guest editorial: studying work practices in global software engineering

6
Editorial Guest editorial: Studying work practices in Global Software Engineering 1. Introduction Software-driven applications increasingly pervade our lives, be it related to work or to leisure activities. This trend is rather obvi- ous in the industrialised part of the world, but also present in emerging and developing countries. Since the societal need for software production is on the rise, the software industry is growing and getting more globally interconnected. Compared to material products, software has some specific characteristics facilitating the globalisation of its production and distribution. Due to its immateriality, software code is easy to copy and to transport. While the distributed production of material products always requires shipping between locations, distributed work in the software industry can happen quasi synchronously with minor transportation costs. While the duplication of material products requires about the same input of resources, software can be duplicated at very low costs. Due to its growth over the past decades, the software industry in the industrialised world is characterised by a shortage of quali- fied labour, and as a result, a rather high level of wages. Given the specific characteristics of software, it is not surprising that the soft- ware industry in industrialised countries tried to move parts of its development work to countries that have a large pool of talent available and lower income levels. At the same time, the domestic market for IT products in these countries began to grow at an even faster rate. While these structural conditions seem to favour higher levels of globalisation in the software industry, there are also extremely relevant factors impeding globalisation that need to be taken into account. First, the application of software is linked to the local con- text of usage. Many software artefacts are deeply interrelated with the users’ social practices. These practices are shaped by history and culture. Therefore, the design of most software artefacts needs to take local requirements into account. In these cases, a process of Global Software Development needs to be grounded in the local practices of users. Secondly, the development of large size software applications depends on complex and highly cooperative work. Differences in language, cultural and educational backgrounds, as well as time zones can have a substantially disruptive effect on the software development work. Global Software Engineering seems to be an important develop- ment line in the software industry. However, a careful analysis of the phenomenon is required to identify projects that, based on their characteristics, are suitable for a distributed division of la- bour. For these projects (or parts of them), we need to identify practices that draw on the opportunities offered by distributed work, while at the same time dealing with the related obstacles. This is a challenge for both academia and industry. Companies wishing to take advantage of globalisation need to develop innova- tive techniques, tools and practices to overcome the various diffi- culties of organising and managing globally distributed software development. The academic field of software engineering, as well, needs to develop appropriate methods and perspectives to investi- gate the phenomenon of highly distributed work and to support practitioners in dealing with the related challenges. The field of Global Software Engineering (GSE) – or Global Soft- ware Development (GSD), as it is sometimes called – emerged as a transdisciplinary research arena, bringing together software engi- neers as well as social scientists and organisation theorists in- volved in examining various aspects of how globally distributed software teams function. The International Conference on Global Software Engineering has contributed significantly to the emer- gence of a community made of both scholars and practitioners interested in the field, and to the creation of a body of work related to Global Software Engineering. Given the challenges of understanding globally distributed work in the software industry, we believe that a change in method- ological focus is needed. While there have been many experimen- tal studies on problem-solving in teams, as well as interview studies with management referring to problems in distributed coordination and management, extensive participative field study material on actual workplace practices is still relatively meagre. Thus, despite occasional empirical studies of distributed software development activities over the years, there is still a dearth of well-designed studies in Information Systems, Software Engineer- ing and Computer-Supported Cooperative Work (CSCW) that would provide good examples of field research in the area. A lack of appreciation and understanding of actual work prac- tices led to problematic perspectives of how highly distributed work should be set up in the software industry. Voices from acade- mia and consultancies expressed the opinion that the additional complexity of distributed work could revitalize a vision of software engineering that had not been too successful in practice before. Slogans such as the ‘‘24 h factory’’ propagated detailed process modelling, rigid definition of the interfaces between tasks allo- cated to different sites, and a high level of formalization in the development process. However, practical experiences in distrib- uted work quickly showed the limitation of this type of visions [1]. Building on the CSCW tradition, we believe that a profound understanding of software development work practices – as op- posed to canonical or espoused theories of work practices – can en- rich the software engineering discourse and lead to appropriate 0950-5849/$ - see front matter Ó 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.infsof.2011.01.010 Information and Software Technology 53 (2011) 949–954 Contents lists available at ScienceDirect Information and Software Technology journal homepage: www.elsevier.com/locate/infsof

Upload: gabriela-avram

Post on 26-Jun-2016

216 views

Category:

Documents


3 download

TRANSCRIPT

Page 1: Guest editorial: Studying work practices in Global Software Engineering

Information and Software Technology 53 (2011) 949–954

Contents lists available at ScienceDirect

Information and Software Technology

journal homepage: www.elsevier .com/locate / infsof

Editorial

Guest editorial: Studying work practices in Global Software Engineering

1. Introduction

Software-driven applications increasingly pervade our lives, beit related to work or to leisure activities. This trend is rather obvi-ous in the industrialised part of the world, but also present inemerging and developing countries. Since the societal need forsoftware production is on the rise, the software industry is growingand getting more globally interconnected.

Compared to material products, software has some specificcharacteristics facilitating the globalisation of its production anddistribution. Due to its immateriality, software code is easy to copyand to transport. While the distributed production of materialproducts always requires shipping between locations, distributedwork in the software industry can happen quasi synchronouslywith minor transportation costs. While the duplication of materialproducts requires about the same input of resources, software canbe duplicated at very low costs.

Due to its growth over the past decades, the software industryin the industrialised world is characterised by a shortage of quali-fied labour, and as a result, a rather high level of wages. Given thespecific characteristics of software, it is not surprising that the soft-ware industry in industrialised countries tried to move parts of itsdevelopment work to countries that have a large pool of talentavailable and lower income levels. At the same time, the domesticmarket for IT products in these countries began to grow at an evenfaster rate.

While these structural conditions seem to favour higher levelsof globalisation in the software industry, there are also extremelyrelevant factors impeding globalisation that need to be taken intoaccount. First, the application of software is linked to the local con-text of usage. Many software artefacts are deeply interrelated withthe users’ social practices. These practices are shaped by historyand culture. Therefore, the design of most software artefacts needsto take local requirements into account. In these cases, a process ofGlobal Software Development needs to be grounded in the localpractices of users. Secondly, the development of large size softwareapplications depends on complex and highly cooperative work.Differences in language, cultural and educational backgrounds, aswell as time zones can have a substantially disruptive effect onthe software development work.

Global Software Engineering seems to be an important develop-ment line in the software industry. However, a careful analysis ofthe phenomenon is required to identify projects that, based ontheir characteristics, are suitable for a distributed division of la-bour. For these projects (or parts of them), we need to identifypractices that draw on the opportunities offered by distributed

0950-5849/$ - see front matter � 2011 Elsevier B.V. All rights reserved.doi:10.1016/j.infsof.2011.01.010

work, while at the same time dealing with the related obstacles.This is a challenge for both academia and industry. Companieswishing to take advantage of globalisation need to develop innova-tive techniques, tools and practices to overcome the various diffi-culties of organising and managing globally distributed softwaredevelopment. The academic field of software engineering, as well,needs to develop appropriate methods and perspectives to investi-gate the phenomenon of highly distributed work and to supportpractitioners in dealing with the related challenges.

The field of Global Software Engineering (GSE) – or Global Soft-ware Development (GSD), as it is sometimes called – emerged as atransdisciplinary research arena, bringing together software engi-neers as well as social scientists and organisation theorists in-volved in examining various aspects of how globally distributedsoftware teams function. The International Conference on GlobalSoftware Engineering has contributed significantly to the emer-gence of a community made of both scholars and practitionersinterested in the field, and to the creation of a body of work relatedto Global Software Engineering.

Given the challenges of understanding globally distributedwork in the software industry, we believe that a change in method-ological focus is needed. While there have been many experimen-tal studies on problem-solving in teams, as well as interviewstudies with management referring to problems in distributedcoordination and management, extensive participative field studymaterial on actual workplace practices is still relatively meagre.Thus, despite occasional empirical studies of distributed softwaredevelopment activities over the years, there is still a dearth ofwell-designed studies in Information Systems, Software Engineer-ing and Computer-Supported Cooperative Work (CSCW) thatwould provide good examples of field research in the area.

A lack of appreciation and understanding of actual work prac-tices led to problematic perspectives of how highly distributedwork should be set up in the software industry. Voices from acade-mia and consultancies expressed the opinion that the additionalcomplexity of distributed work could revitalize a vision of softwareengineering that had not been too successful in practice before.Slogans such as the ‘‘24 h factory’’ propagated detailed processmodelling, rigid definition of the interfaces between tasks allo-cated to different sites, and a high level of formalization in thedevelopment process. However, practical experiences in distrib-uted work quickly showed the limitation of this type of visions [1].

Building on the CSCW tradition, we believe that a profoundunderstanding of software development work practices – as op-posed to canonical or espoused theories of work practices – can en-rich the software engineering discourse and lead to appropriate

Page 2: Guest editorial: Studying work practices in Global Software Engineering

950 Editorial / Information and Software Technology 53 (2011) 949–954

process models, levels of formalization, and tool support. Againstthis background, the purpose of this special section is to examinepractices of Global Software Engineering and to reflect upon thestrengths and limitations of empirical research methods being de-ployed in the field.

Dittrich et al. [2] have already demonstrated the relevance ofqualitative research in Software Engineering. They indicated howmethods and theoretical frameworks can vary significantly acrossstudies, used ‘‘under different epistemological paradigms, and withdifferent theoretical underpinnings’’. Methods are not simply tech-niques to be chosen and deployed at will, but are constructed fromparticular conceptual worldviews, and entail theoretical commit-ments. The actual use of methods also requires training and sensi-tivity to the local situation. Since these issues are not very oftenadequately dealt with, we have asked the authors to carefullyreflect on them in this volume.

2. Background of the special section

The collaboration between the guest editors started with theorganisation of a workshop in conjunction with the InternationalConference on Global Software Engineering (ICGSE, 2007), togetherwith Liam Bannon and Alexander Boden. The workshop generateda lot of interest, and some of the participants suggested following itup with a journal special issue. The guest editors approached theInformation and Software Technology Journal and received ClaesWohlin’s support to distribute an open call in order to elicit papersfor a potential special issue on Studying Work Practices in GlobalSoftware Engineering in early 2009.

We received 13 submissions in total; six were selected after afirst round of reviews, six others have been rejected and one waswithdrawn. Because we were interested in field studies of actualin situ practices of software engineers, papers that were presentingexperiments or/and were considering students as proxies for soft-ware developers were not retained for the special section.

3. Review process

During the first round of reviews, we ensured three reviews foreach paper, one of them coming from an author of another submit-ted article, and two provided by experts in the field. We tried to re-cruit reviewers who were experts in the topic, and who couldgenuinely support the authors to improve their papers. The firstselection was based merely on the fulfilment of the criteria speci-fied in the call for papers: in situ studies of work practices of soft-ware engineers, while also taking into account the strengths andweaknesses of each paper.

The second round of reviews provided two reviews by externalexperts for each paper. Five of the six papers were retained for thespecial section. This second round of reviews focused on:

– Ensuring that the papers are comprehensible for the intendedaudience – by including enough details about the study itselfand its major findings, besides the focus on the methodologicalapproach.

– Ensuring an appropriate level of reflection on the strengths andlimitations of the empirical research methods deployed.

A third round of reviews was undertaken by the editors of thespecial section, in order to ensure consistent discussions on themethodological aspects in all the papers included.

4. The papers included in the special section

The first article by Johri, ‘‘The Creation of Location-SpanningWork Practices by Global Software Developers’’ belongs to a ser-ies of practice-based studies of work, looking at work practices ofglobal software developers in two sites of a multinational companysituated in US and Ireland, respectively.

Taking an interpretivist stance and applying ethnographically-informed methods, Johri has looked at how workers used technol-ogy as part of their work practices, focusing on their choices to usemultiple media in personalised ways.

Based on Orlikowski’s [3] notion of sociomateriality, the articleargues that practices are ensembles of artefacts and social behav-iour – brought together by human actors through bricolage. Theauthor identifies this process as ‘‘sociomaterial bricolage’’.

The ‘‘Methodological Reflections on a Field Study of a Glob-ally Distributed Software Project‘‘ of Patil, Kobsa, John and Selig-mann contemplate an ethnographically-informed studyundertaken in a multinational corporation with four sites in USand one in India. The authors provide the rationale behind theirmethodological choices and list insights that could inform futurestudies of Global Software Development. Their argument favoursmethodological breadth and a plurality of methods, in order to gaina broad understanding of a global project.

While setting to study the tension between the team members’awareness needs regarding an individual’s activities, and the indi-vidual’s desire for privacy, the authors have taken into account allwork practices they were able to observe.

Sigfridsson and Sheehan’s article ‘‘On Qualitative Methodolo-gies and Dispersed Communities: Reflections on the Process ofInvestigating an Open Source Community’’ is an overview ofthe methodological issues encountered during a longitudinal studyof an Open Source community named PyPy over a period of twoyears. Positioned as an interpretive study based on ethnographi-cally-informed data collection methods, their article focuses onthe multidimensionality of the Open Source phenomenon andthe challenges of qualitative interpretation of activities in a long-term context.

The direct involvement of multiple researchers in the study, to-gether with its extended duration and the direct participation ofthe first author in the community for the purpose of this study leadto the deployment of various ethnographically-inspired methodswith the purpose of developing an understanding of the commu-nity and of their way of achieving distributed software develop-ment in practice.

Clear and MacDonell, in their article ‘‘Understanding Technol-ogy Use in Global Virtual Teams: Research Methodologies andMethods’’, rely on structuration theory, action research andGrounded Theory methodologies to construct their own approachon studying technology-use mediation in a collaboration betweenfaculty, support staff and teams of students located in New Zea-land, Sweden and the US, respectively. The framework they devel-oped for researching technology-use mediation in global virtualteams is labelled ‘‘Technology-use Mediated Adaptive Structur-ation Theory’’ (TUMAST).

Having chosen to study their own practices by analyzing avail-able email data for an extended period of time, the authors haveidentified ‘‘episodes’’ as their units of analysis, and devised six dif-ferent elements of analysis for studying each of these episodes.Their article represents a longitudinal interpretive field study ofglobal virtual teams focusing on the actions and interactions ofthese teams and presenting the authors’ effort to develop the the-ory by introducing a more structured approach for analysing fac-tual data generated in GSD.

Page 3: Guest editorial: Studying work practices in Global Software Engineering

Editorial / Information and Software Technology 53 (2011) 949–954 951

The article authored by Boden, Müller and Nett, ‘‘Conducting aBusiness Ethnography in Global Software Development Projectsof Small German Enterprises’’ introduces an action research var-iation labelled ‘‘Business Ethnography’’. Its particularity consistsin defining a joint project for the research team and the employeesto work together, in an attempt to improve existing organisationalpractices. In this case, the aim was the initiation of an organisa-tional development process. The researchers went onto collect datavia interviews, field site observation and document analysis; thisdata was analyzed, articulated and presented to the practitionersin a detached, more impersonal, but multi-faceted manner- meantto support the ability for self-organisation and to enrich the per-spective of project participants on their own practices.

Two consecutive studies were undertaken: the first one, anexploratory field study, included a number of different GermanSMEs involved in offshoring; the second was conducted in one ofthese German SMEs collaborating with a development group inRussia, and the Business Ethnography framework was adopted.

5. Commonalities and variations

5.1. Theoretical underpinnings and methodological choices

All the studies included in this special section position them-selves as interpretive studies and, to a certain extent, deploy eth-nographically-inspired methods for data collection, while relyingon Grounded Theory-derived methods for coding and analysis.

The first three articles acknowledge using ethnographically-in-formed methods, while the two other papers position themselvesas action research; the Clear paper is building on structuration the-ory, trying to devise a method for analysing selected episodes oftechnology-use mediation, while the Boden study introduces theBusiness Ethnography concept as action research lead via a jointproject and meant to improve existing practices.

The studies included in this special section also position them-selves as having merely an exploratory and explanatory character.

5.2. Issues under discussion

The study of work practices in GSE being the very topic of thisspecial section, all the articles included here are focusing on workpractices, while the angles they take on the issue vary. The role ofarticulation work – and not that of ‘‘coding or writing documenta-tion per se’’ – is central to the Sigfridsson paper. Boden looks atarticulation practices involved in organisational development,while Johri is dividing practices into local and location-spanningpractices. Patil’s attention is drawn to the use of collaboration toolsand to the different subcultures and practices around their use indifferent locations. Clear is setting to investigate the role of tech-nology-use mediation in supporting the work of global virtualteams; for this purpose, they develop and apply their own novelframework.

Another set of issues brought into discussion by most of thearticles included here are problems of gaining access to multiplefield sites and obtaining funding for extended periods of study.Gaining access to the field and maintaining a good relationshipwith the practitioners involved in the study is an omnipresenttheme. Johri and Patil had to negotiate access starting from thetop management level, but afterwards, it was paramount to gainthe support of local managers in each location. The access to multi-ple locations can be further complicated by language and culturaldifferences. Patil also shows how other details, like the researcher’saffiliation, the incentives offered to the practitioners (contributingto their work, sharing the findings), the various project phases and

temporal rhythms are extremely important in undertaking such astudy.

Sigfridsson warns against the apparently ready availability ofdata from Open Source projects and the temptation to solely usepublic resources for data mining without getting access to the hu-man actors and their specific social context. Taking data out of con-text can be extremely misleading, and without getting familiarwith the context, the researcher can easily get lost in the vastamounts of data. Getting involved in the community is vital forthe researchers, especially when adopting the interpretive para-digm, and this involvement can range from non-participant obser-vation to becoming an active contributor to the project, especiallywhen the barriers for newcomers are low, as in the case of PyPy.

The Clear study illustrates the special situation of the research-ers–practitioners, when access to the field is not a problem, havingthe advantage of being already fully immersed in it.

Funding sources for such a study and getting access to the nec-essary resources are issues that have an impact on the methodschosen for a study. In some situations, the researchers were fortu-nate to receive funding for an extended period of time with nostrings attached (Sigfridsson), while others had to elicit the finan-cial support of the organisations under study (Johri). Granting ac-cess is typically a political decision of the top management, andit usually happens if the organisation under study has either aninterest in the research topic or the power to steer the study in adirection that interests them. The study findings and specific rec-ommendations are sometimes offered in exchange for access andfunding. In the case of Boden, the researchers had to apply forfunding and obtain it in order to finalise the study; the lack offinancial support from the government for the SME under studymade their involvement more difficult.

An aspect often neglected by researchers is that of opportunitycosts; the time spent by practitioners being interviewed or other-wise supporting the research can be seen as a financial burdenby the organisations targeted. Maintaining a good relationshipand an open communication style with their contacts in the organ-isations under study, as well as offering some value in exchange fortheir time are important factors mentioned in the Johri, Patil, Sig-fridsson and Boden studies as being very important for thecooperation.

5.3. Establishing the boundaries of the research

Defining the unit of analysis is an important decision whendesigning a study, and it is related with that of choosing the appro-priate methodology. Patil selected a distributed project team; Johrilooked at an organisation with sites in several US locations and inIreland; Sigfridsson studied an Open Source community, whileBoden looked at the inter-organisational aspects of an offshoringarrangement. Finally, Clear have defined episodes (‘‘relevant tem-porally-bound sequences of events with antecedent conditionsand outcomes that stand apart from others’’) as their unit ofanalysis.

5.4. Adapting established techniques

Undertaking research in a distributed environment has a num-ber of particularities, as an important part of the communicationbetween sites is mediated by technology and can remain inacces-sible to the researchers.

While interviews, participant and non-participant observationand document analysis were the techniques deployed in almostevery study, a number of other specific issues are raised in thearticles included in this special section.

Page 4: Guest editorial: Studying work practices in Global Software Engineering

952 Editorial / Information and Software Technology 53 (2011) 949–954

Travelling to various locations appears to be essential for such astudy. The location, office layout, the transport facilities, the can-teen, the languages spoken at work are all relevant details of thecontext of work. Talking about funding and time constraints, Johrisuggests that ‘‘short visits are better than no visits’’. Participationin informal events – lunches, social gatherings, coffee breaks, is an-other important opportunity to understand the context in whichwork is taking place.

Data collection methods such as shadowing, photo taking,sketching, video and audio recording, together with field noteson ‘‘location, décor, services available to workers, formal and infor-mal interactions’’ (Johri) were deployed when visiting field sites.Sigfridsson, who joined a PyPy sprint as an active participant,speaks about the tension between enacting practices on one side,and observing and documenting them on the other side, echoingthe findings of Schulze [4].

The importance of observing interactions by participating inproject meetings is pointed out by Patil, Johri and Sigfridsson. Alot can be learned about people’s roles, artefacts and practices inthese meetings. Later on, the cues picked up in meetings can beused for searching and analyzing documents and for conductingboth formal and informal interviews.

The study of existing documentation, mailing lists, project man-agement data and team room content is another rich source ofdata. Johri speaks about studying ‘‘archival data’’ – in context ornot. Boden shows how he used interview data to guide him inselecting and accessing project documentation. Sigfridsson talksabout the dangers of wasting time and getting lost in the hugeamount of data publicly available from Open Source projects, andabout how members of the PyPy community pointed him to self-reflection materials that were relevant – such as blog posts, mes-sages on mailing lists or code committing visualizations.

The issue of virtual ethnography [5] or ‘‘computer-supportedethnographical observation’’ [6] is brought up by Sigfridsson. Online observation of the practices of a community is particularly rel-evant when most of the interaction in the community happens inthe virtual space.

In two of the studies, on line questionnaires were also applied.While Patil deployed them for ‘‘delving deeper in issues’’ andreaching more employees, Sigfridsson used them for mirroringback their findings to the members of the community and attempt-ing to validate these findings.

One difficulty pointed out by Patil and Boden is that of con-necting data collected with different instruments and recon-structing events that haven’t been witnessed by the researcher(because they happened in a different time or a different space).Triangulation is not always possible, and the mission of theresearcher resembles to resolving a puzzle where many piecesare missing.

Studying the practices in multiple locations is essential, accord-ing to Patil. Local work practices and the specific work environ-ments in other locations remain often unknown, hidden, toindividual project members. These globally invisible, but locallyrelevant work practices cannot be discovered unless a visit is madeto the location. The situation is further complicated by the fact thatmany employees nowadays work from home- either full time, or acouple of days a week.

The following aspects are pointed out as facilitating the re-searcher’s work. Although they are not necessarily specific to re-search in distributed environments, they are nevertheless relevant:

� Preparing access to the field by consuming available material –mailing lists, archives, blogs, profiles/intranet or public sources.

� Using the same paths as a new hire for getting familiarised withthe culture of the organisation and that of the site.

� Aligning data collection practices with organisational workpractices.

� Adapting to different time zones, to the dynamics of the organi-sation (project life cycle, employee turnover, different degrees ofemployee involvement in a project at different points in time)(Patil) and to the happenings in the field(Johri).

The necessity to mix and match various methods is emphasizedby the majority of studies included in this special section, in linewith [7,8]. Global Software Engineering – and related work prac-tices – are part of a very complex phenomenon, and ‘‘no single datasource or data gathering technique could have provided the coher-ent description’’(Patil). Some of the authors use creatively well-established methods, while others (Clear, Boden) go a step furtherand suggest new theoretical and methodological frameworks.

5.5. Data analysis

With no exception, the authors of the studies included in thisspecial section applied Grounded Theory – inspired techniquesfor coding and analysis. Combining ethnographically-informeddata collection methods that allow covering a wider range ofissues (and are traditionally deployed in CSCW for the studyof work practices), with a more formalized approach likeGrounded Theory for coding and analysis seems to illustratean existing trend. Ethnography is ‘‘an inductive, open-endedmode of inquiry’’ [9], resulting in long descriptive stories tryingto reflect the complexity of the phenomenon being studied. Apossible explanation for replacing these stories with more struc-tured accounts could be the need for rigor and scientific validitywhile doing interpretive studies in Software Engineering.

Different flavours of Grounded Theory are used: Sigfridssonbuilds on Strauss and Corbin [10], while Johri uses a Strauss varia-tion[11]. Patil chooses not to adopt ‘‘a strict Glaserian or Straussianstance of Grounded Theory’’, in line with many other studies ininformation fields [12]. Clear’s six mutually reinforcing strategiesfor episode analysis are building on Grounded Theory, but aremeant to provide a ‘‘more holistic analysis informed by the inti-mate knowledge of data gained from the earlier grounded formsof analysis in the episode’’.

Another interesting issue is the evolution of interpretationsover time, as the researchers delved deeper into the field. Clearshows how ‘‘the episodes presented themselves progressively asthe data were prepared for analysis’’, emphasizing the ‘‘emergentnature of the processing of data in an interpretive study’’. Sigfrids-son tackles the same issue, illustrating their focus shift from onephase of the study to another, while Boden presents a study thatinvolved two projects deploying different methods, though onewas the continuation of the other.

5.6. The motivation for this type of research

The studies included in this special section describe a widerange of motivations for studying work practices. For Sigfridsson,the motivation is to ‘‘uncover previously unrecognized problemsof Global Software Development’’, unveil hidden complexitiesand ‘‘pose questions that can be applied for the improvement ofpractice’’, in line with Robinson and Sharp [12].

Regarding the potential generalisation of findings, Johri’smotivation is to present the case and serve it to others to con-tinue building on it. Sigfridsson finds that the generalisation po-tential lies rather in evolving from empirical findings totheoretical propositions – and not from one case to a larger pop-ulation [13].

Page 5: Guest editorial: Studying work practices in Global Software Engineering

Editorial / Information and Software Technology 53 (2011) 949–954 953

Clear aimed to develop a framework for analysing the activitiesof global virtual teams and an ‘‘interpretive approach based on rig-orous methods’’, while Boden’s motivation was to develop an ac-tion research approach that would ‘‘prove beneficial for both theorganisation and the research team’’.

6. Conclusion

The studies agree that Global Software Engineering practiceshave to be approached from a dynamic, holistic perspective –which is the only alternative for gaining a broad understandingof the distributed work settings. By providing a rationale for theirmethodological choices, the authors offered extremely useful in-sights into the research process. The research work was presentedin its development, admittance of problems, indications of changesin tactics, and evolutions due to given opportunities. Several of thestudies emphasized the importance of approaching the field withan open mind, and not imposing a model or hypothesis and forcingthe reality into it.

Patil showed the importance of paying attention to routinework practices – regardless of the fact that they might seem irrel-evant to the research question. Both Johri and Boden spoke aboutthe benefits and dangers of the researcher’s familiarity with thefield. In both Clear’s and Sigfridsson’s view, the intimate knowl-edge of data can make new dimensions of analysis emerge.

In doing interpretive studies, the choice of methods is made in ac-cord with the research questions at hand. However, in such longitu-dinal studies, the research questions can evolve as well and the focuscan shift, requiring different methods to be deployed. As Johri puts it,new ways of doing qualitative research emerge, with the researchersthemselves engaging in ‘‘sociomaterial bricolage’’. The researchersget to develop their own version of the methodology, by adaptingand adjusting it to the particular circumstances. Clear’s study is prob-ably the best example of evolving the theory (in this case, the theoryof technology-use mediation) in a global context and developing newtechniques for data analysis, their ambition being to later generalizethis approach across multiple contexts.

Reflexivity is an essential characteristic of interpretive research.No such study is complete without a reflection of the researcherson their own role and on how their backgrounds and researchinterests might have influenced both the selection of the topic,and the methodological choice.

In order to ensure the study’s validity, researchers have to‘‘strive to truthfully account for why decisions are made and expla-nations suggested throughout the study’’ (Sigfridsson).

Regarding the relationship with the organisation(s) understudy, the studies show that these can benefit from work practicesstudies in multiple ways. Sometimes the researchers come up witha different perspective, ‘‘making the familiar unfamiliar’’[14] as inSigfridsson’s case, or providing a detached, multi-faceted and oftencontradictory image of the situation (Boden). The account of localpractices that are not transparent for the other locations can pro-vide inspiration for practitioners and trigger knowledge sharingprocesses. Researchers can often be seen as boundary spanners –bridging locations, cultures, uncovering hidden practices – whoprovide inspiration for practitioners and contribute to cross-pollination.

Some of the challenges exposed by the studies included here arethe danger for the research to be mistaken for consulting, and thelack of sync between the practitioners’ expectations and the re-search team’s output. Such studies tend to span over long periodsof time, and the researchers’ feedback tends to arrive too late. Bod-en also spoke about the inherent danger for the researchers to getinvolved in organisational micro-politics and have their neutralrole compromised.

The papers collected in the special section provide in depth in-sights into distributed work practices in the software industry.They uncover a multitude of interesting insights into complexcooperative settings and their socio-material conditions. Docu-menting the details of these cases enables learning by comparisonand transfer. An appropriate understanding of work practices maylead to their appreciation and prevent academics, consultants andmanagers from relinquishing them following overly simplistic vi-sions of potential futures.

We hope that this special section will provide inspiration for fu-ture research on work practices in distributed software engineer-ing. Enjoy the reading!

Acknowledgements

We would like to thank all authors for their hard work and resil-ience in bringing their studies to the final format included in thisspecial section.

We would like to thank the reviewers for their important con-tribution to this special section. The special section would havenot existed without their kind support!

� Ban Al Ani, University of California Irvine, US.� Liam Bannon, University of Limerick, Ireland.� Yochai Ben-Chaim, Tadiad, Israel.� Marcelo Cataldo, Carneggie-Melon University, Pittsburg, US.� Ricardo Colomo-Palacios, Universidad Carlos III, Madrid, Spain.� Daniela Damian, Victoria University, Canada.� Yvonne Dittrich, IT University of Copenhagen, Denmark.� Nicolas Ducheneaut, XEROX PARC, Palo Alto, California.� Amar Gupta, University of Arizona, US.� Armin Heinzl, Mannheim University, Germany.� Shihong Huang, Florida Atlantic University.� Sabrina Marczak, Victoria University, Canada.� Allen Millewski, Monmouth University, US.� Rafael Prickladnicki, PUCRS, Brazil.� David Redmiles, University of California, Irvine, US.� Helen Sharp, Open University, UK.� Cleidson de Souza, IBM Research, Brazil.� Gunnar Stevens, University of Siegen, Germany.� Jayakanth Srinivasan, MIT, Boston, MA, US.� Kathleen Swigger, University of North Texas, US.� Rainer Todtenhoefer, University of Applied Sciences, Fulda,

Germany.

References

[1] A. Boden, B. Nett, V. Wulf, Trust and social capital – revisiting an offshoringfailure story of a small German software company, in: Proceedings of theEleventh European Conference on Computer Supported Cooperative Work(ECSCW 2009), Springer, London, 2009, pp. 123–142.

[2] Y. Dittrich, M. John, J. Singer, B. Tessem (Eds.), Special issue on qualitativesoftware engineering research, Information and Software Technology 49(2007).

[3] W.J. Orlikowski, S.V. Scott, Sociomateriality: challenging the separation oftechnology, work and organization, The Academy of Management Annals 2 (1)(2008) 433–474.

[4] U. Schulze, A confessional account of an ethnography about knowledge work,MIS Quarterly 24 (1) (2000) 3–41.

[5] C. Hine, Virtual Ethnography, SAGE, London, 2000.[6] N. Ducheneaut, Socialization in an open source software community: a socio-

technical analysis, Computer Supported Cooperative Work (CSCW): AnInternational Journal 14 (2005) 323–368.

[7] W.J. Orlikowski, J.J. Baroudi, Studying information technology in organizations:research approaches and assumptions, Informations Systems Research 2 (1)(1991) 1–28.

[8] J. Mingers, Combining IS research methods: towards a pluralist methodology,Information Systems Research 12 (3) (2001) 240–259.

Page 6: Guest editorial: Studying work practices in Global Software Engineering

954 Editorial / Information and Software Technology 53 (2011) 949–954

[9] S.E. Sim, Evaluating the evidence: lessons from ethnography, in: Proceedings ofthe Workshop on Empirical Studies of Software Maintenance, Oxford, England,1999, pp. 66–70.

[10] A.L. Strauss, J.M. Corbin, Basics of Qualitative Research: Grounded TheoryCoding Procedures and Techniques, Sage, London, 1990.

[11] A. Strauss, Qualitative Analysis for Social Scientists, Cambridge UniversityPress, 1987.

[12] R. Matavire, I. Brown, Investigating the use of ‘‘Grounded Theory’’ ininformation systems research, in: SAICSIT‘08: Proceedings of the 2008Annual Research Conference of the South African Institute of ComputerScientists and Information Technologists on IT Research in DevelopingCountries, ACM, New York, NY, USA, 2008, pp. 139–147.

[13] H. Robinson, J. Segal, H. Sharp, Ethnographically-informed empirical studiesof software practices, Information and Software Technology 49 (2007) 540–551.

[14] A.S. Lee, R.L. Baskerville, Generalizing generalizability in information systemsresearch, Information Systems Research 14 (3) (2003) 221–243.

Gabriela Avram⇑Interaction Design Centre,

University of Limerick, ER1003,

Plassey Campus, Limerick,Ireland

Tel.: +353 (0) 61 202782; fax: +353 (0) 61 213484.E-mail address: [email protected]

Volker WulfInformation Systems and New Media,

University of Siegen, H-B 8420,Hölderlinstr. 3, 57076 Siegen,

GermanyTel.: +49 (0) 271/740 2910; fax: +49 (0) 271/740 3384.

E-mail address: [email protected]