Automated analysis of feature models: Quo vadis?

Computing manuscript No. (will be inserted by the editor) Automated analysis of feature models: Quo vadis? José A. Galindo ·David Benavides ·Pablo Trinidad · Antonio-Manuel Gutiérrez-Fernández ·Antonio Ruiz-Cortés. Received: date / Accepted: date Abstract Feature models have been used since the 90’s to describe software product lines as a way of reusing common parts in a family of software systems. In 2010, a systematic literature review was published summarizing the advances and settling the basis of the area of Automated Analysis of Feature Models (AAFM). From then on, different studies have applied the AAFM in different domains. In this paper, we provide an overview of the evolution of this field since 2010 by performing a systematic mapping study considering 423 primary sources. We found six different variability facets where the AAFM is being applied that define the tendencies: product configuration and derivation; testing and evolution; reverse engineering; multi-model variability–analysis; variability modelling and variability–intensive systems. We also confirmed that there is a lack of industrial evidence in most of the cases. Finally, we present where and when the papers have been published and who are the authors and institutions that are contributing to the field. We observed that the maturity is proven by the increment in the number of journals published along the years as well as the diversity of conferences and workshops where papers are published. We also suggest some synergies with other areas such as cloud or mobile computing among others that can motivate further research in the future. Keywords Software product lines ·Automated analysis ·Feature models ·Variability-intensive systems. 1 Introduction Software Product Lines (SPLs) are about developing a set of different software products that share some common functionality [7]. Documented benefits of SPLs are the increment of reuse, quality and reduction of time to market to achieve mass customization of software products in order to satisfy the customer needs as well as decreasing the effort of personalization. Feature Models (FMs) [15] are tree–like structures that define the set of products that belongs to a SPL and have become the de facto standard to represent common and variable characteristics in a SPL. It is easy to find different graphical and textual notations in the literature to represent FMs [4]. The information that can be obtained from FMs is extensive, and the mechanisms for obtaining it are likewise varied. In fact, this area known as Automated Analysis of Feature Models (AAFM) [3] has recently been identified as one of the most important areas in the SPL [12] community. The AAFM is the computer–aided extraction of information from feature models [4] and can be summarized in three steps. First, FMs are translated to a logical representation. Second, an off–the–shelf solver or specific algorithm is used to perform a given analysis operation (such as counting the number of products or checking the consistency of a FM). Finally, the result is obtained and used in a determined context to perform other tasks such as product configuration or derivation. A comprehensive list of proposals and operations for the AAFM was presented by Benavides et al. [4] in 2010 that settled the conceptual underpinnings of the discipline. In particular, thirty analysis operations José A. Galindo ·David Benavides ·Pablo Trinidad ·Antonio-Manuel Gutiérrez-Fernández ·Antonio Ruiz-Cortés Dept. Lenguajes y Sistemas Informáticos, University of Seville, Avda. Reina Mercedes s/n, 41012, Seville, Spain. E-mail: {jagalindo,benavides,ptrinidad,amgutierrez,aruiz}@us.es 2 José A. Galindo et al. were presented, and further formalized in [8], different automated mechanisms identified such as SAT, CSP or BDD solvers, a conceptual framework described and directions for future work established. Operations range from determining if a product is valid with respect to a FM to the calculation of the number of different products in a product line. AAFM has been applied in different activities along the SPL process such as product configuration and derivation, reverse engineering or SPL testing. In this paper, we present a Systematic Mapping Study (SMS) to identify the evolution and trends in the application of the AAFM since 2010. Concretely, we have performed a search on different databases of AAFM–related papers. We selected 423 primary sources (papers) that followed the defined inclusion and exclusion criteria. The primary sources were classified according to different variability facets that were found during the reading and key–wording phase. It is important to remark that before 2010, AAFM was not well defined and it was referenced using an amalgam of names and concepts. Therefore, we consider that in 2010 the concept of AAFM was coined and then used in different domains and scenarios. This paper studies how AAFM has been used since its definition. We discovered six different variability facets where the AAFM is being applied: i)product configuration and derivation. Automated support is used to guide the configuration process and the derivation of specific products. This is the most traditional usage of automated analysis mechanisms; ii)testing and evolution. Specific configurations are selected for testing purposes using automated mechanisms. These can support the automated-guided evolution of feature models; iii)reverse engineering. Extracting feature models from product descriptions and, in some situations, from logical formulas; iv)multi-model variability–analysis. Traditionally, automated analysis operations have been proposed over a single model. However, there are situations where analyses are performed over more than one model; v)variability modelling. The basic modelling constructs are not enough in some specific situations and other information such as attributes are used for modelling different situations and analysis are performed with these new modelling elements, and; vi)variability–intensive systems. AAFM is used in other application domains not directly related with SPLs. We also observed that there are only a few industrial and real evidences of the application of AAFM techniques in most of the cases. We detect in detail where and when the papers have been published and who are the authors and institutions that are contributing to the field. We saw that the maturity is proven by the increment in the number of journals published along the years as well as the diversity of conferences and workshops where papers are presented. Finally, we devise some research opportunities and applications in the future as well as synergies with other research areas. The remainder of this paper is organized as follows: Section 2 presents the background required to understand the scope of this paper; Section 3 discusses the previous related works that are related to this research area; Section 4 covers the methodology used in this mapping study; Section 5 presents results from analyzing the primary studies; and Section 6 presents concluding remarks. 2 Background 2.1 Feature models Feature Models (FMs) are one of the most used artefacts to describe the set of products in a SPL in terms of features and relationship among them. In FMs, features are hierarchically arranged in a tree–like structure. In addition, cross–tree constraints can be used to connect features. Figure 1 shows an example of a FM describing a mobile-phone SPL, using the most common notation (Figure 1a) and its represented products (Figure 1b). There are different proposals for FM notations (see [26] for a detailed survey) but most of the proposals have the following common elements: –Mandatory: a child feature has a mandatory relationship with its parent feature when it appears in a product whenever its parent does. In Figure 1, Screen has a mandatory relationship with Mobile Phone, i.e. any mobile phone must have a screen. –Optional: a child feature has an optional relationship with its parent feature when the child can appear or not in a product whenever its parent does. In the example in Figure 1, Media has an optional relationship with Mobile Phone, i.e. A mobile phone can have support for media features or not depending on the configuration chosen. Automated analysis of feature models: Quo vadis? 3 Only one One or more OptionalMandatory ExcludesRequires Mobile Phone Calls GPS Color Basic Screen Media Camera MP3High Resolution Requires Excludes (a) A sample feature model of a SPL for mobile phones using a FODA–like notation. (Source: [4]) # Features 1 Mobile Phone, Calls, Screen, Basic 2 Mobile Phone, Calls, Screen, Color 3 Mobile Phone, Calls, Screen, High Resolution 4 Mobile Phone, Calls, Screen, High Resolution, Media, Camera 5 Mobile Phone, Calls, Screen, Basic, Media, MP3 6 Mobile Phone, Calls, Screen, Color, Media, MP3 7 Mobile Phone, Calls, Screen, High Resolution, Media, MP3 8 Mobile Phone, Calls, Screen, High Resolution, Media, Camera, MP3 9 Mobile Phone, Calls, Screen, Color, GPS 10 Mobile Phone, Calls, Screen, High Resolution, GPS 11 Mobile Phone, Calls, Screen, Color, GPS, Media, MP3 12 Mobile Phone, Calls, Screen, High Resolution, GPS, Media, MP3 13 Mobile Phone, Calls, Screen, High Resolution, GPS, Media, Camera 14 Mobile Phone, Calls, Screen, High Resolution, GPS, Media, Camera, MP3 (b) The set of products depicted by the feature model Fig. 1: A sample feature model among it set of represented products –Or–relationship (also known as OneOrMore): a set of child features has an or–relationship with its parent when one or more child features can be selected when the parent is. Figure 1 contains an or–relationship between Camera and MP3 with Media. Whenever Media is present in a product, Camera, MP3, or both have to be present. –Alternative (also known as OnlyOne): a set of child features have an alternative relationship with their parent when one and only one of them can be selected in a given product whenever their parent is selected. Figure 1 shows an alternative relationships among Basic,Colour and High Resolution, so a given mobile phone can only have a specific type of Screen in a product. –Requires,Excludes: Cross–tree relationships like A requires B means that whenever feature Aappears in a product, feature Bmust also appear. Also, a relationship like A excludes B means that both features cannot appear in the same product at the same time. Figure 1 shows two examples of these kinds of relationship: the Camera requires a High Resolution screen and the GPS excludes the Basic screen. 2.2 Automated analysis of feature models The AAFM deals with extracting information from FMs by using computer–aided mechanisms. SPL engineers use the information to improve their business strategies as well as to take technical decisions. The process to extract such information is shown in Figure 2. It starts by translating the features and relationships encoded in the FM and any other additional information (e.g. market share [10]) to a knowledge base described in a logic paradigm. Later, queries to the knowledge base can be performed using existing solvers or tools thus, obtaining the analysis results. In [4], different analysis operations on FMs were reported. According to that study, we present some of them: –Finding out if a product is valid. This operation checks if a product (i.e. set of features) belongs to the set of products represented by a FM or not. It is helpful for SPL engineers and managers to determine whether a given product is available in a SPL. –Obtaining all products. This operation lists the products represented by a FM. It allows practitioners to identify the final products that they can manage in their SPL. For example, the model shown in Figure 1a represents the products in Figure 1b. –Calculating the number of products. This operation counts the number of products of a FM. This provides information about the size and complexity of the SPL represented by a FM. It is commonly used to perform more complex operations such as calculating the amount of reuse metrics of a SPL. For example, in Figure 1a there are fourteen products. –Detecting errors. The large number of different features used in a FM increases its complexity as well as the probability of introducing errors. There are several types of errors that can be detected by using the AAFM. For example, it is important to determine if a FM is void, i.e. whether it represents no product at all because of contradicting relationships. Another common error is the detection of dead features, i.e. features that cannot appear in any of the products derived from the model. Dead features are clearly undesired since they are the result of a wrong domain modelling. –Explaining errors. As shown before, there are circumstances where FMs contain errors. In such situations, it is important to assist on resolving them. An explanation operation takes a FM as input and a set of previously identified errors, trying to provide insights to correct them. 4 José A. Galindo et al. Logic&Paradigm& Solver& or& Tool& Transla3on& Knowledge) Análisis)Opera2ons) Filter) Void) #Pr) List) Dead) Expl) Corr) Expl)Var) …) Data) Feature)Model) Analysis) result) Fig. 2: AAFM framework. These operations are performed automatically using different approaches. Most of them translate FMs into specific logic paradigms such as propositional logic, constraint programming or description logic. Others propose ad–hoc algorithms and solutions to perform these analyses [4]. Finally, these analysis capabilities can also be found in several commercial and open source tools such as pure::variants1,SPLOT [21], FaMa [5], FeatureIDE [245] or FAMILIAR [1] 2.3 Literature review methods To crawl the existing knowledge in the literature there are different kinds of reviews [11], but we focus on two of them. Figure 3 depicts the main difference among them and how they are related.  Primary sources heterogeneity  Broad scope  Large set of papers  Visualization  Focused on directions and trends of an area FACET FACET ...  Primary sources homogeneity  Concrete scope  Small set of papers  Focused on the details of results for an area SMS SLR Mapping Comparison between related research  Conferences, journals...  Temporary evolution  Institutions and authors Fig. 3: Systematic mapping studies and literature reviews 1http://www.pure-systems.com/ Automated analysis of feature models: Quo vadis? 5 –Systematic Literature Reviews (SLRs) take primary sources, i.e. papers under study, that are homogeneous in terms of comparability and compare them to get conclusions [17]. All papers of the review are comparable among different aspects. SLRs provide a synthesis of the knowledge existing in the core content of primary sources of a specific field and, to do so, researchers have to read, understand and classify the whole content of the studies. SLRs provide as output as many details and comparisons as possible among the content of primary sources. In other words, they are focused on “how” the things have been done in a concrete field. This is, SLRs target a reduced set of papers and extract all the insights possible from them. –Systematic Mapping Studies (SMSs) take primary sources that are heterogeneous in terms of comparability but are related to a broader area and provide a mapping and categorization of the different facets detected in the studies [24]. Researchers read the title, abstract, and optionally other parts of the paper stepwise. The idea is to get the whole picture of a broad research area. SMSs provide visual outputs that allow an easy identification of research gaps. For example, most of SMSs provide bubble plots or heatmaps that show how well covered is an intersection between the categories analysed. Note that SMSs do not analyse specific research but allow the characterization and classification of more heterogeneous papers. In other words, the main focus of an SMS is to detect “what have been done”. Petersen et al. [24] detailed the process of building systematic mappings and compared them with SLRs. SMSs take as input as many papers as possible reducing the bias of only reading portions of the papers in the case of a mistake in the classification. There are potential relations between SMSs and SLRs as shown in Figure 3. The results of an SMS can be the basis for a more in depth SLR of one or several of the research facets that are part of the mapping. For instance, an SLR of testing in mobile phone applications can then serve as the basis for performing an SLR of techniques for automated unit tests in that context, which is a more concrete field. SLRs and SMSs usually provide common outputs such as the fora where authors have published, temporary evolution of primary sources as well as authors and institutions contributing to the area of study. 3 Related work As explained in Section 4 there are different kinds of review methods. In this Section we go through other systematic reviews identifying the context, year and period. In the area of SPLs, several SMSs or SLRs can be found [4, 9, 18, 19, 22, 23] as shown in Table 1, most of them being published in recent years. The period of time used for the reviews varies depending on the topic, ranging from 1990 (where SPLs started to be popular) to the date where the study was performed 2. Also, the number of analysed contributions of our work reviewed more studies. As explained in Section 4.2, we used 423 papers for extracting some general data and 242 for more concrete insights. Review work Kind Year Context Period # papers studied Benavides et al. [4] SLR 2010 Automated analysis of feature models 1990–2010 53 Da Mota et al. [23] SMS 2011 SPL testing 1993–2009 120 Engström and Runeson [9] SMS 2011 SPL testing 1990–2008 64 Laguna and Crespo [18] SMS 2013 SPL evolution 1990–2011 74 López-Herrejón et al. [19] SMS 2015 Search-based SPLs 2011–2014 77 Montalvillo and Díaz [22] SMS 2016 Requirement-driven evolution in SPLs 1990-2015 107 This paper SMS 2018 Automated analysis of feature models 2010–2017 423/242 Table 1: Comparison of related reviews Heradio et al. [13] presented a bibliometric analysis of 20 years of SPLs, from 1995 to 2014. One of the conclusions was that "feature modelling has been the most important topic for the last fifteen years, having the best evolution behaviour in terms of number of published papers and received citations"[13]. 2López-Herrejón et al. [19] reduce the period of time due to the topic handled. 6 José A. Galindo et al. Also, in the 2009-2014 period, the AAFM was the most influential topic according to the number of papers produced and the centrality and density of papers. Note that [13] is not an SMS nor an SLR and only performs data analysis based on automated mechanisms. 3.1 Need for a systematic mapping study Benavides et al. [4] analysed in detail the relevant literature that builds up the body of knowledge of the AAFM. That work analysed and categorized concrete methods for performing automated analysis of feature models. Also, authors identified different analysis operations and several automated mechanisms for performing the operations (see Section 2.2). We conjecture that in [4] the grounds of the field of AAFM were settled and from then, most of the papers are using analysis operations to solve different problems in SPL engineering instead of defining new operations or proposing new analysis mechanisms or algorithms. In this paper, we want to detect the main trends where the AAFM is being used, this is, we want to include in this study any paper that uses the AAFM in any area of the software product line engineering process (e.g. testing, requirements, derivation,...). In this context, we performed an SMS which is the most suitable review method to cover a wider set of publications and understand the current and future state of this area. Section 4.1.1 shows the set of questions that will guide our study and will help us to achieve our goal. The authors of this work belong to one of the most active research groups in the area of feature modelling, and automated analysis according to [13], being these topics on the ones that attracted more attention in recent years. As the field evolves, there is a need to evaluate the trends in the area and assess their maturity. With this work, we aim to serve as evidence for research opportunities both in the kind of research (i.e. more empirical vs more theoretical) and in the topics to be covered. 4 SMS process definition This SMS follows a process inspired by the one proposed by Petersen et al. [25] summarized in Figure 4. The main steps to perform a systematic mapping study are: 1. Planning the review, which includes the definition of the search protocol, the survey of the literature and the definition of the research questions; 2. Study identification; where the databases are crawled and the primary studies selected; 3. Data extraction and classification, where the mapping is developed and conclusions are obtained. Phase 2: Study identification Phase 3: Data extraction and classification Phase 1: Planning the review Fig. 4: Systematic mapping study process [22, 25] Next, we detail the process that guides this SMS. Automated analysis of feature models: Quo vadis? 7 4.1 Phase 1: Planning the Review This phase comprises the execution of three process steps as shown in Figure 5: i) protocol definition, where we decide how to do the review and how to minimize the threats to validity; ii) literature survey, that consists of getting the base of the body of knowledge of the area, and; iii) the definition of research questions, where the questions to be answered are stated. Next, we detail how we performed each of these steps. Protocol definition Literature survey Research question definition Protocol and data collection form Set the grounds Review scope Phase 1: Planning the review Phase 2: Study identification Phase 3: Data extraction and classification Process step Outcome Legend Fig. 5: Planning the review phase Protocol definition. The protocol that we defined follows the guidelines of Petersen et al. [25] to perform systematic mapping studies. Then, we compared the research questions with Jia et al. [14] 5W+1H model to check their accuracy (see Section 4.1.1). Also, we modified the protocol to obtain not only the results from databases but also considering the papers quoting Benavides et al. [4] (see Section 4.2). Literature survey. In order to survey the current state of the art in the AAFM, we read the papers published in the main conferences by means of authors expertise and journals as well as taking the following papers as the main input defining the body of knowledge [4, 8]. Also, we relied on the expertise of some of the authors of this paper that have been working in the AAFM and collaborating with the community for the last ten years or so. 4.1.1 Definition of research questions Defining a set of research questions that guide an SMS is a difficult task that could lead to a biased study, which is aggravated if the authors have experience in the area. To guide and reduce the arbitrariness of the questions definition, we were inspired by the 5W+1H pattern proposed by Jia et al. [14] using it to verify the completeness of our proposed questions. This pattern relies on the 5W+1H model, a pillar of journalism to report stories originally proposed by Kipling [16]. 5W+1H is an abbreviation of Who,Why, What,Where,When and How, the six questions to be answered in order to know the most important aspects about a story. Thus, we defined the following research questions for our study: RQ1 Where are the papers published? Which fora are being targeted by practitioners in the last six years? We think that this question can help researchers to know which communities are interested in which topics. Therefore, to target the most suitable forum for a new contribution. RQ2 Who are the authors and institutions that make research into AAFM? Which authors have been more active regarding the AAFM area and which institutions are currently hosting them? This question aims at fostering researchers collaboration and to determine which institutions hold the knowledge of a topic. RQ3 What are the areas for which AAFM has been applied? What topics have attracted more researchers attention in the last years? This research question aims at discovering “Why” the researchers have applied AAFM techniques in the sense of the applications where AAFM has been attracting researchers attention. 8 José A. Galindo et al. RQ4 Which kind of publications are used to address the challenges? This question gets insights about the maturity of the area by relying on Wieringa et al. [27] taxonomy research which includes validation research, evaluation research, solution proposal, philosophical papers, opinion papers and experience papers. RQ5 When have the papers been published? This question tries to determine the temporary evolution of the publication types and fora. This question aims at helping researchers to determine the current state of a topic, thus, to see if there are chances of further collaboration in that area or not. RQ6 How are the interrelationships among the papers? This question tries to find research gaps whether to invest more research efforts in the future by analysing the already covered areas. Note that most of the questions could be rewritten differently and that the pattern [14] is used for the sake of coverage w.r.t. the research questions. 4.2 Phase 2: Study identification In this section we explain how we identified the studies to include in the mapping study as shown in Figure 6. Please note that in Figure 7 we provide details about the evolution of the selection of papers used in this review: i) conducting search, where we perform a raw search in different data sources for primary studies; ii) filtering studies, that consists in applying inclusion and exclusion criteria. iii) deep search, where new studies are added manually to improve the quality of the primary sources. iv) evaluate search, where we get the final set of primary sources after an evaluation. Next, we detail how we performed each of these steps. Phase 3: Data extraction and classification Phase 1: Planning the review Conducting search Filtering studies Deep search All studies Preliminary primary studies Improved preliminary primary studies Evaluate search Primary studies Phase 2: Study identification Fig. 6: Study identification phase 4.2.1 Conduct search for primary sources The process followed to conduct the search for primary studies is shown in Figure 7 (step 1). We decided to crawl Google Scholar 3and Scopus 4databases because both of them provide easy mechanisms to export the resulting data ordered by number of citations. In each database, we performed two different queries 5: one with the following search string “feature model” AND (“reasoning” OR “analysis” OR “automated” OR “analyses”) and the other with the papers citing [4]. We specified that the publication date should be between 2010 and 2017. We obtained 15,300 (using the search string) and 945 (citing [4]) for Google scholar queries and 1167 (using the search string) and 572 (citing [4]) for Scopus queries. Then, we selected only the first 200 most cited papers from each data source and removed duplicates getting a total of 445 initial studies. Also note that in this paper, we want to determine in which scenarios the AAFM has has been used. It is important to remark that [4] was published in 2010 and that is the reason why we start the search for papers from that date. 3http://scholar.google.com 4http://www.scopus.com 5Note that, as well as the standard process proposed defined by [25] to query bibliographic databases, we added a second group of papers citing the paper that settle the body of knowledge of AAFM [4] as justified in Section 3.1. Automated analysis of feature models: Quo vadis? 9 Step 1: Conducting search Google scholar Search (15300) 01/2010 – 12/2017 by citations Google scholar Citing (945) 01/2010 – 12/2017 by citation Scopus Search (1167) 01/2010 – 12/2017 by citation Scopus Citing(572) 01/2010 – 12/2017 by citation Merged Studies (800) Initial set of studies (445) Step 2: Filtering studies Step 4: Evaluating the search Primary studies First Batch [RQ1] – [RQ2] – [RQ5] (423) Primary studies Second Batch [RQ3] – [RQ4] – [RQ6] (242) Step 3: Snowball reading Google scholar Search (6500) 01/2015 – 12/2017 Google scholar Citting (91) 01/2015 – 12/2017 Scopus Search (316) 01/2015 – 12/2017 Scopus Citting(219) 01/2015 – 12/2017 Manually review relevant studies from 2015-2017 (19) Filtered studies (404) Removed Duplucates. Filtered by title and abstract Selected conferences (VaMoS, SPLC, ASE, ICSE) and the top 10% percentiles according to SciVal (21). To minimize the impact of the ordering by citation reading the first 50 of each result. 200 most cited papers from each data source. Legend Data source Process step Filtering Fig. 7: Search citing publications process 16 José A. Galindo et al. been extensively documented (e.g. [245]). Also, researchers proposed the improvement of the scalability of product configuration [216]. Finally, there are works that focus on configuration, such as Asadi et al.[47], who provide mechanisms to configure feature models while optimizing non-functional properties, and [245], where a tool to guide practitioners while configuring SPLs is presented. Research opportunities Traditionally, configuration technologies are applied in a closed SPL context, this is, all or most of the features are available in a controlled and closed environment, where all the decisions for configuring a product are taken by a very small group of people. An emerging challenge in this area is the configuration of a diversity of distributed SPL descriptions, sometimes known as software ecosystems. As SPLs grow, configuring and maintaining them become an unfeasible task for a small group. Also, different stakeholders and privacy policies encourage the use of visibility restrictions for the configurable parts. We think that there is still work to be done in this area like, for example, to enable the parallel configuration of those open and distributed SPLs. Testing and evolution. When a large number of products is encoded in a FM the testing and evolution processes become expensive and tedious. Two main approaches have been followed to reduce the testing costs. First, combinatorial testing and more concretely T-wise methods to narrow the number of products [130]. Second, test prioritisation to order the execution of critical tests in a time or resource constrained environment [214]. These techniques become also relevant when coping with software evolution. SPL evolution happens when it is required to add, remove or modify features or relationships to an existing SPL. In this scenario, it is required to test if the SPL is error-free before, during and after the evolution actually happens. In this case, the rationale for grouping evolution and testing is that, after the keywording process, some papers were in both sub-areas, probably, because testing is a key aspect to consider when evolving variability-intensive systems. We have identified sixty-two papers referring to SPL testing and evolution. We also noticed that the works coping with testing costs reduction are focusing in scenarios where there is more than one objective –maybe contradictory– to be satisfied at the same time. This is, to optimize different aspects of the same test-suite [130]. For example, to find the test-suite that minimises the testing cost while maximising the market-share. Recently, researchers have been looking for the most convenient evolutionary algorithm to test and select best SPL products, finding that IBEA was returning better results than NSGA-II [216] when coping with multi-objective testing objectives. In terms of evolution management different automated analysis have been proposed to guarantee the safe transition between the different evolution phases of the SPL. For example, White et al. [258], propose the use of CSP solvers to grant the validity of so-called FM drifts. Research opportunities A challenging task is to evolve a SPL while maintaining support for existing products. This is, to verify that the products already being used are still valid in the next evolution of the SPL. SPL researchers use information encoded in variability models when selecting and prioritising testcases. However, other information related to SPL activities such source code management, bug tracking system can be considered to select and prioritise test-cases. The exploration of different testing techniques such as metamorphic, mutation, graph-based among others can still explored in the SPL context. Reverse engineering. There are two main strategies to adopt an SPL approach. First, a proactive approach when a company already starts by planning the construction of the SPL. Second, a reactive approach when individual products are first developed one after another and at a certain point, when the number of similar products is big enough, the company transitions to an SPL engineering approach. To help in this transition, researchers have proposed several reverse-engineering methods. The last years have been fruitful in this area with a total of twenty-two papers classified. Researchers have extracted variability encoded in product comparison matrices[215] and CNF formulas [39]. Moreover, Becan et al. [54] worked on benefiting from ontological knowledge to help in the task of reverse engineering. Automated analysis of feature models: Quo vadis? 17 Research opportunities Nowadays, the variability description of variability-intensive systems is getting more complex. This is done by introducing non-boolean information [145], and by using several variability models [118] in a multi-layer fashion. However, we found no proposals to reverse engineer the existing variability information of variability-intensive systems with more than one variability model or non-boolean information. Also, researchers had not relied on low level assets such as source code or package management systems descriptions when reverse engineering feature models. This kind of assets can provide valuable information because they can contain implicit variability information to be exploited in the reverse engineering task. Multi-model variability analysis. The AAFM started by only considering one FM description at a time. However, the more the SPL becomes larger, the more complex the variability description becomes. Nowadays, it is common to find the variability description of a variability-intensive system in several variability models that can be depicted using different formats. A total of twenty-five papers cope with multi-model SPLs. For example, Dhungana et al. [90], proposed the distributed modelling of SPLs providing examples from the industry. Also, new tools have been provided to determine the set of implementation artifacts that supports a concrete feature specification [29]. But also, because of the need of describing a variability-intensive system in smaller artifacts so it can be maintained by different practitioners, new operations to merge models appeared, thus introducing new FM operations such as model merging or slicing [30]. Research opportunities There is a lack of support for quality attributes when coping with multi-model SPL descriptions. For example, we did not find techniques showing how quality attributes domains are impacted by the selection of certain features at specification level. Also, it is possible to implement distributed analyses of FMs when having multiple models. For example, to first slice the model, and then distribute the analysis in different computation nodes and finally merge the results. This can be helpful in scenarios with large models such as the Linux kernel containing more than 8,000 features. Variability modelling. Encoding the variabilities and commonalities of a SPL requires to find a trade– off between the expressivity of the language and its usability [20]. There are multiple ways of encoding variability such as the one described in [4]. Some variability modelling languages in real context have been analysed [85] to learn how variability is managed and modelled in realistic systems [61]. This area has gained momentum in the last years (see Figure 9), grouping a total of sixty papers. Several proposals appeared to cover different domain specific requirements. For example, new cardinalities regarding features have been considered [175] to enable the proper description of systems when a feature is present multiple times in a configuration. Existing languages for feature modelling have been documented[101] and new language constructs have been introduced to improve the analysis of FMs. For example, FAMILIAR [35] introduced different constructs and translations to solvers that make the AAFM scaling over previous approaches. Research opportunities Nowadays, AAFM is applied beyond product lines in scenarios like cloud computing or the operating systems domain [94] (see the next challenge). We think that the use of new available information in those systems can be used to leverage the AAFM in other domains by means of the definition of new analysis operations or the composition of existing ones. This motivates the need of new constructs encoding this information and thus, improving the AAFM. 18 José A. Galindo et al. Variability intensive systems analysis. The AAFM was initially developed thinking in SPLs. However, nowadays it is used for a widespread number of applications. Currently, the scenarios where AAFM is used range from variability analysis in other domains such as mobile apps testing [117], cloud systems [119] or bioinformatics [84]. Note that there is a clear trend of using AAFM for systems beyond SPL however, authors tend to use it for a concrete scenario such as testing [117] or configuration [119]. A total of twenty-three papers referring to this trend were identified within the corpus of papers. Different variability models benefit from automated analysis techniques to extract information in a similar way. While diverse models such as OVM [209] and DOPLER [170] models had automated analysis support in 2010, now, there are other models such as BPMNs [87], that use similar techniques to extract information and metrics. In the open source domain, the AAFM was applied to existing systems such as the Linux [232] kernel. Also, the AAFM has profited from the open-source community which helped to understand the nature of models and the constraints flavours existing in the wild [180]. Research opportunities The number of different scenarios where AAFM can be applied have been proved to be very large and we envision that there will be more and more scenarios where it will be applied in the future. Ecosystems, cyber-physical systems, robotics, big data or the internet of the things are only some examples where variability is a first-class citizen and can use AAFM techniques in their development. 5.3.1 Quantitative analysis The variability context dimension of this mapping study aims to identify the current trends in the area of AAFM. In the previous Section, we already presented them while highlighting some of the most interesting findings. Now we present quantitative data that provides the weight of each trend in the area. Table 8 presents the distribution of the different trends of the full set of papers. We see that the areas that more attracted researchers attention were testing and evolution, product configuration and variability modelling. When we take a look to the list of papers we see that the product configuration was more present within the time lapse considered for this research while the use of AAFM in a diversity of systems is more recent. Also, we see that, because of this new AAFM usage the variability, modelling papers started to introduce new forms of variability in the reasoning process such as features cardinalities. Variability context Citations #papers Product configuration and derivation [35, 37, 42, 46, 47, 49, 50, 53, 55, 56, 57, 58, 59, 60, 65, 66, 68, 72, 74, 75, 76, 77, 80, 82, 83, 89, 93, 95, 98, 99, 100, 103, 105, 106, 107, 108, 109, 110, 112, 114, 116, 118, 119, 120, 123, 125, 126, 128, 133, 134, 135, 136, 138, 140, 144, 145, 146, 149, 153, 154, 155, 156, 161, 169, 171, 172, 173, 174, 176, 177, 178, 182, 183, 185, 189, 190, 191, 194, 195, 197, 198, 199, 201, 204, 205, 206, 208, 209, 210, 216, 219, 221, 222, 224, 234, 236, 237, 238, 242, 244, 245, 246, 247, 249, 250, 251, 252, 253, 259, 260, 263, 264, 265, 266, 267] 115 Testing and evolution [28, 33, 37, 38, 41, 43, 44, 45, 48, 52, 64, 67, 69, 72, 73, 83, 91, 92, 97, 104, 105, 107, 111, 113, 114, 115, 117, 121, 122, 124, 128, 129, 130, 132, 134, 135, 139, 141, 143, 144, 150, 152, 159, 160, 161, 162, 164, 165, 166, 168, 170, 176, 183, 186, 187, 188, 192, 193, 195, 196, 200, 207, 212, 213, 214, 222, 225, 226, 227, 228, 229, 230, 235, 239, 241, 244, 248, 250, 253, 254, 255, 256, 257, 258, 261] 85 Reverse engineering [31, 34, 36, 39, 48, 54, 56, 70, 86, 126, 127, 131, 134, 144, 148, 149, 151, 157, 163, 164, 180, 181, 191, 215, 233, 244, 268] 27 Multi-model variability analysis [29, 30, 32, 33, 34, 50, 65, 71, 73, 81, 89, 90, 100, 113, 118, 121, 134, 144, 158, 179, 182, 206, 210, 220, 222, 237, 241, 246, 262] 29 Variability modelling [40, 47, 49, 51, 54, 55, 56, 58, 61, 62, 63, 64, 70, 71, 78, 79, 80, 81, 84, 85, 88, 91, 94, 101, 102, 108, 109, 110, 118, 134, 136, 137, 139, 141, 142, 144, 145, 146, 147, 149, 157, 158, 167, 169, 175, 179, 183, 184, 194, 197, 199, 202, 203, 210, 211, 217, 218, 219, 221, 223, 224, 225, 231, 232, 240, 242, 243, 244, 252, 261, 262, 265, 266, 269] 74 Variability intensive systems analysis [28, 46, 55, 56, 57, 65, 68, 73, 84, 86, 87, 91, 94, 119, 120, 134, 153, 156, 167, 178, 190, 204, 213, 214, 232, 244, 259, 267, 269] 29 Table 8: Classification of papers based on the variability context facet Automated analysis of feature models: Quo vadis? 19 ● ● ● ● ● ● ● ● 0% 20% 40% 60% 2010 2011 2012 2013 2014 2015 2016 2017 Year of publication # Variability context facet papers/ # year papers Variability context ●Product configuration and derivation Multi−model variability analysis Variability modeling Reverse engineering Testing and evolution Variability−intensive systems analysis Fig. 9: Temporary distribution of the variability context facet Figure 9 shows the evolution of the different trends depending on the year of publication. In the y axis, the percentage of papers per year of each variability context facet is presented which can show the interest of the facet in the corresponding year. We see that there are some areas that stayed with minor variations, such as multi-model variability analysis and variability modelling, while there are others that clearly have hyped during the last years, such as reverse engineering. We want to highlight that “Product configuration and derivation” has a good percentage of papers over the time span of the review. This is probably due to the fact that it is one of the initial and more established usages of AAFM. Also, the contributions targeting “Multi-model variability analysis” had an stable percentage of papers every year, and has not hyped yet. We conjecture that this is due to the fact of the complexity of the underling analysis on multiple models and the lack of well recognized case studies and examples on this field. Finally, another observation is that before 2014 there were two groups of trends that evolved similarly. First, “Multi-model variability analysis”,“Reverse engineering” and “Variability intensive systems analysis”; and second, “Product configuration and derivation”,“Testing and evolution” and “Variability modelling”. In the last years, we see the hype of “Variability modelling”. We conjecture that the hype of this last trend is due to industrial needs of specifying specific variability properties within their domains [2]. 5.4 RQ4: Which kind of publications are used to address the challenges? The research facet classifies the papers depending on its research type. In Table 9 we observe the number of papers within our set of papers that fits each research type. We observe that there is a clear higher amount of papers on the evaluation, philosophical and solution proposal papers categories. This indicates that even this area is maturing and we can foresee some papers focusing on real-world problems, most papers are not related to the industry. We notice that most of the research lacks validation and focuses on evaluating concrete theoretical approaches. This pinpoints the need for this community to work closer to the industry and to validate methods and techniques in a more practical way. 20 José A. Galindo et al. Research facet Citations #papers Opinion Paper [66, 73, 75, 144, 164, 193, 248] 7 Philosophical Paper [51, 74, 99, 101, 134, 137, 171, 175, 196, 201, 244] 11 Solution Proposal [30, 35, 44, 58, 60, 65, 72, 78, 79, 80, 81, 93, 102, 106, 140, 141, 142, 145, 151, 158, 162, 165, 172, 176, 184, 185, 191, 195, 197, 198, 199, 200, 203, 205, 207, 210, 217, 219, 220, 225, 236, 239, 243, 246, 247, 262, 264, 265, 268] 49 Evaluation Research [28, 29, 32, 33, 36, 37, 38, 39, 40, 41, 42, 43, 45, 46, 47, 48, 49, 50, 52, 54, 55, 56, 59, 67, 68, 69, 70, 71, 77, 82, 83, 85, 86, 87, 88, 89, 90, 92, 94, 96, 97, 98, 100, 103, 104, 107, 109, 110, 111, 112, 113, 114, 115, 116, 117, 118, 119, 120, 121, 122, 123, 124, 125, 126, 127, 128, 129, 130, 131, 132, 133, 135, 138, 139, 143, 146, 147, 148, 149, 150, 152, 153, 154, 155, 157, 159, 160, 161, 163, 166, 167, 168, 169, 170, 173, 174, 177, 178, 179, 180, 181, 182, 183, 186, 188, 189, 190, 192, 202, 204, 206, 208, 209, 212, 214, 215, 216, 218, 222, 223, 224, 226, 227, 228, 229, 230, 231, 233, 234, 235, 237, 238, 240, 241, 245, 249, 250, 252, 255, 256, 257, 258, 260, 261, 263, 266, 267, 269] 148 Validation Research [31, 34, 57, 61, 62, 63, 84, 91, 105, 136, 156, 187, 194, 211, 213, 221, 232, 251, 253, 254, 259] 21 Experience Report [53, 64, 76, 95, 108, 242] 6 Table 9: Classification of papers based on the research facet 5.5 RQ5: When have the papers been published? In this question, we want to determine the dates where the papers were published. Also, we want to see how the number of papers published in the different fora types used across the paper evolved across the different years. To answer this question we relied on the first batch of papers containing 423 contributions. Note that the papers are normalized by the number primary studies published that year. 0.38 0.51 0.11 0.36 0.45 0.19 0.23 0.47 0.3 0.3 0.51 0.19 0.22 0.44 0.34 0.19 0.53 0.29 0.08 0.52 0.4 0.13 0.4 0.47 0.0 0.2 0.4 2010 2011 2012 2013 2014 2015 2016 2017 Year of publication # papers of a kind / # papers of a year Type of paper Journal Non Variability Related Conference Variability Related Conference Fig. 10: Temporary distribution of papers. We observe in Figure 10 that there is an increment of papers published in journals. This is specially remarkable in the years 2017 were the papers published in journals were more than in conferences and Automated analysis of feature models: Quo vadis? 21 workshops. This points out that AAFM research works have matured when published. Also, the number of papers published in SPLC, VAMOS, ICSE, and ASE is lower than the number of papers published in the rest of conferences when taking into account the publication year. The difference between variability related conferences and the non related ones stays mostly unaltered in the last three years covered by this study. However, we can see that the percentage of variability related conferences paper is decreasing since 2013. We conjecture that this change in the trend is due to the fact that the AAFM is now more widespread in other communities and therefore, non variability related conferences accept more papers on this topic. 5.6 RQ6: How are the papers related among them? This question aims at finding research gaps to drive future work in the area of AAFM. By identifying the variability-context and type of papers where there are fewer publications we can infer the lacks of current research. To answer this question we relied on the reduced dataset only considering variability-related conferences and journals. It is interesting to see the global distribution (see Figure 11) of the number of papers based on the two main dimensions considered in this mapping study. This heatmap shows the research gaps where the SPL community should invest more efforts. 0 0 5 1 3 0 2 1 6 3 6 2 1 2 3 5 1 2 1 8 26 11 21 3 18 17 66 59 33 18 7 1 9 6 10 2 Variability−intensive systems analysis Multi−model variability analysis Product configuration and derivation Testing and evolution Variability modeling Reverse engineering Philosophical Paper Opinion Paper Solution Proposal Evaluation Research Validation Research Experience Report Research facet Variability context facet 0 20 40 60 # papers Fig. 11: Visualization of the systematic map Firstly, we observe that there is still room for improvement regarding experience reports because only eleven contributions fall in this category. Moreover, we almost found papers about product configuration and derivation and variability modelling in such research facet. This is a clear gap that evidences the distance to industry of the other trends. Secondly, we observe that opinion and philosophical papers have low weight which may show that the discipline is getting mature because there are taxonomies, literature reviews or mapping studies on the different trends. Finally, we see that most of the papers are on the evaluation research category and testing and evolution context facet, i.e. the papers evaluate the solution in front of small examples or datasets. We conjecture that this is due to the fact of the availability of well known examples and datasets on these trends. Generally speaking, we observe that the area is getting mature but, still misses some more collaboration with industry. In fact, those variability contexts that lately attracted the most researchers attention did not have time to report on experience reports. Also, we observe a tendency of publishing conceptual ideas that are later evaluated in terms of efficiency (evaluation) instead of customer satisfaction (validation). 22 José A. Galindo et al. 6 Conclusions In this paper we went through different research questions to understand the current state of the area of AAFM by applying a systematic mapping method. We identified where relevant papers are being published; who are the authors and institutions that currently holds the know-how of AAFM; which trends in the usage of AAFM attracted more researchers attention in the last years; what kind of publications are being proposed; when and where the papers were published. The surveyed primary sources point out that the AAFM is a subject that is getting mature and that has driven some other subjects such as product configuration, derivation, testing, evolution and reverse engineering. However, regarding the distribution depending on the nature of primary sources, the main fact detected is that FM practitioners are not validating the research as much as they evaluate it. This actually should encourage the community to work closer to industry and provide better ready-to-the- market solutions instead of toy techniques and prototypes. The aim of this study is to guide future research on the application of the AAFM in new domains such as images creation, cloud management, mobile computing, operating systems dependencies or internet of the things among others. We believe that it is time to stop gleaming luster to the analysis techniques that are in general mature enough and find application domains and real evidence where the AAFM can be successfully applied. Acknowledgements This work was supported, in part, by the European Commission (FEDER), by the Spanish government under BELi (TIN2015-70560-R) project and by the Andalusian government under the COPAS (TIC-1867) project. You can find all the material used in this paper in the website https://isa-group.github.io/ aafm-quo-vadis/. References 1. Acher M, Collet P, Lahire P, France RB (2013) FAMILIAR: A domain-specific language for large scale management of feature models. Science of Computer Programming (SCP) 78(6):657–681 2. Alférez M, Acher M, Galindo JA, Baudry B, Benavides D (2018) Modeling variability in the video domain: language and experience report. Software Quality Journal DOI 10.1007/s11219-017-9400-8 3. Batory D, Benavides D, Ruiz-Cortes A (2006) Automated analysis of feature models: Challenges ahead. Commun ACM 49(12):45–47, DOI 10.1145/1183236.1183264 4. Benavides D, Segura S, Ruiz-Cortés A (2010) Automated analysis of feature models 20 years later. Information Systems 35(6):615–636 5. Benavides D, Trinidad P, Cortés AR, Segura S (2013) FaMa, Springer Berlin Heidelberg, chap FaMa, pp 163–171. DOI 10.1007/978-3-642-36583-6-11 6. Capilla R (2013) Variability realization techniques and product derivation. In: Systems and Software Variability Management, Springer, pp 87–99 7. Clements P, Northrop L (2002) Software product lines. Addison-Wesley Boston 8. Durán A, Benavides D, Segura S, Trinidad P, Ruiz-Cortés A (in press) FLAME: a formal framework for the automated analysis of software product lines validated by automated specification testing. Software & Systems Modeling pp 1–34, DOI 10.1007/s10270-015-0503-z 9. Engström E, Runeson P (2011) Software product line testing – a systematic mapping study. Information and Software Technology 53(1):2 – 13, DOI http://dx.doi.org/10.1016/j.infsof.2010.05.011 10. Galindo J, Turner H, Benavides D, White J (2014) Testing variability-intensive systems using automated analysis: an application to android. Software Quality Journal pp 1–41, DOI 10.1007/ s11219-014-9258-y 11. Grant MJ, Booth A (2009) A typology of reviews: an analysis of 14 review types and associated methodologies. Health Information & Libraries Journal 26(2):91–108 12. Heradio R, Perez-Morago H, Fernandez-Amoros D, Cabrerizo FJ, Herrera-Viedma E (2015) A science mapping analysis of the literature on software product lines. In: Fujita H, Guizzi G (eds) Intelligent Software Methodologies, Tools and Techniques, Communications in Computer and Information Science, vol 532, Springer International Publishing, pp 242–251, DOI 10.1007/978-3-319-22689-718 Automated analysis of feature models: Quo vadis? 23 13. Heradio R, Perez-Morago H, Fernandez-Amoros D, Cabrerizo FJ, Herrera-Viedma E (2016) A bibliometric analysis of 20 years of research on software product lines. Information and Software Technology 72:1 – 15, DOI http://dx.doi.org/10.1016/j.infsof.2015.11.004 14. Jia C, Cai Y, Yu YT, Tse TH (2016) 5W+1H pattern: A perspective of systematic mapping studies and a case study on cloud software testing. Journal of Systems and Software 116:206–219, DOI 10.1016/j.jss.2015.01.058 15. Kang KC, Cohen SG, Hess JA, Novak WE, Peterson AS (1990) Feature-oriented domain analysis (foda) feasibility study. Tech. rep., DTIC Document 16. Kipling R (1902) Just so stories. MacMillan, London, UK 17. Kitchenham B, Brereton OP, Budgen D, Turner M, Bailey J, Linkman S (2009) Systematic literature reviews in software engineering – a systematic literature review. Information and Software Technology 51(1):7 – 15, DOI http://dx.doi.org/10.1016/j.infsof.2008.09.009, special Section - Most Cited Articles in 2002 and Regular Research Papers 18. Laguna MA, Crespo Y (2013) A systematic mapping study on software product line evolution: From legacy system reengineering to product line refactoring. Science of Computer Programming 78(8):1010 – 1034, DOI http://dx.doi.org/10.1016/j.scico.2012.05.003 19. Lopez-Herrejon RE, Linsbauer L, Egyed A (2015) A systematic mapping study of search-based software engineering for software product lines. Information & Software Technology 61:33–51, DOI 10.1016/j.infsof.2015.01.008 20. Méndez-Acuña D, Galindo JA, Degueule T, Combemale B, Baudry B (2016) Leveraging software product lines engineering in the development of external dsls: A systematic literature review. Computer Languages, Systems & Structures 46:206–235, DOI 10.1016/j.cl.2016.09.004 21. Mendonca M, Branco M, Cowan D (2009) S.p.l.o.t.: Software product lines online tools. In: Proceedings of the 24th ACM SIGPLAN Conference Companion on Object Oriented Programming Systems Languages and Applications, ACM, New York, NY, USA, OOPSLA ’09, pp 761–762, DOI 10.1145/1639950.1640002 22. Montalvillo L, Díaz O (2016) Requirement-driven evolution in software product lines: A systematic mapping study. Journal of Systems and Software 122:110 – 143, DOI http://dx.doi.org/10.1016/j.jss. 2016.08.053 23. da Mota Silveira Neto PA, do Carmo Machado I, McGregor JD, de Almeida ES, de Lemos Meira SR (2011) A systematic mapping study of software product lines testing. Information and Software Technology 53(5):407 – 423, DOI http://dx.doi.org/10.1016/j.infsof.2010.12.003, special Section on Best Papers from {XP2010} 24. Petersen K, Feldt R, Mujtaba S, Mattsson M (2008) Systematic mapping studies in software engineering. In: Proceedings of the 12th International Conference on Evaluation and Assessment in Software Engineering, British Computer Society, Swinton, UK, UK, EASE’08, pp 68–77 25. Petersen K, Feldt R, Mujtaba S, Mattsson M (2008) Systematic mapping studies in software engineering. In: Proceedings of the 12th International Conference on Evaluation and Assessment in Software Engineering, BCS Learning & Development Ltd., Swindon, UK, EASE’08, pp 68–77 26. Schobbens P, Heymans P, Trigaux J, Bontemps Y (2007) Generic semantics of feature diagrams. Computer Networks 51(2):456–479, DOI 10.1016/j.comnet.2006.08.008 27. Wieringa R, Maiden N, Mead N, Rolland C (2006) Requirements engineering paper classification and evaluation criteria: a proposal and a discussion. Requirements Engineering 11(1):102–107 24 José A. Galindo et al. Appendix A. Primary sources 28. Abal I, Brabrand C, Wasowski A (2014) 42 variability bugs in the linux kernel: a qualitative analysis. In: ASE, DOI 10.1145/2642937.2642990 29. Abele A, Papadopoulos Y, Servat D, Törngren M, Weber M (2010) The cvm framework-a prototype tool for compositional variability management. In: VAMOS, vol 10, pp 101–105 30. Acher M, Collet P, Lahire P, France R (2011) Slicing feature models. In: ASE, pp 424–427, DOI 10.1109/ASE.2011.6100089 31. Acher M, Cleve A, Perrouin G, Heymans P, Vanbeneden C, Collet P, Lahire P (2012) On extracting feature models from product descriptions. In: VAMOS, pp 45–54, DOI 10.1145/2110147.2110153 32. Acher M, Collet P, Gaignard A, Lahire P, Montagnat J, France R (2012) Composing multiple variability artifacts to assemble coherent workflows. SQJ 20(3-4):689–734, DOI 10.1007/ s11219-011-9170-7 33. Acher M, Collet P, Lahire P, France R (2012) Separation of concerns in feature modeling: Support and applications. In: AOSD, pp 1–12, DOI 10.1145/2162049.2162051 34. Acher M, Baudry B, Heymans P, Cleve A, Hainaut JL (2013) Support for reverse engineering and maintaining feature models. In: VAMOS, ACM, p 20, DOI 10.1145/2430502.2430530 35. Acher M, Collet P, Lahire P, France RB (2013) Familiar: A domain-specific language for large scale management of feature models. SCP 78(6):657–681, DOI 10.1016/j.scico.2012.12.004 36. Acher M, Cleve A, Collet P, Merle P, Duchien L, Lahire P (2014) Extraction and evolution of architectural variability models in plugin-based systems. SOSYM 13(4):1367–1394, DOI 10.1007/ s10270-013-0364-2 37. Ajoudanian S, Hosseinabadi SH (2015) Automatic promotional specialization, generalization and analysis of extended feature models with cardinalities in alloy. LAMP 84(5):640–667, DOI 10.1016/ j.jlamp.2014.11.005 38. Al-Hajjaji M, Thüm T, Meinicke J, Lochau M, Saake G (2014) Similarity-based prioritization in software product-line testing. In: SPLC, ACM, pp 197–206, DOI 10.1145/2648511.2648532 39. Andersen N, Czarnecki K, She S, Wa¸sowski A (2012) Efficient synthesis of feature models. In: SPLC, vol 1, pp 106–115, DOI 10.1145/2362536.2362553 40. Antkiewicz M, B¸ak K, Murashkin A, Olaechea R (2013) Clafer tools for product line engineering. In: SPLC, DOI 10.1145/2499777.2499779 41. Apel S, Speidel H, Wendler P, Von Rhein A, Beyer D (2011) Detection of feature interactions using feature-aware verification. In: ASE, IEEE Computer Society, pp 372–375, DOI 10.1109/ase.2011. 6100075 42. Apel S, Von Rhein A, ThüM T, KäStner C (2013) Feature-interaction detection based on featurebased specifications. CNJ 57(12):2399–2409, DOI 10.1016/j.comnet.2013.02.025 43. Arcaini P, Gargantini A, Vavassori P (2015) Generating tests for detecting faults in feature models. In: ICST, DOI 10.1109/ICST.2015.7102591 44. Arcaini P, Gargantini A, Vavassori P (2017) Automated repairing of variability models. In: SPLC, vol 1, pp 9–18, DOI 10.1145/3106195.3106206 45. Arrieta A, Sagardui G, Etxeberria L, Zander J (2017) Automatic generation of test system instances for configurable cyber-physical systems. SQJ 25(3):1041–1083, DOI 10.1007/s11219-016-9341-7 46. Asadi M, Mohabbati B, Gröner G, Gasevic D (2014) Development and validation of customized process models. JSS 96:73–92, DOI 10.1016/j.jss.2014.05.063 47. Asadi M, Soltani S, Gasevic D, Hatala M, Bagheri E (2014) Toward automated feature model configuration with optimizing non-functional requirements. IST 56(9):1144–1165, DOI 10.1016/j. infsof.2014.03.005 48. Assunção W, Lopez-Herrejon R, Linsbauer L, Vergilio S, Egyed A (2017) Multi-objective reverse engineering of variability-safe feature models based on code dependencies of system variants. ESE 22(4):1763–1794, DOI 10.1007/s10664-016-9462-4 49. Bagheri E, Asadi M, Gasevic D, Soltani S (2010) Stratified analytic hierarchy process: Prioritization and selection of software features. In: SPLC, vol 6287 LNCS, pp 300–315, DOI 10.1007/978-3-642-15579-6_21 50. Bagheri E, Di Noia T, Ragone A, Gasevic D (2010) Configuring software product line feature models based on stakeholders’ soft and hard requirements. In: SPLC, Springer, pp 16–31, DOI 10.1007/978-3-642-15579-6_2 Automated analysis of feature models: Quo vadis? 25 51. Bagheri E, Noia TD, Gasevic D, Ragone A (2012) Formalizing interactive staged feature model configuration. JSEP 24(4):375–400, DOI 10.1002/smr.534 52. Bagheri Eb, Gasevic D (2011) Assessing the maintainability of software product line feature models using structural metrics. SQJ 19(3):579–612, DOI 10.1007/s11219-010-9127-2 53. Baresi L, Guinea S, Pasquale L (2012) Service-oriented dynamic software product lines. Computer 45(10):42–48, DOI 10.1109/MC.2012.289 54. Bécan G, Acher M, Baudry B, Nasr SB (2015) Breathing ontological knowledge into feature model synthesis: an empirical study. ESE pp 1–48, DOI 10.1007/s10664-014-9357-1 55. ter Beek M, Legay A, Lafuente A, Vandin A (2016) Statistical model checking for product lines. In: ISOLA, vol 9952 LNCS, pp 114–133, DOI 10.1007/978-3-319-47166-2_8 56. ter Beek M, Reniers M, de Vink E (2016) Supervisory controller synthesis for product lines using cif 3. In: ISOLA, vol 9952 LNCS, pp 856–873, DOI 10.1007/978-3-319-47166-2_59 57. ter Beek MH, Legay A, Lafuente AL, Vandin A (2015) Statistical analysis of probabilistic models of software product lines with quantitative constraints. In: SPLC, ACM, pp 11–15, DOI 10.1145/ 2791060.2791087 58. Beek MHt, De Vink EP (2014) Using mcrl2 for the analysis of software product lines. In: FormaliSE, ACM, pp 31–37, DOI 10.1145/2593489.2593493 59. Benavides D, Felfernig A, Galindo J, Reinfrank F (2013) Automated analysis in feature modelling and product configuration. In: ICSR, vol 7925 LNCS, pp 160–175, DOI 10.1007/978-3-642-38977-1_ 11 60. Berger T, She S, Lotufo R, Czarnecki K, Wasowski A (2010) Feature-to-code mapping in two large product lines. In: SPLC, Citeseer, pp 498–499, DOI 10.1007/978-3-642-15579-6_48 61. Berger T, She S, Lotufo R, Wa¸sowski A, Czarnecki K (2010) Variability modeling in the real: A perspective from the operating systems domain. In: ASE, pp 73–82, DOI 10.1145/1858996.1859010 62. Berger T, She S, Lotufo R, Wasowski A, Czarnecki K (2013) A study of variability models and languages in the systems software domain. TSE 39(12):1611–1640, DOI 10.1109/TSE.2013.34 63. Berger T, Lettner D, Rubin J, Grünbacher P, Silva A, Becker M, Chechik M, Czarnecki K (2015) What is a feature? a qualitative study of features in industrial software product lines. In: SPLC, vol 20-24-July-2015, pp 16–25, DOI 10.1145/2791060.2791108 64. Bezerra C, Andrade R, Monteiro J (2017) Exploring quality measures for the evaluation of feature models: a case study. JSS 131:366–385, DOI 10.1016/j.jss.2016.07.040 65. Boškovi M, Bagheri E, GaŠevi D, Mohabbati B, Kaviani N, Hatala M (2010) Automated staged configuration with semantic web technologies. IJSEKE 20(4):459–484, DOI 10.1142/S0218194010004827 66. Boucher Q, Perrouin G, Heymans Pb (2012) Deriving configuration interfaces from feature models: A vision paper. In: VAMOS, pp 37–44, DOI 10.1145/2110147.2110152 67. Brabrand C, Ribeiro M, Tolêdo T, Borba P (2012) Intraprocedural dataflow analysis for software product lines. In: AOSD, ACM, pp 13–24, DOI 10.1145/2162049.2162052 68. Buchmann T, Dotor A, Westfechtel B (2013) Mod2-scm: A model-driven product line for software configuration management systems. IST 55(3):630–650, DOI 10.1016/j.infsof.2012.07.010 69. Bürdek J, Kehrer T, Lochau M, Reuling D, Kelter U, Schürr A (2016) Reasoning about product-line evolution using complex feature model differences. ASEJ 23(4):687–733, DOI 10.1007/ s10515-015-0185-3 70. Bécan G, Behjati R, Gotlieb A, Acher M (2015) Synthesis of attributed feature models from product descriptions. In: SPLC, vol 20-24-July-2015, pp 1–10, DOI 10.1145/2791060.2791068 71. Bąk K, Diskin Z, Antkiewicz M, Czarnecki K, Wąsowski A (2016) Clafer: unifying class and feature modeling. SOSYM 15(3):811–845, DOI 10.1007/s10270-014-0441-1 72. Camacho C, Llana L, Núñez A (2016) Cost-related interface for software product lines. LAMP 85(1):227–244, DOI 10.1016/j.jlamp.2015.09.009 73. Capilla R, Bosch J (2011) The promise and challenge of runtime variability. Computer 44(12):93–95, DOI 10.1109/MC.2011.382 74. Capilla R, Bosch J, Trinidad P, Ruiz-Cortés A, Hinchey M (2014) An overview of dynamic software product line architectures and techniques: Observations from research and industry. JSS 91(1):3–23, DOI 10.1016/j.jss.2013.12.038 75. Capilla R, Ortiz Ó, Hinchey M (2014) Context variability for context-aware systems. Computer 47(2):85–87, DOI 10.1109/mc.2014.33 32 José A. Galindo et al. 221. Schroeter J, Mucha P, Muth M, Jugel K, Lochau M (2012) Dynamic configuration management of cloud-based applications. In: SPLC, ACM, pp 171–178, DOI 10.1145/2364412.2364441 222. Schröter R, Thüm T, Siegmund N, Saake G (2013) Automated analysis of dependent feature models. In: VAMOS, ACM, p 9, DOI 10.1145/2430502.2430515 223. Schröter R, Siegmund N, Thüm T, Saake G (2014) Feature-context interfaces: tailored programming interfaces for software product lines. In: SPLC, ACM, pp 102–111, DOI 10.1145/2648511.2648522 224. Schröter R, Krieter S, Thüm T, Benduhn F, Saake G (2016) Feature-model interfaces: The highway to compositional analyses of highly-configurable systems. In: ICSE, IEEE Computer Society, pp 667–678, DOI 10.1145/2884781.2884823 225. Schubanz M, Pleuss A, Botterweck G, Lewerentz C (2012) Modeling rationale over time to support product line evolution planning. In: VAMOS, pp 193–199, DOI 10.1145/2110147.2110169 226. Segura S, Benavides D, Ruiz-Cortés A (2011) Functional testing of feature model analysis tools: a test suite. IET 5(1):70–82, DOI 10.1049/iet-sen.2009.0096 227. Segura S, Hierons R, Benavides D, Ruiz-Cortés A (2011) Automated metamorphic testing on the analyses of feature models. IST 53(3):245–258, DOI 10.1016/j.infsof.2010.11.002 228. Segura S, Galindo JA, Benavides D, Parejo JA, Ruiz-Cortés A (2012) Betty: benchmarking and testing on the automated analysis of feature models. In: VAMOS, ACM, pp 63–71, DOI 10.1145/ 2110147.2110155 229. Segura S, Parejo J, Hierons R, Benavides D, Ruiz-Cortés A (2014) Automated generation of computationally hard feature models using evolutionary algorithms. ESA 41(8):3975–3992, DOI 10.1016/j.eswa.2013.12.028 230. Segura S, Durán A, Sánchez A (2015) Automated metamorphic testing of variability analysis tools. STVR DOI 10.1002/stvr.1566 231. Seidl C, Schaefer I, Assmann U (2014) Capturing variability in space and time with hyper feature models. In: VAMOS, ACM, p 6, DOI 10.1145/2556624.2556625 232. She S, Lotufo R, Berger T, Wasowski A, Czarnecki K (2010) The variability model of the linux kernel. In: VAMOS, vol 10, pp 45–51 233. She S, Lotufo R, Berger T, Wa¸sowski A, Czarnecki K (2011) Reverse engineering feature models. In: ICSE, pp 461–470, DOI 10.1145/1985793.1985856 234. Siegmund N, Kolesnikov SS, Kästner C, Apel S, Batory D, Rosenmüller M, Saake G (2012) Predicting performance via automated feature-interaction detection. In: ICSE, IEEE Press, pp 167–177, DOI 10.1109/icse.2012.6227196 235. Soltani S, Asadi M, Hatala M, Gašević D, Bagheri E (2011) Automated planning for feature model configuration based on stakeholders’ business concerns. In: ASE, pp 536–539, DOI 10.1109/ASE. 2011.6100118 236. Soltani S, Asadi M, Gašević D, Hatala M, Bagheri E (2012) Automated planning for feature model configuration based on functional and non-functional requirements. In: SPLC, ACM, pp 56–65, DOI 10.1145/2362536.2362548 237. Stein J, Nunes I, Cirilo E (2014) Preference-based feature model configuration with multiple stakeholders. In: SPLC, ACM, pp 132–141, DOI 10.1145/2648511.2648525 238. Strickler A, Prado Lima J, Vergilio S, Pozo A (2016) Deriving products for variability test of feature models with a hyper-heuristic approach. ASCJ 49:1232–1242, DOI 10.1016/j.asoc.2016.07.059 239. Tanhaei M, Habibi J, Mirian-Hosseinabadi SH (2016) Automating feature model refactoring: A model transformation approach. IST 80:138–157, DOI 10.1016/j.infsof.2016.08.011 240. Tawhid R, Petriu D (2011) Automatic derivation of a product performance model from a software product line model. In: SPLC, pp 80–89, DOI 10.1109/SPLC.2011.27 241. Teixeira L, Borba P, Gheyi R (2013) Safe composition of configuration knowledge-based software product lines. JSS 86(4):1038–1053, DOI 10.1016/j.jss.2012.11.006 242. Ter Beek M, Fantechi A, Gnesi S (2015) Applying the product lines paradigm to the quantitative analysis of collective adaptive systems. In: SPLC, vol 20-24-July-2015, pp 321–326, DOI 10.1145/ 2791060.2791100 243. Thüm T, Kästner C, Erdweg S, Siegmund N (2011) Abstract features in feature modeling. In: SPLC, pp 191–200, DOI 10.1109/SPLC.2011.53 244. Thüm T, Apel S, Kästner C, Schaefer I, Saake G (2014) A classification and survey of analysis strategies for software product lines. ACMCS 47(1), DOI 10.1145/2580950 Automated analysis of feature models: Quo vadis? 33 245. Thüm T, Kästner C, Benduhn F, Meinicke J, Saake G, Leich T (2014) Featureide: An extensible framework for feature-oriented software development. SCP 79:70–85, DOI 10.1016/j.scico.2012.06. 002 246. Tërnava X, Collet P (2017) Early consistency checking between specification and implementation variabilities. In: SPLC, vol 1, pp 29–38, DOI 10.1145/3106195.3106209 247. Vierhauser M, Grünbacher P, Egyed A, Rabiser R, Heider W (2010) Flexible and scalable consistency checking on product line variability models. In: ASE, ACM, pp 63–72, DOI 10.1145/1858996.1859009 248. Vogel-Heuser B, Fay A, Schaefer I, Tichy M (2015) Evolution of software in automated production systems: Challenges and research directions. JSS 110:54–84, DOI 10.1016/j.jss.2015.08.026 249. Von Rhein A, Apel S, Kästner C, Thüm T, Schaefer I (2013) The pla model: on the combination of product-line analyses. In: VAMOS, ACM, p 14, DOI 10.1145/2430502.2430522 250. Von Rhein A, Grebhahn A, Apel S, Siegmund N, Beyer D, Berger T (2015) Presence-condition simplification in highly configurable systems. In: ICSE, vol 1, pp 178–188, DOI 10.1109/ICSE.2015. 39 251. Walter R, Felfernig A, Küchlin W (2017) Constraint-based and sat-based diagnosis of automotive configuration problems. JIIS 49(1):87–118, DOI 10.1007/s10844-016-0422-7 252. Wang S, Ali S, Gotlieb A (2013) Minimizing test suites in software product lines using weight-based genetic algorithms. In: GECCO, ACM, pp 1493–1500, DOI 10.1145/2463372.2463545 253. Wang S, Buchmann D, Ali S, Gotlieb A, Pradhan D, Liaaen M (2014) Multi-objective test prioritization in software product line testing: an industrial case study. In: SPLC, ACM, pp 32–41, DOI 10.1145/2648511.2648515 254. Wang S, Ali S, Gotlieb A, Liaaen M (2016) A systematic test case selection methodology for product lines: results and insights from an industrial case study. ESE 21(4):1586–1622, DOI 10.1007/s10664-014-9345-5 255. Wang S, Ali S, Gotlieb A, Liaaen M (2017) Automated product line test case selection: industrial case study and controlled experiment. SOSYM 16(2):417–441, DOI 10.1007/s10270-015-0462-4 256. Wang Sb, Ali S, Gotlieb A (2015) Cost-effective test suite minimization in product lines using search techniques. JSS 103:370–391, DOI 10.1016/j.jss.2014.08.024 257. White J, Benavides D, Schmidt D, Trinidad P, Dougherty B, Ruiz-Cortes A (2010) Automated diagnosis of feature model configurations. JSS 83(7):1094–1107, DOI 10.1016/j.jss.2010.02.017 258. White J, Galindo J, Saxena T, Dougherty B, Benavides D, Schmidt D (2014) Evolving feature model configurations in software product lines. JSS 87(1):119–136, DOI 10.1016/j.jss.2013.10.010 259. Wittern E, Zirpins C (2016) Service feature modeling: modeling and participatory ranking of service design alternatives. SOSYM 15(2):553–578, DOI 10.1007/s10270-014-0414-4 260. Xue Y, Zhong J, Tan T, Liu Y, Cai W, Chen M, Sun J (2016) Ibed: Combining ibea and de for optimal feature selection in software product line engineering. ASCJ 49:1215–1231, DOI 10.1016/j. asoc.2016.07.040 261. Yu W, Zhang W, Zhao H, Jin Z (2014) Tdl: a transformation description language from feature model to use case for automated use case derivation. In: SPLC, ACM, pp 187–196, DOI 10.1145/ 2648511.2648531 262. Zaid L, Kleinermann F, De Troyer O (2011) Feature assembly framework: Towards scalable and reusable feature models. In: VAMOS, pp 1–9, DOI 10.1145/1944892.1944893 263. Zhan Z, Luo W, Guo Z, Liu Y (2017) Feature selection optimization based on atomic set and genetic algorithm in software product line. AISC 686:93–100, DOI 10.1007/978-3-319-69096-4_14 264. Zhan Z, Zhan Y, Huang M, Liu Y (2017) Product configuration based on feature model. AISC 686:101–106, DOI 10.1007/978-3-319-69096-4_15 265. Zhang G, Ye H, Lin Y (2014) Quality attribute modeling and quality aware product configuration in software product lines. SQJ 22(3):365–401, DOI 10.1007/s11219-013-9197-z 266. Zhou F, Jiao J, Yang X, Lei B (2017) Augmenting feature model through customer preference mining by hybrid sentiment analysis. ESA 89:306–317, DOI 10.1016/j.eswa.2017.07.021 267. Zhu H, Wu L, Huang K, Zhou Z (2016) Research on methods for discovering and selecting cloud infrastructure services based on feature modeling. MPE 2016, DOI 10.1155/2016/8194832 268. Ziadi T, Frias L, da Silva MAA, Ziane M (2012) Feature identification from the source code of product variants. In: ECSMR, IEEE, pp 417–422, DOI 10.1109/CSMR.2012.52 269. Štuikys V, Burbait˙e R, Bespalova K, Ziberkas G (2016) Model-driven processes and tools to design robot-based generative learning objects for computer science education. SCP 129:48–71, DOI 10. 34 José A. Galindo et al. 1016/j.scico.2016.03.009 Automated analysis of feature models: Quo vadis? 35 B Conferences and Journals This section contains all the material used in this mapping study. Rank #Papers Conferences 1 40 Software Product Line Conference 2 33 International Workshop on Variability Modelling of Software-intensive Systems 3 12 International Conference on Software Engineering 4 11 International Conference on Automated Software Engineering 5 9 International Conference on Model Driven Engineering Languages and Systems 6 9 International Conference on Software Reuse 7 7 International Conference on Generative Programming 8 6 International Conference in Software Testing 9 6 International Conference on Advanced Information Systems Engineering 10 5 International Conference on Fundamental Approaches to Software Engineering 11 4 Joint Meeting on Foundations of Software Engineering 12 4 International Symposium On Leveraging Applications of Formal Methods, Verification and Validation 13 4 Symposium on Applied Computing 14 3 International Symposium Search-Based Software Engineering 15 3 Annual Conference on Genetic and Evolutionary Computation 16 3 International Symposium on Software Engineering for Adaptive and Self-Managing Systems 17 3 Electronic Proceedings in Theoretical Computer Science 18 3 International Workshop on Modeling in Software Engineering 19 3 Working Conference on Reverse Engineering 20 3 Workshop on Scalable Modeling Techniques for Software Product Lines 21 3 International Conference on Evaluation of Novel Approaches to Software Engineering 22 2 International Conference on Software Language Engineering 23 2 Variability for You Workshop 24 2 International Joint Conference on Software Technologies 25 2 International Conference on Service Oriented Computing 26 2 International Conference of Software Engineering and Knowledge Engineering 27 2 Requirements Engineering Conference 28 2 Workshop on Product Line Approaches in Software Engineering 29 2 International Conference on Research Challenges in Information Science 30 2 European Conference on Software Architecture 31 2 International Design Engineering Technical Conferences and Computers and Information in Engineering Conference 32 2 International Conference on Web Services 33 2 International Conference on Progress in Informatics and Computing 34 2 International Symposium on Software Reliability Engineering 35 2 Workshop on Configuration 36 2 International Workshop on Requirements Engineering Practices on Software Product Line Engineering 37 2 Software Testing Verification and Validation Workshop 38 2 International Conference on Conceptual Modeling 39 2 Requirements Engineering: Foundation for Software Quality 40 2 International Conference on Cloud Computing 41 2 Advances in Transdisciplinary Engineering 42 2 Software Engineering and Advanced Applications 43 2 IEEE International Conference on Software Maintenance 44 2 International Conference on Aspect Oriented Software Development 45 2 International Conference on Distributed Multimedia Systems 46 2 International Conference on Services Computing 47 2 International Conference on Software Engineering Companion 48 2 Workshop on Feature-Oriented Software Development 49 1 Modellierung 50 1 Brazilian Symposium on Software Components, Architectures and Reuse 51 1 Workshop on Formal Methods and Analysis in Software Product Line Engineering 52 1 International Conference on Industrial Informatics 53 1 International Workshop on Description Logics 54 1 International Conference on Fundamentals of Software Engineering 55 1 International SDL Forum 56 1 Brazilian Symposium on Software Engineering, SBES 2015 57 1 International Conference on Engineering Design 58 1 International Conference on Cloud Computing 59 1 European Dependable Computing Conference 60 1 International Workshop on the Verification of Model Transformation (VOLT) 36 José A. Galindo et al. 61 1 International Workshop on Automated Configuration and Tailoring of Applications 62 1 International Conference on Software Engineering and Applications 63 1 International Workshop on Smalltalk Technologies 64 1 International Symposium on Software Reliability Engineering Workshops 65 1 Engineering Interactive Computing Systems 66 1 European Conference on Software Maintenance and Reengineering 67 1 International Colloquium on Theoretical Aspects of Computing 68 1 International Conference on Information and Communication Systems 69 1 Information Reuse and Integration 70 1 International Conference on Information, Communication and Computing Technology 71 1 International Conference on Formal Methods for Components and Objects 72 1 International Semantic Web Conference 73 1 International Working Conference on Source Code Analysis and Manipulation 74 1 International Conference on Computational Science and Engineering 75 1 FME Workshop on Formal Methods in Software Engineering 76 1 International Workshop on Variability and Complexity in Software Design 77 1 International conference on Principles and practice of constraint programming 78 1 Workshop on Recommendation Systems for Software Engineering 79 1 Central and East European Conference on Software Engineering Techniques 80 1 European Conference on Modelling Foundations and Applications 81 1 International Conference on Tests and Proofs 82 1 Modularity 83 1 Working Conference on Software Visualization 84 1 International Symposium on Object/Component/Service-Oriented Real-Time Distributed Computing Workshops 85 1 Conference on Telematics and Information Systems 86 1 International Conference of the Chilean Computer Science Society 87 1 International Symposium on New Ideas, New Paradigms, and Reflections on Programming and Software 88 1 International Conference on Model-Driven Engineering and Software Development 89 1 Workshop on Knowledge-Oriented Product Line Engineering 90 1 International Conference on Computing, Communication and Automation 91 1 Conference on Programming Language Design and Implementation 92 1 International workshop on Early aspects 93 1 European Conference on the Applications of Evolutionary Computation 94 1 International Conference on Program Comprehension 95 1 International Conference on Software and Systems Process 96 1 Workshop on software product line analysis tools 97 1 Latin American Computing Conference 98 1 The International Symposium on Software Testing and Analysis 99 1 First International Workshop on Multi Product Line Engineering 100 1 International Symposium on Formal Methods 101 1 International Conference on Software Composition 102 1 International Conference on Engineering of Complex Computer Systems 103 1 International Conference on Integrated Formal Methods 104 1 Conference on Cloud Computing and Services Science 105 1 International conference on Object oriented programming systems languages and applications 106 1 International Workshop on Search-Based Software Testing 107 1 International Conference on Software Engineering and Formal Methods 108 1 International Conference on Software Analysis, Evolution, and Reengineering 109 1 SCArVeS: Services, Clouds, and Alternative Design Strategies for Variant-Rich Software Systems 110 1 Colombian Computing Conference 111 1 International Symposium on Theoretical Aspects of Software Engineering 112 1 International Workshop on Formal Aspects of Component Software 113 1 International Conference on Industrial Engineering and Engineering Management 114 1 International Conference on Bio-Inspired Information and Communications Technologies 115 1 International Workshop on Dynamic Software Product Lines 116 1 International Conference on Enterprise Information Systems 117 1 Combining Modelling and Search-Based Software Engineering 118 1 International Conference on Artificial Intelligence and Statistics 119 1 International Conference on Concept Lattices and Their Applications 120 1 Proceedings of the on Future of Software Engineering Table 10: Conferences and number of papers from the survey Automated analysis of feature models: Quo vadis? 37 Rank #Papers Journals 1 13 Journal of Systems and Software 2 11 International Journal on Software and Systems Modeling,Software and Systems Modeling 3 9 Software Quality Journal 4 7 IEEE Transactions on Software Engineering 5 6 Information and Software Technology 6 6 Science of Computer Programming 7 5 International Journal on Software Tools for Technology Transfer 8 4 Expert Systems with Applications 9 3 Computer 10 3 Empirical Software Engineering 11 2 International Journal of Software Engineering and Knowledge Engineering 12 2 Applied Soft Computing Journal 13 2 Communications in Computer and Information Science 14 2 Journal of Universal Computer Science 15 2 Indian Journal of Science and Technology 16 2 Advances in Intelligent Systems and Computing 17 2 Requirements Engineering 18 2 Journal of Logical and Algebraic Methods in Programming 19 1 Mathematical Problems in Engineering 20 1 MDPI Entropy 21 1 Electronic Notes in Theoretical Computer Science 22 1 Journal of Intelligent Information Systems 23 1 Journal of Biomedical Informatics 24 1 Computer Languages, Systems and Structures 25 1 MDPI Open Access Journal Sustainability 26 1 Journal of Software: Evolution and Process 27 1 Scientific Annals of Computer Science 28 1 Artificial Intelligence for Engineering Design, Analysis and Manufacturing 29 1 Transactions on Software Engineering and Methodology 30 1 International Journal of Information System Modeling and Design 31 1 ACM Computing Surveys 32 1 Information Systems 33 1 Computational Ecology and Software 34 1 Frontiers in Artificial Intelligence and Applications 35 1 Frontiers of Computer Science 36 1 Electronic Communication of the European Association of Software Science and Technology 37 1 IEEE Computational Intelligence Magazine 38 1 IET Software 39 1 Journal of Computer Science 40 1 Computer Networks 41 1 Journal of King Saud University 42 1 ISRN Software Engineering 43 1 China - Journal of Software 44 1 Automated Software Engineering 45 1 Future Generation Computer Systems 46 1 Cluster Computing 47 1 IEEE Transactions on Automation Science and Engineering 48 1 Information Systems Frontiers 49 1 IFAC Paper Online 50 1 IEEE Access 51 1 Advanced Information Systems Engineering 52 1 European Journal of Operational Research 53 1 Formal Aspects of Computing 54 1 IEEE Transactions on Industrial Electronics 55 1 Transactions on Autonomous and Adaptive Systems 56 1 Studies in Computational Intelligence 57 1 Information technology and control 58 1 Software testing, verification and reliability Table 11: Journals and number of papers from the survey