Software Development based on Server-Driven UI Architecture

Systematic Mapping Study – Research Protocol Software Development based on Server-Driven UI Architecture X Y Z September 31, 2025 Contents 1 Introduction 1 2 Review Protocol 1 2.1 Research Questions ............................... 2 2.2 Search Procedure ................................ 2 2.3 Study Selection ................................. 3 2.4 Keywording of Abstracts ............................ 4 2.5 Data Extraction and Mapping ......................... 5 2.6 Threats to validity ............................... 6 3 Report the Review 7 A Primary Studies 9 i List of Tables 1 Selected Databases ............................... 3 2 Search string .................................. 4 3 Data extraction form .............................. 5 4 List of primary studies analyzed in the mapping ............... 9 ii 1 Introduction This study aims to provide a fair and comprehensive overview of the development of software based on server-driven user interfaces. To do so, we have adopted a mixed approach that combines a systematic mapping study (SMS) with an investigation of the gray literature. This SMS was conducted in strict accordance with the guidelines proposed by Petersen et al. (2015). In summary, these guidelines were methodically organized into a five-step process, which is outlined as follows: •Step 1 – Definition of Research Questions.This step reports the research objectives for this systematic mapping. In short, a research protocol must have a predetermined plan that describes research questions and guides how the systematic mapping must be conducted; •Step 2 – Conduction of Search.This step describes how the primary studies must be searched and identified; •Step 3 – Screening of Papers.This step reports how the identified studies must be evaluated according to the inclusion and exclusion criteria defined in this protocol; •Step 4 – Keywording of Abstracts.This step shows how the included studies must be analyzed and classified; and •Step 5 – Data Extraction and Mapping of Studies.This step describes how the final representation must be built, based on the final scheme created in Step 4. Regarding the use of gray literature, the search string originally employed in the mapping process was adapted for the search engines available on the Web1. Next, the approach adopted in our investigation will be detailed in the following sections. 2 Review Protocol This section explains the steps of the proposed SMS. In short, a research protocol is a plan that describes research questions and how to conduct the study. The essential steps of an SMS process are (i) definition of research questions; (ii) searching for relevant papers; (iii) screening of papers; (iv) key wording of abstracts; (v) data extraction and mapping; and (vi) threats to validity. Next, we reported a description of each step. 1https://www.google.com 1 2.1 Research Questions The main purpose of this study is to identify primary studies that report some evidence concerning the development of software based on server-driven user interfaces. This study also aims to get a comprehensive overview of the features of this software type. To do so, we established the following research questions (RQ): RQ1: How has the server-driven UI architecture been used in software design? The purpose of this question is to identify the primary studies published in the literature (or gray literature) that reported some evidence of the development of software based on server-driven user interfaces. RQ2: What are the demographic characteristics of these studies? The purpose of this question is to understand some aspects related to primary studies. To gather more information on these studies, we elaborated four sub-questions, namely: RQ2.1: When were the studies published? RQ2.2: Where were these studies published? RQ2.3: What is the nature of these studies? RQ2.4: What is the mature level of these studies? RQ3: How has the development of software based on server-driven user interfaces been organized? This question aims to provide substantiation of the organizational structure implemented in the development of software founded on server-driven user interfaces. RQ4: What are the key elements of a server-driven UI architecture? The purpose of this question is to understand the principal elements (or essential elements) of the server-driven UI architecture so that they can be used in software development. 2.2 Search Procedure Considering the RQs presented in Section 2.1, we defined the search strategy, which is composed of source selection criteria, sources list, studies language, and keywords and their related terms. •Sources selection criteria: We adopted the following criteria to select search sources (i.e., publication databases and search engines) to find primary studies: (i) content update (i.e., publications are regularly updated); (ii) availability (i.e., the full text of the papers are available); (iii) quality of results (i.e., the accuracy of the results returned by the search); and (iv) versatility to export the results (i.e., since much 2 information are returned through the search, a mechanism to export the results is required) (Dieste et al.,2009); •Sources list: The publication databases selected in our study are shown in Table 1. The ACM Digital Library, IEEE Xplore, Science Direct, Scopus, SpringerLink, and Web of Science is an efficient set of publication databases, namely the most used in computer science and widely used in the context of software engineering (Petersen et al.,2015). To avoid missing any relevant study, we applied the backward snowballing technique to the selected studies (i.e., second phase), and included the subsequent results as gray literature in our SMS (Petersen et al.,2015). Table 1: Selected Databases Source Location ACM Digital Library https://dl.acm.org IEEE Xplore https://ieeexplore.ieee.org ScienceDirect https://sciencedirect.com Scopus https://scopus.com SpringerLink https://link.springer.com Web of Science https://www.webofscience.com Besides the publication databases listed in Table 1, we adapted the search string (Table 2) for use in Google Scholar2. •Studies language: Only primary studies written in English will be considered in this investigation, since it is the most common language in scientific papers; •Keywords: According to the preliminary investigation, the term “server-oriented user interface” is the most suitable to represent the research interest. Next, this term was further divided into the following keywords: server-driven, server driven, user interface, UI. •Search string: We identified one term related to four known keywords to answer the RQs. Thus, we used the Boolean operator OR to link the main terms and their synonyms, and the Boolean operator AND to combine these terms, as shown in Table 2. 2.3 Study Selection Another important element of planning is to define the Inclusion Criteria (IC) and Exclusion Criteria (EC). These criteria enable us to include primary studies that are relevant 2https://scholar.google.com 3 Table 2: Search string (“server-driven” OR “server driven”) AND (“user interface” OR “UI”) to answer the research questions and exclude studies that do not answer them. Thus, we defined one inclusion criterion for our SMS, namely: •IC1: The primary study presents some evidence about server-driven user interfaces in software development. This criterion (IC1) serves as the basis for our selection of relevant studies for this review. Similarly, it is equally important to outline our exclusion criteria. The exclusion criteria established are: •EC1. The study did not provide any evidence in relation to the use of server-driven user interfaces in software development; •EC2. The study is not written in English; •EC3. The study only provides a summary or is not available in full-length; •EC4. The study is similar to previous work developed by the same author. Here, only the most recent study or the most complete version must be considered; •EC5. The study is a secondary study; and •EC6. The study was written by an author of our research group. 2.4 Keywording of Abstracts This summarizes the conduction phases of our study, also showing the number of selected studies at the end of each phase. Phase 0 represents the total number of studies returned by the search strings (i.e., 73 studies), as well as the number returned for each publication base (ACM Digital Library – 27, IEEE Xplore – 1, ScienceDirect – 1, Scopus – 4, SpringerLink – 40, and Web of Science – 0). We encountered duplicate studies within the databases, prompting our first refinement step. This step involved removing the duplicates, resulting in 70 studies (Phase 1). Next, we unified all the studies into a single file and eliminated any duplicate studies that were found in more than one database, resulting in 77 studies (Phase 2). We selected 6 studies for full reading after applying our selection criteria. Regarding this phase, it is worth highlighting that divergences in 4 criteria were resolved through discussion among review participants to standardize inclusion or exclusion from the study. After carefully reading the six studies, we selected five to compose the final list of our study (Phase 4). In Phase 5, we conducted a gray literature search using the Google3search engine. As a result, 15 studies were included in our investigation, totaling 20 studies. Appendix A presents the complete list of studies selected in our mapping. 2.5 Data Extraction and Mapping To increase the reliability of our results, we reviewed all the studies selected before proceeding to the data synthesis (Kitchenham and Charters,2007). The forms (see Table 3) filled in the second selection were analyzed again and, whenever necessary, important parts of each study or its full text were read to identify potential classification errors. During the review, we extracted more information from some studies, particularly the ones that did not have enough information extracted in the previous step, and did not remove any study, maintaining the 20 primary studies selected. Table 3: Data extraction form Research Topic Description RQ ID Integer – Article title Title of the article – Author Name Set of names of the authors – Author Country Set of names of the countries – Design Set of evidence on the development of applications based on SDUI architecture. RQ1 Year Calendar year RQ2.1 Venue Name of the source of publication that retrieved the study RQ2.2 Nature Academic, Industrial, or Mixed RQ2.3 Mature Proof of concepts, case studies, experiments, among others RQ2.4 Organization Applications’ organization RQ3 Key elements Set of elements used in the SDUI architecture RQ4 During the data collection phase, we will collect a set of variables that describe each primary study. Data collection is going to be handled by members of the team and conflicts will be resolved by a supervisor. Each researcher involved will perform data extraction for every primary study separately. Then the two values of each variable obtained by 3https://www.google.com 5 each researcher will be compared to each other and the final value of the variable will be assigned to the primary study after a discussion on every researcher’s opinion. If both researchers assigned the same value to one variable, we will assign this value to the variable without further discussion. On the other hand, after a debate among the researchers and their supervisor, a value will be assigned to each variable. We will not apply an agreement measure as the number of researchers involved in the review is not significantly large. However, all conflicts will be recorded. We extracted assigned values to the attributes for each study, as presented in Table 3. 2.6 Threats to validity To enhance the trustworthiness of our study and reduce the biases that could be inserted by authors of the selected studies, we report and discuss the main aspects that might represent threats to validity, and the actions performed to mitigate them. We have strictly followed the guidelines proposed by Kitchenham and Charters (2007), Kitchenham et al. (2010), and Petersen et al. (2015) to achieve an impartial review. These guidelines have been widely adopted in SLR, surveys, and SMS (Filisbino Passini et al.,2022;Souza et al.,2018;Wedyan and Abufakher,2020;Zavala et al.,2019). Next, a description of the threats of our study is reported: Omission of papers. As reported in Section 2.2, six publication databases were used to extract the data analyzed in this paper. According to Dybå et al. (2007), Kitchenham and Charters (2007), and Petersen et al. (2015), such databases are sufficient and the most relevant sources in the software engineering area. Despite our efforts to make this process as reliable as possible, it is still possible that primary studies were missed and not reported. Although we have conducted evaluations to calibrate and ensure the efficiency of our search string, our study depends on the search engines of each research database. Thus, in order to minimize problems in the search phase, we adapted the general search string to meet the constraints/limitations of each publication database. Reviewers’ reliability. Only one researcher was responsible for the conduction of our study, which can result in bias risk. Therefore, to ensure the reliability of our study and that it was conducted unbiasedly, software engineering specialists supported the researcher in this process during the conduction meeting to solve questions as to the inclusion or exclusion of studies. Moreover, none of the primary studies included was developed by our research group or researchers related to us. Therefore, we are not aware of any bias we may have introduced during the analysis. Data extraction. We have analyzed or interpreted data extracted from the primary studies because not all data was clear. Since this activity involves human judgment, 6