Factors affecting adoption of Open source software by individuals
Young, Alison, Griffith University, Nathan 4111 Brisbane, Australia, Alison.Young@student.griffith.edu.au
This research aims to identify factors that affect an individual when adopting open source software. An examination of the literature includes a definition of open source software and brief accounts of its use within governments and organisations. The research question is to identify what factors influence the decision to adopt open source software in end-users, or individuals. The diffusion of innovation model was used to interpret findings from previous research to identify factors. The primary factor identified was regarding ease of use and complexity of open source software. Research that utilises the multi-level framework in addition to a MIS discipline theory will provide a sound basis for continuing research on this topic.
Keywords: adoption, end-users, open source software.
This paper presents an investigation of literature demonstrating that end-user adoption of open source software has been stalled by an assortment of issues. Issues affecting the rate of end-user adoption of open source software include factors such as accessibility, ease of use, and knowledge of its existence. The paper will also explain routes for future research to better understand whether any change in the rate of adoption or issues affecting rates of adoption have occurred in the period since the literature was published.
The purpose of this paper is to inspire investigation into the opportunities that use of open source software can bring to everyday end-users. With such opportunity freely available, information systems researchers need to explore why end-users are not utilising this source of software. The purpose of this paper is to provide insight into what factors are attributed to the rate of adoption of open source software among end-users. The research question being addressed in this paper is what are the factors that affect end-user adoption of open source software? With knowledge of these factors in hand, it will be possible for developers and distributors of open source software to find ways to circumvent issues and thus enable their end-user numbers to grow.
The scope of this paper is to present findings from research and articles from literature that describe and explain why the rate of adoption of open source software is lagging behind that of proprietary software. The background literature on open source software will provide some factors, or issues, that can be used to analyse the affects on the rate of end-user adoption.
Preliminary literature synthesis will then enable further research to be done on this topic which may include development of a case study to analyse the current situation and make conclusions as to whether factors identified in the literature are still present. Explanations from open source software literature will provide some background details about end-user adoption.
Use of a case study in association with the theoretical framework will allow factors and issues to be identified that are relevant to the adoption of open source software. It is important to remember that for the purpose of this paper, open source software will be deemed an innovation.
The following paper will contain a literature review comprising background information about open source software and where it has been used in governments and organisations. Adoption of open source software within those environments will also be addressed and then related to individual adoption. Theoretical frameworks will be used to analyse case studies and introduce the use of a multi-level framework for researching open source.
- Literature review
Due to the nature of this research topic, much of the literature that has been published does not fall under the classification of an academic quality publication. This is primarily due to the nature and actions of the open source community and the belief that information should be freely available.
Another reason for the lack of academic papers may be that there has simply been a lack of academic level research on this particular topic. This research will serve to highlight the importance and interest of this topic. An exploratory study has revealed that there is a recent but growing interest in open source software and how it may be used to provide software in developing countries due to its low or zero cost to use (Evans and Reddy, 2003). This is a growing interest area and an important topic, however only very recently have approaches been devised to study open source as a wider concept by use of a multi-level framework (Niederman, Davis, Greiner, Wynn and York, 2006a). Further details pertaining to this framework and how levels can be used to examine different aspects of open source will be discussed in the research framework and method section (Niederman et al., 2006a). In addition different discipline theories and their relevance to researching open source will be discussed, in particular diffusion of innovation which is the theoretical framework used to examine this topic (Niederman et al., 2006a).
- What is Open Source Software?
The first and most important thing to do for this research topic is to establish a definition for open source software. The Open Source definition version 1.9 contains 10 clauses which specify the criteria which must be met if software is to be classed as open source (Open Source Initiative, 2006).
The clauses are as follows (Open Source Initiative, 2006):
1. Free Redistribution – There shall be no restriction to any party from selling or giving away the software and also not require payment of royalties or fees for the sale.
2. Source Code – Source code must be included, and must allow distribution in source code as well as in compiled form; if code is unavailable there must be a well advertised way of obtaining code that is able to be modified by the programmer.
3. Derived Works – Modification of software must be allowed and the modified program must be distributed under the same license as the original software.
4. Integrity of the Author’s Source Code – There may be restrictions on source-code from being distributed in modified form only if the license allows the distribution of "patch files", else modified works may be required to change the name or version number from the original.
5. No Discrimination Against Persons or Groups – No groups or persons may be discriminated against.
6. No Discrimination Against Fields of Endeavour – There must be no restrictions on the field in which the software is used.
7. Distribution of License – the license must apply to all who receive the program through redistribution without the need to agree to another license.
8. License Must Not Be Specific to a Product – the license must not be specific to a program that is part of a particular distribution.
9. License Must Not Restrict Other Software – no restrictions on software that is distributed with the licensed software must exist.
10. License Must Be Technology-Neutral – the licence must not be dependant on any particular technology or interface style.
While the details of the criteria that ensures software is in fact open source may appear exhaustive. It is ensuring that not just the developers of software have access to the source code and the right to make modifications that will enhance the software for every user. This freedom to alter, customise and redistribute is often in association with the freedom of obtaining the product without a monetary exchange taking place.
- Organisations and Open Source Software
Organisations have long been aware of open source software and its benefits. Use of open source software is increasing in organisations and organisational awareness is growing. In particular recent literature is discussing how open source software may be used in organisations and what benefits can be found from development and use of software in-house in conjunction with sharing of developed software between organisations (Skidmore, 2005).
How open source software is adopted and used in organisations is outside the focus of the research. The acknowledgement of organisational involvement and use of open source software is present to demonstrate that the advantages of open source software have been discovered and utilised by the corporate sector.
- Governments and Open Source Software
Governments are increasingly adopting open source software for their departments. This is especially prevalent in European countries which are adopting open standards and deliberately avoiding proprietary software developers (Evans and Reddy, 2003).
Evans and Reddy (2003) surveyed government proposals and initiatives that concern open source software and found that many countries were adopting the use of open source software. Their findings revealed that Germany and France were at the forefront of promoting open source (Evans and Reddy, 2003). In particular the French public sector is moving towards a complete open-source infrastructure (Evans and Reddy, 2003). Brazil, Italy, Spain and Venezuela have all passed resolutions that either force or encourage the government of each nation to use open source software (Evans and Reddy, 2003).
Demonstration of government involvement and public resolution for their governments to use open source software is a promising sign that individuals are becoming increasingly aware of open source software.
- Adoption of Open Source Software
Literature shows that organisations are adopting open source software but more interestingly lists have been produced of technical and management requirements that the open source software must satisfy (Wang and Wang, 2001). While such lists may be useful for determining important factors within an organisation, similar factors will likely not arise when analysing adoption on an individual scale, however using organisational requirements allows for comparison of individual equivalents. Other studies have used a grounded theory approach to investigate adoption, but again this has taken place on an organisational level (Dedrick and West, 2003).
While quite a number of conference papers have focussed on organisational adoption of open source software, there are a few papers which have taken an individual perspective. An end-user is also known as an individual user, somebody much like you or me. The needs, perceptions and capabilities of an end-user differ from those of an organisation. It is for this reason that organisational based research can not be directly applied to end-users.
The article by Sullivan (2001) presents an end-user perspective piece on the perceived factors that are affecting adoption of open source software. This type of paper is useful as research in this topic as while it contains author bias, it does give a list of issues and concerns (Sullivan, 2001). Using such a list as a starting point it is then possible to seek out similar factors in related literature which will assist in standardising factors and aiding to remove bias. The advantage of end-user authored articles is also that they present an interesting picture of the perceptions that some end-users may have towards open source software.
Research conducted by Hussein (2003) in a Malaysian university sought to discover why, despite the growing phenomenon of open source software in tertiary level institutions across a number of countries, the rate of adoption among Malaysian computer science students was minimal. A survey was conducted and the results show that while many students have used open source software operating systems, only approximately 30% of students surveyed enjoyed it (Hussein, 2003). Most students who owned a computer didn’t have an open source operating system installed, those who did were doing so as it was required for school assignments (Hussein, 2003). The factor identified that describes why students didn’t like open source software is that it is hard to use and not user-friendly, the next most stated factor was that the interfaces were unattractive (Hussein, 2003).
While the research by Hussein (2003) was a very small study, it does give insight into what an end-user finds most important in software that they are using. Benefits of open source software identified by students were primarily related to learning by source code analysis. This may not be indicative of a typical end-user whom is not a student.
- research framework and method
The research method utilised for this paper is that of a literature review. The results and findings of the research conducted by others will be used to introduce and substantiate the importance of the research topic. The topic of research is analysis of documents commenting on the adoption of open source software by individuals and factors affecting adoption. Due to the nature of open source software development and use, there has been little research on its use outside of organisational contexts. Individual use and adoption has not been a topic with much accompanying research. Research that has been conducted on this topic rarely falls under the classification of ‘academic research’ and thus demonstrates a need for academic level research to be conducted on this topic.
Most recently a series of papers have been published that describe a research agenda for studying open source (Niederman et al., 2006a, Niederman, Davis, Greiner, Wynn and York, 2006b). This series of three papers commences by describing a multi-level framework of which the discrete levels can be used to analyse information systems through open source software (Niederman et al., 2006a). There are five levels described in the framework; the artifact, the individual, the group/ the project/ the community, the organisation and the broader societal perspective (Niederman et al., 2006a). Of these levels the one addressed in this paper will be that of the individual, subsequent studies may wish to address how the individual corresponds to the broader societal perspective. Niederman et al (2006a) state that examining a level individually lends insight to that level but to gain the bigger picture, relationships between the levels must be examined also. For this particular research paper, the level of the individual will be focussed on. However within a single level descriptive studies may be done that “examine the nature of the variables at that level” (Niederman et al., 2006a, p.133). It must be noted that in some studies the role of an individual is typically that of a developer and not a user (Niederman et al., 2006a). Research is easier when focussing on developers rather than users, however using a diffusion and adoption model allows for consideration of other research topics as given in Niederman et al. (2006a). These research topics include such areas as examination of patterns of the number of users changing over time and what decision making process does an individual undertake when deciding between proprietary or open source software (Niederman et al., 2006a).
With a level of analysis in mind it is then possible to use the second paper to determine which discipline theories from information systems are most relevant to the level in question. The theories that have been included in the paper by Niederman et al. (2006b) include adaptive structuration theory, agency theory, complexity theory, diffusion theory, game theory, social network theory and transaction cost theory. Each of the abovementioned theories was selected based on “the potential for examining open source issues and their existing base of application within the MIS literature”(Niederman et al., 2006b, p.152). The theory selected for use in this research paper is diffusion theory which was defined by Rogers (1995) and the suitability of using this theory to analyse open source has been explained by Niederman at al. (2006b). While there have been no studies from classical diffusion literature, the concept of innovation has enabled IS researchers to adopt alternate methods of examining innovation (Niederman et al., 2006b). From the literature differentiation in communication of innovations occurs between open source and so called typical approaches; where in open source software the communication travels in a bottom-up style from the users towards the developers (Niederman et al., 2006b).
In attempt to gain an understanding of the topic in light of the theoretical framework, a similar case study will be used and comparisons will be drawn from the case study and research topic. This will aim to show relevance of the research topic to the theoretical framework and allow areas of further investigation to be identified. The theoretical framework section contains a description of the diffusion of innovation framework by Rogers (1995) and the following method section demonstrates how the framework is used to interpret a case study from a university environment.
- Theoretical Framework
The theoretical framework used in this research is Diffusion of Innovation (DOI) theory (Rogers, 1995). This theory was selected due to the ability to assess both individual and prior condition variables which affect the innovation decision and to also trace factors that affect adoption or rejection through the four phases shown in the model. A more in depth description of this model follows.
Other models that were considered but later rejected include the Technology Acceptance Model (Szajna, 1996), and DeLone and McLean’s Information Systems Success Model (DeLone and McLean, 1992). The reasons for rejection of these two models are given in the following paragraphs.
The Technology Acceptance Model investigates user acceptance of information systems based on perceptions about ease of use and usefulness then later the actual ease of use and usefulness (Szajna, 1996). While user acceptance is related to the research topic, information systems specifically nor simply focussing on the ease of use and usefulness encompasses the full scope of the study.
The DeLone and McLean Information Systems Success Model was considered due to its user satisfaction components and how net benefits affect user satisfaction which in turn drives use of the system (DeLone and McLean, 1992). While some aspects of the model may be relevant to the research topic, the focus of the model is not on what factors drive adoption. This model may be useful in future studies that examine how successful open source software is among end-users following adoption.
- Innovation-Decision Process Model
The Innovation-Decision Process Model shows the process an individual goes through when deciding whether to accept or reject an innovation as found by diffusion scholars (Rogers, 1995, Rogers, 2006). A diagram of the stages in the Innovation-Decision process can be found in Figure 1 has been based on the model found in Rogers (1995, pg.163).
- Figure 1. A Model of Stages in the Innovation-Decision Process (Rogers, 2006)
This model was deemed appropriate for the research as the focus is on an individual end-user and the factors that play a role in determining whether that individual will accept or reject an innovation (Rogers, 1995). Despite the model encompassing all processes of the innovation acceptance/rejection decision, only the first and second processes will be directly applicable to the research topic.
Factors from the model that are expected to coincide with factors found in the literature are the prior conditions, characteristics of the decision-making unit and perceived characteristics of the innovation (Rogers, 1995).
The research question of this paper is what factors from the literature that can be used to explain end-user adoption or rejection of open source software. Using an aspect from the multi-level framework it is possible to focus the research in such a way to gain maximum benefit and relevance (Niederman et al., 2006a). For this research the individual level has been selected as the primary focus however Niederman et al. (2006) do stress that to gain a complete understanding of open source, the entire multi-level framework is required. This ensures that all subtleties that occur between the levels are acknowledged and the effects and influences each level has on the others is understood (Niederman et al., 2006a).
The article by Sullivan (2001) presented a list of factors that have been personally and socially identified describing issues and concerns about open source software, specifically the Linux operating system. The most important issue from Sullivan’s (2001) article is that Linux is presently quite difficult and complex for new users, in particular as most users converting to Linux are likely to be Windows users. The paper by Hussein (2003) discussed research that was conducted among students and presented results describing actual use statistics and issues that have negatively affected the adoption of open source software by students.
The common factor between both papers is the importance of ease of use, in both articles this is actual ease of use as participants have actually used open source software and were not merely stating their perceptions. The factor about aesthetics while not mentioned by Sullivan (2001) may play an important part to other end-users who are not so technically minded. Whereas some of the more specialist issues raised by Sullivan (2001), such as Windows networking and cohesion of vision and standards, may not be mentioned if a variety of end-users are surveyed. Whereas an end-user survey revealed that the lack of end-user support and aesthetic features of open source software were genuine issues raised by end-users (Hussein, 2003, Sullivan, 2001).
The theoretical framework selected is deemed to be appropriate as factors identified in the literature are able to be mapped onto the Model of Stages in the Innovation-Decision Process (Rogers, 1995). This is demonstrated in the factors associated with usability and how they can be attributed to characteristics of the decision-making unit, in this example, the end-user(Rogers, 1995). These characteristics form part of the first stage in the model, knowledge. Usability is also linked to complexity, which is a perceived characteristic of the innovation in the persuasion stage of the model (Rogers, 1995). Based on feedback from end-user research complexity is a key issue of the persuasion stage where individuals are receptive to the innovation and will use particular characteristics to evaluate the innovation prior to moving to the decision stage (Rogers, 1995). Of the existing attributes within the persuasion stage of the DOI model additional attributes have been identified in the study of technology diffusion (Niederman et al., 2006b).
These attributes include “critical mass, cost and social approval” (Niederman et al., 2006b, p.160). Consideration of the additional attributes is important for this research as issues of cost and social approval may prove to play a larger part than previously anticipated. Factors such as social approval may depend heavily on the software aesthetics for it to be deemed acceptable by society. For each of the stages within the DOI model, each must be examined using an individual basis for analysis. In purely considering an individual, certain elements of the DOI framework may become more or less important to that particular level from the multi-level framework. Due to the complexity of comprehending multiple layers of analysis Niederman et al. (2006b) recommended that research of the impact of features and complexity be done for specific open source software applications on the consequences of implementing software in varying contexts. By cumulating several separate studies it will be possible for researchers to understand successful adoption of open source software and the preconditions required for that success (Niederman et al., 2006b).
Based on the literature review and analysis of the findings it can be concluded that the research topic investigating end-user adoption of open source software is important and interesting. Due to the lack of academic quality literature there is much room for additional research to be conducted. Research by Niederman et al. (2006a, 2006b) has highlighted the levels in which can be used to study open source software and have provided seven discipline theories which may be used to research and analyse findings based on open source software. The theoretical framework selected for this research is one that has been identified as being suitable for further research in this topic area.
This research has identified the primary factor that is important to end-users when deciding to adopt open source software as a new innovation. This factor is related to complexity and ease of use from a end-user perspective and has been identified following analysis of a case study and also a critique article. Open source software is a relatively new innovation and research on this topic is valuable for not only developers and current end-users of open source software but also users who are considering the opportunities and benefits open source software may allow them.
This research has demonstrated that identification of factors influencing the adoption decisions of end-users is a topic that is worth pursuing. Diffusion of the open source innovation is growing and developing a more thorough and complete analysis of factors affecting the rate of diffusion will provide the means to increase growth. The publication of research by Niederman et al (2006a, 2006b) demonstrates that research in this area is worth pursuing and that by using a number of existing theories, each of the levels within the multi-level framework in addition to the relationships between levels may be fully understood. Further research in this domain will increase understanding and also clarify and expand on several approaches to investigating open source software.
Dedrick, J. & West, J. (2003) An Exploratory Study into Open Source Platform Adoption. HBS - MIT Sloan Free/Open Source Software Conference: New Models of Software Development. Boston.
Evans, D. S. & Reddy, B. J. (2003) Government Preferences for Promoting Open Source Software: A Solution in Search of a Problem. Michigan Telecommunications and Technology Law Review, 9,2, 313-394.
Hussein, N. (2003) The Acceptance of Open Source Among Computer Science Students In Universiti Sains Malaysia. Free & Open Source Software Conference 2003. Selangor, Malaysia.
Niederman, F., Davis, A., Greiner, M. E., Wynn, D. & York, P. T. (2006a) A Research Agenda for Studying Open Source I: A Multi-Level Framework. Communications of the Association for Information Systems, 18, 129-149.
Niederman, F., Davis, A., Greiner, M. E., Wynn, D. & York, P. T. (2006b) Research Agenda for Studying Open Source II: View Through the Lens of Referent Discipline Theories. Communications of the Association for Information Systems, 18, 150-175. Open Source Initiative (2006) The Open Source Definition version 1.9. Open Source Initiative. Rogers, E. M. (1995) Diffusion of innovations, Free Press New York. Rogers, E. M. (2006) Chapter 5: The Innovation-Decision Process, Diffusion of Innovations: Part I. Good Ideas do not Sell themselves.
Skidmore, D. (2005) The future of software as a business artefact. IN SCOTTO, M. & SUCCI, G. (Eds.) First International Conference on Open Source Systems. Genova. Sullivan, P. (2001) What’s Holding Linux Back? Anandtech.com. Szajna, B. (1996) Empirical Evaluation of the Revised Technology Acceptance Model. Management Science, 42,1, 85-92.
Wang, H. & Wang, C. (2001) Open Source Adoption: A Status Report. IEEE Software. Conference Information
This paper will be submitted to the Open Source Software miniconf which will be taking place as part of LinuxConf 07. LinuxConf 07 will be taking place in Sydney at the University of New South Wales on January 15-20 2007. The call for papers for the main conference has closed however the call for papers for the miniconf I’m applying to have not closed. The LinuxConf 07 website can be found at http://lca2007.linux.org.au/.
Presentation guidelines for the miniconf are as follows:
“Proposals for miniconfs can be up to 2000 words, and should detail the community involved, the expected number of attendees, proposed activities and detail any support you would like from the organising committee. This proposal needs to convince us that you can organise an interesting, successful miniconf that people are going to want to go to!” (http://lca2007.linux.org.au/cfp)
As there was no clear template given for the miniconf I’m submitting to, I’ve used the template given as part of the paper submission deadlines for the 15th European Conference on Information Systems (ECIS). The website for ECIS can be found at http://www.ecis2007.ch/.