
Moreover, we study whether a cohesive and motivated TSD i. Considering Subgroup A , the global effect size of the perceived explaining quality is positive. This means that using a GSD has a positive effect on the perceived explaining quality.
Similarly, the global effect size of the understanding i. Nevertheless, by considering distributions of the scores we neither find a statistically significant difference in the quality of explaining Observation 1 nor in the quality of understanding Observation 2 between the two groups: G and T.
While analyzing the recorded, and further transcribed, discussions between the Explainers and Receivers , we interestingly observed a difference in the explaining approach between the Explainers of the two groups.
Figure 7 provides an illustration of the observed explaining approaches in the two groups. On the one hand, the Explainers of a TSD tended to explain the three modules of the MVC sequentially: Firstly the model entities, then the controllers, and lastly the views, as these modules are orderly presented in the textual document.
We think that this trend is intrinsically imposed by the nature of textual descriptions where the knowledge is presented sequentially on a number of consecutive ordered pages.
On the other hand, the Explainers of the GSD had more freedom in explaining the design. Indeed according to their explaining preferences, the Explainers of the GSD tended to jump back and forth between the three MVC modules when explaining the design. However, developers might not have this advantage when explaining many GSDs e. This is actually inline with Meade et al. One of the recall questions that we used to measure the recall ability of the participants is concerned with the relationships between the entities of the software architecture design.
We compared the score interval variable; min is 0 and max is 1 point of the two groups on this question. However, this difference is not statistically significant Sig. The Chinese Whispers game is often invoked as a metaphor for miscommunication. Keywords Software engineering Globalization Cost benefit analysis Research initiatives systematic literature review Software engineering Globalization Cost benefit analysis Research initiatives systematic literature review. Additional information Data set: ieee.
Publisher IEEE. Fields of science No field of science has been suggested yet. You have to log in to notify your friend by e-mail Login or register account. Download to disc. High contrast On Off. Close window. Assign to yourself. Assign to other user Search user Invite. Assign Wrong email address.
0コメント