Seeing that a chapter of a period that once filled a limited number of individuals curious about computers, different operating systems, and technological developments with great joy and hope is coming to a close, I realized that the time had come to publish an article I had penned in the past; but continuously postponed its publication due to an excessive sense of meticulousness or perfectionism. The initial version of the article, written on June 6, 2010, was essentially a “policy document” directed towards an operating system that aimed to enter the market and prove itself. However, it became evident that, at this point, a text featuring different versions but with an unchanged beginning and illustrating the saying “start like a Turk, finish like a German/English” – which I do not particularly like – was correct once again.
There are three main characters in this topic: First, the “National Operating System” developed by TÜBİTAK (The Scientific and Technological Research Council of Turkey) in Turkey. The operating system, a GNU/Linux distribution, is named PARDUS[1], derived from the scientific name of the extinct Anatolian leopard, Panthera pardus tulliana. The roots of our National Operating System Pardus as a project can be traced back to 2003, and it’s evident that the project bore its first fruit in February 2005, after the project team was formed in 2004 to determine the roadmap and start work. Following this period, Pardus continued to develop steadily and with increasing momentum, and in 2010, it was released with the name “Pardus 2009.2 Geronticus eremita” in memory of our endangered species, “Bald Ibis.” Pardus also released two corporate versions in 2007 and 2011.
In 2011, following the release of Pardus versions “Dama dama” (Fallow Deer) and “Cervus elaphus” (Red Deer), the management and development team left the Pardus project. To evaluate the current state of the Pardus operating system from different perspectives and with a wide range of stakeholders and to create a roadmap for future development and sustainability, the “Pardus Workshop for the Future” [2] was organized on March 23-24, 2012, and a final report was prepared[3].
The second character in this story is the commercial operating system installed on approximately 90% of personal computers worldwide, which predates the existence of the “panther.” We will briefly refer to it as “Windows.”
As for the third character mentioned in the first paragraph, we are the ones who haven’t given up our characteristics mentioned in the idiom above.
The purpose of this work is to examine the Vision of Pardus[4] and its development process from the perspective of “innovation economics and policies” and to evaluate the fundamental question: “What could (or couldn’t) have been done for the development of Pardus?”
However, at the initial stage, the following assumptions should be considered:
- While the essay compares two competing operating systems, it doesn’t make any judgments about one being superior or recommending one over the other. In other words, it is observed that countries like Germany and Brazil have either created their operating systems or turned to an open-source operating system, at least in government and educational institutions, due to national security concerns or costs. However, whether a state should have a national operating system for reasons of national security, cost, or otherwise is beyond the scope of this essay. The essay focuses on the processes following a state’s decision to take this direction for “a reason.”
- This essay aims to point out the glitches in the process of implementing the decision made by the state and how the process can function more effectively to achieve its goals.
- Since this is a policy document and does not claim to be a scientific study, unofficial opinions on user habits, preferences, motivations, etc., and experiences gained within the normal course of life are considered. It would be useful to test these in advance experimentally in the case of a policy draft.
I. Innovation Processes and Software
It is possible to say that innovation in any field fundamentally goes through three stages. These stages are, in order, (a) the creation of innovation, (b) the introduction of innovation for use, and (c) the widespread adoption of innovation.
The primary goal of the innovation creation process is to develop a new product, introduce a new production technique, open up new markets, discover new resources like raw materials, or establish new organizational forms. In this context, innovation can also be seen as bringing existing products together in a novel way to serve a new purpose. Therefore, innovation involves intensive research and development processes, along with associated costs.
The second stage involves the commercialization and release of innovation. This includes transitioning from a prototype to mass production or replacing a software’s beta version with a stable version for users. The production and market release phase signals both the competence of the new product and a shift in cost functions. This is because production costs replace research and development costs, and there are specific costs associated with each product brought to the market. In traditional markets, these costs decrease initially but increase after reaching a certain scale. Here, the software industry differs significantly from manufacturing. The cost of duplicating and delivering a software product to users is almost negligible once the final version is produced.
The third stage is the dissemination and adoption of innovation. Initially, an application or process may be adopted by a limited number of users. Over time, it will become more widely adopted or disappear from the market. When looking at this phase in terms of time, it often follows an “S”-shaped curve. In this context, it means that in the initial stages after the launch of the innovation, it is used by fewer people. The number of users increases significantly at a certain point, and at some stage – which generally corresponds to the product becoming “outdated” or “out of fashion” as new products are introduced – the number of users starts to decline [5].
There is also a relationship between adoption and the development and quality of the product. As innovation becomes widespread, the problems related to its use will increase. This, in turn, gives the developers opportunities to improve the product, enhance its quality, reach a broader audience, and eventually develop new products. In recent times, one of the reasons companies have placed special emphasis on “listening” to consumer problems is because of this. Therefore, the potential for developing a product that is not widespread is also limited.
From a software perspective, the principles mentioned above hold. In the dissemination phase of innovation, marketing, and sales costs are typically more prominent. However, unlike the previous stages, especially when dealing with a commercial product, producers not only bear the costs but also generate income. Therefore, it is a kind of “harvest time” following research, development, and production activities. Hence, it is equally important as the preceding two processes and is particularly necessary for the sustainability of the research and development activities in the first stage.
For the Pardus Project, this situation aligns with the statement in the Pardus Vision: “Although it does not pursue profit, it will be a sustainable organization with its usage and created business models and its own resources.”
What is crucial in terms of reaping the rewards is reaching the critical mass. Key factors for innovation dissemination and reaching the critical mass in a market include the following:
- Product Cost: When purchasing and trying out a new product, we all know there’s a risk that our money might go to waste. Therefore, offering free samples of products on the market or allowing users to test the software for a certain period before making a purchase is a marketing tactic to minimize this risk. Moreover, especially for software, the cost factor is evaluated differently for various users, as explained below, and can be a crucial factor in the widespread adoption of software.
- Habits / Ease of Use: This is a crucial factor, especially for individuals. Taking into account the psychological principle of “habituation,” we can understand that while improvements in speed and functionality in software are welcome, changes made to menus, shortcuts, and other interface elements slow us down during a certain period of adjustment. Therefore, radical changes are generally not well-received. For individuals, switching from one operating system to another is not easy, mainly because of this principle. To test this, place a proficient Windows user before a Linux distribution or a MacOS system they have not used before.
- Network Effect: Rather than expecting every computer to perform limited functions, the focus today is on computers understanding each other’s language, performing common tasks, and opening files we prepare and send in a way that the recipient can access easily without loss. When we look at markets that exhibit network industry characteristics, we can see that a GSM company with more subscribers or a bank with more branches is preferred over its competitors. Consequently, this imposes an extra effort on their competitors.
- Product Quality: Interestingly, what’s most common isn’t always the best. The best example is the QWERTY keyboard layout, which many use. Although alternative layouts that speed up typing have existed, the QWERTY layout, initially designed to slow down American typists, became the most widespread [6]. So, it is overly optimistic to assume that presenting the “best product” will automatically lead to its adoption and dissemination. Contrary to expectations, the most common product may eventually be relegated to the dustbin of history.
- Competition in the Market: Significant differences exist between a company/product entering a market for the first time and one entering a market where competitors already exist. A person or company introducing a product that has not been offered before generally competes with different (lower-level) solutions, and even if the product has various technical weaknesses, it can quickly win over many users and dominate the market. Examples include the transition from horse-drawn carriages to automobiles or the introduction of the first graphical user interface-based operating systems alongside command-line interfaces. On the other hand, when similar products are in the market and are widely used, the newcomer will have to compete not with lower-level solutions but with similar, more common solutions. Therefore, convincing existing users and reaching the critical mass becomes more challenging.
As can be understood from the explanations above, making the best mousetrap or the best operating system and distributing it for free alone will not be sufficient to reach the critical mass, especially for a product that enters the market later. Therefore, to bring a product to a self-sustaining level, these five factors must be evaluated correctly in their interaction with users, and strategies must be developed accordingly.
II. The Story of Panther and Windows from the Perspective of User Groups
Using what has been discussed so far, let’s assume that there are two operating systems: one called Panther and the other called Windows.
Windows has been in the market for many years and is a commercial operating system used by 90% of computer users. Behind Windows, there is a company with thousands of employees, a wide support network, and hundreds of software companies developing applications for Windows.
Panther, on the other hand, is a relatively new operating system developed by a public institution, still under development in some aspects, and claimed to be superior to Windows in terms of security and other aspects. Panther’s development involves a limited core team and volunteers, and there is no professional support available.
For the purpose of simplifying our analysis, let’s also assume that there are four key user groups in the relevant geographic area:
1. Large Enterprises: These include banks, telecommunications companies, large-scale manufacturing, distribution, and sales companies where information systems play a significant role in daily operations. For them, the most critical factors are uninterrupted operations and security, with cost being a secondary concern. Disruption of operations or a security breach carries the risk of causing much greater financial damage than software costs. What often influences the choice of operating system in these companies is the “switching costs.” Large companies may rely on applications closely tied to their operating systems for managing document flows, stocks, and other business processes. These applications can be more costly than the operating system licenses themselves. Changing the operating system would be a significant financial and operational risk. For this reason, large enterprises often prioritize using licensed software.
2. SMEs (Small and Medium-sized Enterprises): These businesses employ a relatively small number of people who don’t heavily rely on IT systems except for basic office and accounting tasks. For SMEs, cost is the most critical factor. They may not allocate many resources to software, sometimes resorting to using software illegally, and often rely on community knowledge for support. Unfortunately, using common, and sometimes pirated, software may be seen as the most cost-effective solution. Therefore, ubiquity rather than licensing costs is a crucial preference factor for SMEs.
3. Public Institutions: These are organizations with varying sizes but collectively employ thousands of workers. For these institutions, licensing costs, especially regarding the operating system, are not a significant concern. Some of them, unfortunately, prefer using widespread software without licenses and attempting to resolve issues internally. Others opt to use entirely licensed software, but they do not consider the cost of software, as they operate within the framework of the Friedman Matrix [7] (where the government’s money is used to purchase products for the public). For these organizations, there is no direct comparison between the investment in software and the benefits derived from its use, and since many top-level executives may not be well-informed on this matter, procurement and usage decisions are largely under the control of the IT department. From this perspective, the most crucial factor for IT department employees is that the software is capable of meeting their needs at any time and ensuring that business operations continue without disruption. For instance, when it comes to delivering services requested by a minister or undersecretary, the preference of the IT department manager is to purchase a product that offers the “best support services.” Consequently, even if the IT department employs individuals with limited capabilities, external support can help resolve issues. Regarding software acquisition, the focus is on support services, as reducing costs does not carry incentives within public units, but failing to meet the demands of executives can lead to consequences. Additionally, there is no hindrance to the proper allocation of public funds when it comes to spending in the budgetary process. Hence, support services are emphasized, and the preference leans toward an operating system that provides the best support rather than the least costly one. User and departmental preferences and whether proprietary software is used within the organization take precedence over costs.
4. Individuals: Individuals have varying preferences. Unless they are technically competent in IT, average individuals tend to continue their office habits at home, using the operating system that comes with their computer or the widely available one. Often, they don’t consider the cost because they either use work-issued laptops at home or use the same software that they use at the office on their home computers. Additionally, personal habits play a significant role.
Based on the factors mentioned so far [8], we can summarize the preference factors for operating systems in the following table, with a number indicating their influence (from high to low). Areas where the impact is unpredictable but considered to be low are left blank:
Considering these factors, it’s clear that for large enterprises and public institutions, the licensing cost of the operating system is not a primary concern. It’s more about “security, reliability, and support services.” Ubiquity and established habits also need to be considered. As for SMEs and individuals, the ineffectiveness of intellectual property rights enforcement leads to widespread unlicensed software use. For these groups, the licensing cost is not a top priority – ubiquity is more important.
In conclusion, given these user group preferences and considerations, Panther, as a new entrant in the game, must reach a critical mass of users on its own. Even if individual users increasingly prefer Panther, the choices of large enterprises and public institutions will be the determining factors. Cost is not the primary driver in their decisions, and the fact that Panther is free is not enough to change the game’s dynamics. Therefore, to achieve widespread adoption and development, Panther needs to develop a strategy that correctly assesses the interaction of these five elements with users and formulates strategies accordingly.
III. A Strategy Aligned with the Pardus Vision
In terms of the ranking of factors affecting the preference for an operating system and other applications regarding licensing costs, it is possible to categorize what can be done within a framework where the operating system and other applications do not rank high, while “security, reliability, and support services” and the prevalence factor are at the forefront. Generally, what TÜBİTAK UEKAE can do and what the government can do can be divided into two categories.
What the Government Can Do – as we all somewhat estimate – can be summarized as follows:
1. Follow a more effective approach to protecting intellectual property rights: This way, the value of labor is preserved, and the real costs of software are revealed for individuals, private companies, and public institutions.
2. Compel applications, especially custom software solutions used by public institutions, to work independently of the operating system. In the long term, government tenders should be structured to consider different operating systems. This would create a competitive environment between solution providers other than the common operating system and Linux distributions, fostering competition even among Linux distributions.
3. Implement curriculum and education programs in formal education that ensure students learn Pardus with the same proficiency as the common operating system. While steps have been taken in this direction [10], the implementation needs to be more widespread [11].
4. Indeed, when looking at success stories, especially in countries like Brazil [12], it is evident that the government has made radical decisions not only in terms of creating software but also in terms of widespread use. Thousands of computers in schools and public institutions were quickly migrated to the Linux operating system. In this sense, the proliferation has a “revolutionary” rather than an “evolutionary” character, depending on the determined use of public power.
5. From the perspective of TÜBİTAK UEKAE, support should be provided to plan and implement a new business model that will make a breakthrough in user support. Although positive steps were taken in this direction in 2008, such as the “Pardus Migration Partners” and “Pardus Training Centers” [13], it cannot be said that sufficient progress has been made over time.
One of the biggest reasons for this could be that these migration centers may be focused on SMEs. However, as mentioned earlier, when it comes to the prioritization of SMEs, factors such as Pardus not being widespread and there being fewer individuals to solve Pardus-related issues make the cost of support after switching to Pardus higher compared to the support provided by “neighbors” who are familiar with the common operating system, even if the operating system and other software are free.
A similar dilemma applies to training centers: As long as Pardus is not widespread, it is not reasonable to expect individuals to endure the time cost, even if training with Pardus is free. Individuals will prefer training for the common operating system, which offers more job opportunities during the same time frame. As long as there is not enough demand, Pardus Training Centers cannot operate, and the rate of Pardus adoption will not increase. Furthermore, the content, examination system, and certification of the education to be provided are not yet clear.
Therefore, for TÜBİTAK UEKAE, the priority should primarily be to bring large institutional customers, especially public institutions, into the Pardus family. However, as mentioned before, from the perspective of operating system preference in a public institution, an IT manager in a decisive position (i) will consider support services as the most important preference factor to meet the demands of the management as quickly and easily as possible. They will not want to use a system that does not provide support or provides limited support; (ii) will not want to force users who are already accustomed to the common operating system to use Pardus by making its usage mandatory; (iii) will not have any reason to make IT employees who are familiar with applications running on the common operating system, who have mastered hardware and network problems, learn a new operating system, and will not want to risk disruptions in these areas. Given that reducing licensing costs is not rewarded for the IT manager, as it does not bring any return but may instead become a threat to their career, waiting for public institutions to “spontaneously” switch to Pardus is unrealistic. Additionally, even if some employees want to use Pardus, IT departments may not allow it due to possible system incompatibility.
Hence, the most critical factor for public institutions is “support services.” Considering that TUBITAK does not have the resources to cover this with its limited resources, Pardus Migration Partners (PGOs) should be structured to serve public institutions 24/7, in addition to SMEs. They should turn this into a profitable business, report problems and solutions to the Pardus Center, and contribute to the spread and development of Pardus. However, it is essential to note that as most institutions use the common operating system and purchase software and services through tenders, it is almost impossible for PGOs to work with public institutions. It should also be remembered that each tender implies an investment in the common operating system for the relevant public institution for many years. Therefore, the opportunity for Pardus to enter these areas is lost for a long time.
Therefore, regarding Pardus’ development, TÜBİTAK needs to consider a systematic “marketing” strategy. Within this framework, the following can be done:
1. Conduct a survey to inventory the hardware and software used by public institutions. This survey should include questions about the number of desktop/laptop computers, the number of operating system versions by users, the number and version of office applications, the number and version of security and server applications, and the license renewal period, among other things. This survey will provide the first step in finding suitable institutions for pilot projects.
2. Create a “Migration Evaluation Survey” prepared by TÜBİTAK to evaluate the benefits and losses of institutions in terms of software, hardware, security, and financing if they switch to Pardus. It will help understand whether there is special software that will negatively affect the migration project. In this context, a scenario might be envisaged: The evaluation survey can be put online on the Pardus page and, with support from a database of software and hardware, relevant individuals or unit managers can enter their software and hardware information and costs after selecting the best Pardus packages for each item. They can also check if Pardus will support certain hardware. However, since IT departments may not be willing to fill out such a survey voluntarily, incentives should be provided.
3. Teams established within TÜBİTAK should visit institutions close to the licensing period and capable of creating a success story. They should conduct qualification studies related to the migration to Pardus, present a migration plan, match the institution with a migration partner if necessary, and ensure continuous support with TÜBİTAK guarantee. Naturally, the most critical aspect here is to persuade top managers of the institutions. Explaining the benefits of this process, such as cost savings and developing mechanisms that reward institutions, will be beneficial in this context.
Some of those who have patiently read this far might think, “Make a good operating system, offer it for free, hire staff to encourage institutions to use it, and wander after institutions; this doesn’t sound like a profitable business.” However, as mentioned before, Pardus cannot sustain itself and accelerate its development until it reaches a certain number of users, the critical threshold. To achieve the vision of “a sustainable organization that does not pursue profit and can sustain itself with its own resources through business models that will be used and created,” it does not seem possible for Pardus to reach the critical threshold until a “Pardus migration and support services market” can be created for public institutions and private companies, which can compete with other operating systems. Indeed, the current status of Pardus confirms this.
V. Conclusion
From an innovation theory perspective, especially for software that enters the market later, being better or free is not enough to reach a critical threshold regarding user numbers. As long as this critical threshold is not reached, it does not seem possible for the software to accelerate its development, stand on its own feet, and maintain its existence in the long term.
Therefore, just as important as developing the software is actively pursuing a strategy for the widespread adoption of the software and allocating resources to its “marketing,” even if it is not commercial. Marketing requires approaching target audiences with an appropriate strategy. From the perspective of large companies, SMEs, public institutions, and individuals – especially in a country where intellectual property rights are not adequately utilized – it will not be sufficient for Pardus to be less costly and reliable.
Although steps have been taken toward the widespread adoption of Pardus, it is believed that these will be insufficient in the long term. To reach the critical threshold, it is necessary to focus systematically on the migration of public institutions rather than SMEs and individuals and to take steps in this field that go beyond the conventional approach.
Of course, if there is still a will for Pardus to be the national operating system, and if we want to complete the task that started in Turkey, this time disproving the phrase “like a Turk”…
* This article contains recommendations for the development of the Pardus Operating System. Still, it should not be interpreted as stating that Pardus, or any other operating system, is good or bad or that one of them should be favored. The author is grateful to Assistant Professor Dr. A. Talha YALTA for carefully reading and providing feedback on the draft of this article.
** Barış Ekdi: Ph.D. Student in Middle East Technical University Science and Technology Policy Studies Program, Civil Servant. The views and recommendations in this article are not binding on any institution or organization. To contact the author, you can use the contact information on the website www.barisekdi.name.
References:
- http://en.wikipedia.org/wiki/Pardus_(operating_system)
- http://www.pardus.org.tr
- http://www.pardus.org.tr/pardusun-yarini-calistayi-sonuc-metni/
- http://www.pardus.org.tr/hakkinda/
- http://en.wikipedia.org/wiki/Diffusion_of_innovations
- http://home.earthlink.net/~dcrehr/whyqwert.html
- (a) If you buy something for yourself with your own money, you’ll try to get the best quality at the lowest price; (b) If you’re buying something for someone else with your own money, you’ll care about the price, and quality will come second; (c) If you’re buying something for yourself with someone else’s money, you won’t care about the price, but you’ll look at quality; (d) If you’re buying something for someone else with someone else’s money, the relationship between quality and price is disconnected, and you may not care about either or be extravagant (http://en.wikiquote.org/wiki/Milton_Friedman) – Public expenditures are generally of this nature.
- So far, these are purely hypothetical and not based on surveys or empirical studies. Nevertheless, it is thought that it reflects the reality of software usage in our country.
- Security, in this context, refers to the absence of security vulnerabilities in the system, while reliability refers to the system working smoothly.
- See http://pardus.meb.gov.tr/index.php
- These three items may be subject to separate regulations to make computer models with and without the operating system mandatory, encouraging users to choose the cheaper operating system-free computer. However, it should not be expected to be practically beneficial, as the major operating system manufacturer can offer its product to computer manufacturers for free, bridging the price gap. On the other hand, if Pardus is installed on computers, it may be expected that users who consider the cost of applications outside of the operating system would prefer Pardus. In this case, persuading major computer manufacturers to accept this will be a separate cost, and public institutions do not have the means to cover it.
- Detailed information is available at http://en.wikipedia.org/wiki/Linux_adoption
- http://www.ozgurlukicin.com/haber/bir-yol-hikyesi-bilisim-08/