Open source, the internet and national R&D organization: Where to?

THEME #2: Internet science and technology for the 21st century

Johan Eksteen (jeksteen@csir.co.za)
CSIR Mikomtek
South Africa

Carey-Ann Jackson (cjackson@csir.co.za)
CSIR Mikomtek
South Africa

Abstract

Open source is a phenomenon of interest across a wide range of disciplines. Simplistically, it seems to be largely of interest in the Information Technology domain, where the application is obvious. One can, however, argue that the open source phenomenon can be seen as a symptom of larger sociological and organizational changes that might be accelerated by driving forces like the Internet. One cannot examine open source in isolation, but must rather one must look at issues such as the creation and ownership of knowledge, the issues of virtual and real roles and identities, economic drivers and other not-technical effects. These issues have an important impact on organizations and institutions. But, we believe, these issues are especially important for institutionalized research and development organizations, particularly if those organizations are based in developing countries.

In this paper, some of the "traditional" IT issues of open source will be discussed. This will be done in such a way as to illustrate the argument of open source as a symptom of larger forces. We will also demonstrate some of the principles through a case study of the CSIR, a research and development organization in South Africa. Our paper does not attempt to provide specific answers to all of the issues raised, but is aimed primarily at stimulating debate around these issues.

Table of Contents

Introduction

In this paper, we set out to stimulate debate on the importance and potential of open-source and its potential impact upon the organization for which we work. Our paper is essentially a non-technical one but focuses on a phenomenon often relegated to technical journals. Our reason for this is deliberate in that we wish to avoid spending time bickering about the relative technical merits of, for example, Microsoft NT versus Linux, and hence, want to avoid falling into the trap of detracting attention away from more important questions about the open-source phenomenon.

We wish instead to focus on the fundamental effect that the Internet has on the underlying proceses of knowledge creation and use and its effects on, for example, software development practices and business models for software product development. While we are interested in the identification of personal drivers and their potential to attract and retain above-average developers and researchers in the organization in which we work, we do not focus on these issues in this paper.

Our paper's primary intent is to stimulate thought, debate and possible action around interesting and emerging knowledge and research practices, of which the open-source and Linux phenomenon is a symptom. We do this from the perspective of a case-study of a research and development organization in which we examine the reasons for why the open-source phenomenon is important to our organization's continued development and transforming role in South Africa. Our paper begins with some general comments about the major challenges for the open-source movement. We follow this with a consideration of how these challenges are experienced in our particular organization, and as a result, why it is that we, as an organization, need to take notice of the open-source phenomenon and what it represents.

Major challenges faced by the open-source community

One argument often heard in discussions around open-source is that the open-source movement lacks a common and coherent vision and strategy. We believe that while some perceptions about open-source community members may be valid, the claim that open source vision is either lacking or is anti-Microsoft, is simplistic. The ideology and dynamics of the open-source community have been explored in several articles and white papers. A better known effort is Eric Raymond's thought-provoking paper, "The Cathedral and the Bazaar"[1]. Another is cWare's white paper, "The Linux Storm"[2], which provides important insights into the dynamics of the open-source phenomenon, and specifically how open-source relates to the Microsoft factor. The cWare white paper likens the dynamics of the open-source community to that of a storm, where several factors determine the strength and direction of the "storm". One of the enabling factors for the open-source movement is the Internet effect of accessible and low cost communication capability. The other one is the effect of the monolithic nature of Microsoft's sphere of influence, which is compared to the action of a "low pressure system" to fuel the "storm". It is thus too simplistic a view to state that the open-source vision is purely anti-Microsoft. That also fails to recognise that IBM and the mainframe community played a similar role in the early evolution of the open-source movement. If we continue with the storm analogy, it is important to recognise that weather patterns are global phenomena and cannot be studied in isolation of each other. A storm is only a local manifestation of other effects and the implication of this is that what we are dealing with here, goes beyond just open-source, but rather has wider sociological effects. In his article "Cyberpower: The Culture and Politics of Cyberspace"[3], Tim Jordan argues that perceptions of "us" versus "them" are symptoms of a far more complex cyber-sociological phenomenon. We consider this relevant to discussions of the open source community. If this is the case, then we need to question the need for a vision, because the "group" might not be a group or community. If we consider the fundamental changes in the software industry caused by, among others, the Internet effects, the idea of a centralized vision for a group might be in as much danger as the current software and product development principles. Our preferred approach is to focus strongly on solutions and then choose the appropriate technologies to support those solutions. And this, in turn, may call into question the value or longevity of visions such as products like Windows NT from Microsoft.

Another common concern with open source is how best to leverage advantage. It seems difficult for most organizations to understand how to leverage open-source projects, such as Linux, to their advantage. Most businesses are happy when they make money. But the open-source phenomenon poses organizations with a dilemma: By giving away one's intellectual property (the source code in this case), how do you make money? Or in business management terms: "Where are the business models for the open-source phenomenon?" In the January / February 1999 issue of IEEE Software, Frank Hecker from Netscape examined some issues and business models in "Setting up Shop: The Business of Open-Source software". Another source for possible business models can be found from OpenSource.Org[4]. It is clear that adding open-source projects to current practice is not ideal, and calls into question current practice. This obviously makes several organizations uncomfortable.

This brings us to a major consideration: culture. And by "culture" we refer to both organizational culture and the open-source community culture. It is widely accepted that "hackers" tend to be individualistic and, often, fanatical. Concerns with traditional development practices and ease-of-use issues are, apparently, considered of lesser importance to the hacker community. This potentially brings them into direct conflict with established management structures and principles within organizations. This has had negative impact on widely held perceptions of open-source community members. For example, it is increasingly difficult for non-technical people to understand the value of certain technologies or products when these are looked at through confrontational glasses or when the fanaticism filters are too well implemented. This frequently has the consequence that sound advice from an early adopter of the technology is disregarded. In turn, this negative response appears to force the "hacker" to become more fanatical as there is a perception that their contributions are ignored or undervalued. Similarly, the counter-perception that decision makers readily succumb to expensive marketing efforts by commercial companies and do not trust the insights of their own people is not easily dispelled.

Part of the cultural aspects of open-source is driven by considerations with and for money. Ownership and support issues (including aspects such as support personnel, licensing, and so on) are also reasons for why open source has not been as widely adopted as it could have. We frequently see comparisons of the real cost of NT vs. Linux but the mistake often made in these comparisons is a failure to consider the scaling factor. If one compares two stand alone systems, the comparison appears to favour the commercial route. When one scales across several systems, the picture looks quite different. This is especially true for systems where the license fees are determined by the number of users (See [5] for a recent case where a company realized substantial savings through the use of open-source projects).

Securing support for systems such as Linux are thought to be very expensive because support personnel are scarce. If one looks at the remuneration packages demanded by MCSE and NCE qualified professionals, this claim rings false. The growing number of open-source developers and users, especially from the academic environment, need to be absorbed by the industry at some stage. With a growing interest in open-source systems, the scarcity of skills might not be such a major issue in the near future.

Another challenge associated with open source systems, from an organization's point of view, is that if the system falls over, it is not possible to complain at a big company's door. This is a valid concern but it is possible that a false sense of security can be created. Consider the following: Most calls to major help desks originate from misconfiguration and not necessarily from bugs in the product. These misconfiguration issues are often resolved painlessly. When bugs are present in a product it is often the case that the issue is only addressed in the next version of the product. This would be fine if Linus Torvald's philosophy of "Release early, release often" is followed but this is, however, often not the case. If one considers the major growth in installed base of the Apache web server - and as highlighted in the Netcraft Survey[6] - this factor might become less of an issue.

A concern has also been expressed with the fact that developers may change pieces of software to make things work and when the software must be maintained, these developers are nowhere to be found. This is a valid concern but we do need to realise that people who install Linux, for example, do not necessarily compile their own kernels?

Sustainability of open-source projects is, further, a complex and challenging issue for the open-source community. We believe the issue of sustainability raises questions about what is the driving force behind the development and the developers. Several articles are available that explore this issue. Without going into the details, quite a few of the issues addressed Eric Raymond's "Homesteading the Noosphere"[7] are parallel to experiences within our organization.

In the (in)famous Halloween Document[8] the biggest stumbling block for the open-source community is highlighted as the reliance on open standards. According to the document, de-commoditizing protocols and standards will be very effective to prevent third parties (who do not license technologies) from gaining market share. In the official reply to the "Open Source Memo Regarding the Open Source Model and Linux"[9], this point was defended under the banner of "the need to do it to add value to the customer". This is a powerful motivator for large sections of the corporate decision makers. And potentially, this, more than anything else, might deliver a deathblow to the open-source community.

On the surface then, it appears as if the open-source community faces an uphill battle when one considers all these and other issues. Quite a few of these issues are only perceived threats and it is not possible to consider them without reference to the opportunities offered by the open-source movement.

In the remaining section of our paper, we wish to examine these opportunities with reference to the experience of our own organization. We believe that the opportunities provided by the open-source movement extend beyond a Pro-Linux fanaticism and in fact offer benefits into transformations of the what and the how of institutionalised research and innovation.

Case-study

Background

Before proceeding, we need to briefly describe our own organization in order for our debate to be appropriately contextualized. The CSIR is a national organization mandated by the South African government to undertake scientific research and development in the interests of the South African community. Given that one of our key foci is on information and communications technology research and development, and as a science council, the CSIR is required to play a transformative and development role in South Africa's IT industry.

Why do we need to take notice of the open-source phenomenon?

We believe that we need to take notice of the open-source phenomenon for several reasons. The existence of the open-source movement is obviously topical but more importantly it touches on many aspects that are (or should be) of importance to the CSIR, namely:

  1. The implications for product development practices, especially software development;
  2. The interesting implications and commentary on current business models and future alternatives;
  3. The implications for the establishment and sustainability of a local software industry;
  4. A commentary on corporate culture;
  5. Considerations of aspects of human resource management and reward systems; and
  6. The influences upon IT Architecture, including issues such as standardization and business system selection.

Our paper cannot examine all of these issues in detail but we do attempt to highlight some important points in the hope that this will stimulate a healthy debate within and beyond our organization.

The first point we flagged was about the implications for product development practices, generally, and software development, specifically. The open-source community enjoys successes despite the apparent lack of formal project management procedures and software engineering principles. That success is important for an organization doing development work (not exclusively in the software realm) to understand the success drivers employed in the open-source community. This understanding might lead to innovation in the ways in which we approach and execute our projects.

Many of the developments in the open-source community can be classified as equivalent to "Gold Rush" development efforts. Steve McConnell, the editor of the January/February 1999 issue of IEEE Software, took a critical look at the high-risk, high-reward development issues in his editorial "After the Gold Rush" (pp. 6-8). When one reads this article against the backdrop of statements such as "We (the CSIR) undertake projects for or with customers to reduce the technical risks" or "We undertake high-risk, high-reward projects", one begins to see how this type of development is of relevance to our organization: If one recognises that the CSIR has a role to play in establishing or supporting a local software development industry, we believe that we need to understand the success factors of open-source development. This must also be understood in context of our current understanding of development and engineering principles. We might be spending an unnecessary amount of time on debugging and testing of programmes. If a way could be found to make use of the broader Internet community to test our code for us, it might deliver not only savings, but also provide more development time. Time is an important factor here. The reality in the developing world is that we need to drive innovation from a time perspective. In other words: "Innovate faster!" If approached in the correct way, open-source methodologies might deliver on this need.

We believe that the open-source phenomenon, as a methodology, has implications for how we conduct research and development generally. We believe that the open-source methodology is based on principles similar - if not equivalent - those of participatory research principles. Our current, more institutionalized research and development practices follow more formalized, empirical and measurable step-wise processes. In a word: positivism. This point in our argument may require some elucidation as it addresses fundamentally the issue of knowledge and knowledge practice - a topic that does not often sit comfortably with the gung-ho "just do it" approaches of technical people.

It is a truism to state that global economic changes are powerfully present in society, and demand continuously that we examine what role and function organizations, such as universities or science councils, have in society. We are all aware that there are profound changes taking place but many us are not able to pinpoint exactly what those changes are. Information technologies are regarded to be simultaneously products and hasteners of this alleged change. One of the most profound changes which is occurring is at the level of knowledge creation and use. Knowledge, according to post-structuralists, has become a commodity, something which can be controlled and sold so that those who have knowledge are able to achieve more effectively and more efficiently their market edge. This is a critically important realisation because what we experience today is a fundamental shift - facilitated in part by the Internet - of what is considered - reliable knowledge - and who / what can know that knowledge. Hence, we are all forced to ask questions about what knowledge systems are valid, what are institutions doing with that knowledge, and how are they do those things.

We believe that what we are looking at in the 'conflict' between the open-source community and institutionalised approaches to software development is symptomatic of a more profound struggle between new and traditional approaches to knowledge creation and use. Our existing and dominant scientific approaches are, as the philosophers of science will bear testimony, positivist. Positivism, in and of itself, is a dominant methodology but when combined with the economic models on which institutionalised research is based, we believe that the real battle between open-source methodology and institutionalised (positivist) methodology, is at its core, a battle of contesting knowledge paradigms.

Open-source approaches, in our opinion, suggest that IT knowledge creation and use can be based upon more participative approaches and methodologies. And while this may seem an odd claim because participative research approaches generally tend to be applied to rural water scheme development, we believe that while the participants may look slightly different, the dynamics of the open-source movement and how it creates its knowledge are strongly participative and hold out the promise for a more participative - and successful - scientific knowledge creation paradigm.

From the perspective of our organization, we are able to see evidence of a tug-of-war between these paradigms, and the people caught up within them, with the result that we see a degradation of the quality of output achieved. We believe that organizations such as ours should understand this tension in order to, firstly, manage within this understanding, and secondly, to ask whether or not we should be more participative or open to the outside world with how we do things? It also forces us to ask questions about whether or not our current practice is capable of promoting creativity and innovation in our existent development practices? And to ask ourselves challenging questions about examining cultural changes within our organization.

A second consequence of the open-source phenomenon is its effects on current business models and future alternatives. Current initiatives exploring the effect of the Internet on our business show the positive sign that the organization understands the need to revisit its current business models. We do not have a complete picture of the current thinking around this topic and it would be inappropriate to comment on specifics. But one aspect that might need attention is the leveraging of skills and experience in the open-source environment for marketing efforts. We have people that are important contributors to, among others, the FreeBSD project. We will not be able to sell FreeBSD as is, but options such as commercial support might be ways to benefit. The open-source phenomenon might provide valuable input into the lived experience of tension that we deal with between commercial successes on the one hand and research/impact measures on the other.

As we mentioned earlier, the CSIR has an important role to play in the establishment and sustainability of a local software industry. The software industry is one of the industries that has the lowest establishment costs (i.e. in terms of capital, training etc.) Several success stories are to be found from developing countries where concerted efforts to establish a local software industry paid off handsomely. The Indian[10] and Israel examples have been quoted repeatedly. And recently, the local success story of Thawte[11] highlights our argument. Recent surveys suggest that most of the local IT industry focuses on reselling important technology. This is clearly not a sound approach when one looks at the outflow of capital that accompanies this trend. The Department of Arts, Culture, Science and Technology (DACST) has recognizes the shortcomings of this trend and has tried to rectify this trend providing funds through its Innovation Fund, specifically through the "Advanced Software Development" part of the Information Society section. This issue was examined inn some detail in an article which appeared in the Business Day of 22 July 1999, "Age of cybercolonialism has dawned"[12] by Adam Gordon, which made two important comments:

  1. "A more aggressive and creative way for developing states to respond would be to take the IT revolution bull by the horns and to invest heavily in telecommunications technology and subsidise low-cost access to it, making telecommuting a possible and reasonable alternative to migration for the aspirant IT professional."
  2. "Much professional IT work - programming, systems design, software development and so on - is fairly easily done at a distance and there is already precedent for this. In Bangalore, India, and Haifa, Israel, for example, private and government infrastructure allows IT workers to routinely log on for work in the US."

When we looked earlier at some of the drivers for the open-source movement, and in particular "The Linux Storm" [2], we believe that parallels emerge. The telecommunications infrastructure as driver for virtual development teams should not be underestimated. Mib Software's effort of developing an on-line catalogue of open-source projects [13] shows that "two guys in a garage" do have access to five million of code. This is an important factor when entering existing markets and for creating new technologies. We believe that we, the CSIR, could play a major role in the local software industry by learning from the open-source community and ploughing this learning into the local industry. We, as an organization, have the ability to be a driver or play a coordination role in the industry by exploiting this learning.

The relative lack of success in trying to get commitment from major commercial firms such as Microsoft and Oracle to establish local development centres is a concern. We ask whether or not we should take this as a sign to invest in open-source models to achieve better the effect of the development of a local software industry? Given the fact that the major corporations' presence in South Africa is mostly in the form of marketing arms, the CSIR does appear to have a local advantage. We believe that this type of approach should also not be limited to software only. If one considers the efforts of the Open Telecom Group [14], one can probably exploit the availability of technologies to solve problems locally and even export the results. It is easy to pay lip service to this, but a lot is said about one's commitment to this when we look at our corporate culture.

The open-source phenomenon, we believe, does have a potentially positive influence on our corporate culture and aspects of human resource management and reward systems. We group these two issues together because they relate specifically to the CSIR context and because the questions we ask are specific to our organization. Why do we wince when we find out that a number of our fireballs protecting our precious information are based on an open-source operating system (FreeBSD)? Are we making ourselves guilty of discouraging the use of open-source development tools for our corporate systems? Are we trapped into thinking that we would not be able to support products based on open-source, such as Linux and Apache? Do we, through selection of business tools, send the message that we do not support open standards? Do we prefer the opinions from outside entities such as the GartnerGroup over the opinion of our own people? Do we encourage personal development through involvement in hobbies that might be beneficial to our work, such as open-source development? Do we settle for less appropriate technology options because one can attend a course on the maintenance of the system? Do we recognize and reward people who implement and maintain systems that might become important for our business? What message do we send to people by arguing savings on overheads when discontinuing systems, but discounting the personal growth and experience gained through the implementation and maintenance of a system? These are but a few questions we might need to ask ourselves if we want to understand the effect of the open-source phenomenon on our business.

Finally, we believe that the influence of the open-source movement on IT Architecture, including issues such as standardization and business system selection, is potentially enhanced. Again, we have chosen to ask questions throughout this sub-section. Standardization has many important benefits. Standardization also has one or two drawbacks. Standardization in networked communities or corporations might open vulnerabilities that might not have existed if systems were more diversified. Why is it that I cannot synchronize my PDA's diary with our existing scheduling system, but I can do so even from Microsoft's Exchange. Simply stated, support for the open vCalendar protocol is built into most modern scheduling systems. This is not a comment on the appropriateness of GroupWise, but a symptom of a perceived problem. Do we select business products based on their support for open standards?

Conclusion

In this paper, we have highlighted several issues for the open-source movement, generally, and for our organization, specifically. If we are to summarise the message of our paper then it would be this one: "Understand open-source and it's impact. It is important. It also goes beyond software."

We have argued that the open-source phenomenon holds the possibility for a more fundamental change in how knowledge is created and used in the scientific and technical community, and as a consequence thereof, it holds the potential to transform how and why organizations such as the CSIR conduct research and innovation.

Many issues and statements made here might be highly contentious and emotional but we have deliberately chosen it to be that way. We believe that our organization needs to debate these issues and get to grips with them, particularly, as they will impact on what role we will have - and how we will execute that role - within developing South Africa's indigenous IT market.

References

  1. Eric Raymond, The Cathedral and the Bazaar http://www.tuxedo.org/~esr/writings/cathedral-bazaar/
  2. "The Linux Storm" http://www.cwareco.com/linux_storm.html
  3. Tim Jordan,"Cyberpower: The Culture and Politics of Cyberspace" in the Proceedings of the Internet Society INET99, 1999, http://www.isoc.org/inet99/proceedings/3i/3i_1.htm
  4. OpenSource.Org Home Page http://www.opensource.org
  5. http://www.idg.net/go.cgi?id=144039
  6. Netcraft Survey,http://www.netcraft.com/
  7. Eric Raymond,"Homesteading the Noosphere" http://www.tuxedo.org/~esr/writings/homesteading/homesteading.html
  8. Halloween Document http://www.opensource.org/halloween/
  9. Open Source Memo Regarding the Open Source Model and Linux"http://www.microsoft.com/NTServer/nts/news/mwarv/linuxresp.asp
  10. India turning around its economic fortunes http://www.bday.co.za/99/0705/world/w9.htm
  11. VERISIGN ACQUIRES THAWTE CONSULTINGhttp://www.verisign.com/press/1999/corporate/1220_thawte.html
  12. Adam Gordon, "Age of cybercolonialism has dawned",Business Day,22 July 1999http://www.bday.co.za/99/0722/comment/e6.htm
  13. Mib Software's on-line catalogue of open-source projects http://www.mibsoftware.com/reuse/
  14. Open Telecom Grouphttp://www.opentelecom.org/