Roundtable Discussion Part 2: Exploring "Scaling Strategies" for OSS Suited to Japan

Building on the discussion in Part 1, I’d like to shift the conversation to specific “strategies” from here on. Regarding the benefits of the government releasing and utilizing OSS, “reducing procurement costs” has tended to be emphasized so far, but how do those on the front lines actually perceive the situation?

To be honest, the argument that “using OSS reduces costs” is prone to misunderstanding and is a dangerous trap. Even if the license fee is free, there are still costs for maintenance and implementation. Instead, we should focus on “reducing social costs.” By using a common framework, we can eliminate the waste of local governments and ministries by building disparate systems and ensure data integration and interoperability. Consequently, the true ROI (return on investment) for the government in adopting OSS lies in reducing coordination and development costs for society as a whole.

So, the perspective isn’t “saving our organization’s budget,” but rather “reducing waste across society as a whole.”

Exactly. Additionally, for engineers and staff on the front lines, more direct motivations are also important. It comes down to whether they have firsthand experience thinking, “This open-source code really helped me” or “It made my work easier.” Personally, I have a strong firsthand experience of being able to launch a business by utilizing data and tools that the Geospatial Information Authority of Japan had openly shared. It is precisely this “sense of having benefited” that generates the motivation to contribute to the community in turn and becomes the driving force that keeps the ecosystem running.

I agree. There are two major benefits to promoting OSS. One is contributing to the development of systems for society as a whole, and the other is improving the efficiency of the administrative systems themselves. Especially as we look to the future, where the use of AI is expected to further reduce system development costs, I believe the significance of the government utilizing OSS lies not so much in “cost reduction” for administrative system development, but rather in “standardization” across society as a whole. For example, if data models required for tax payments are released as OSS, vendors will find it easier to build systems that comply with them, making it easier to build private-sector services that integrate with government systems. Data integration will also become smoother. This “ease of connection” is the greatest value.
Another important perspective, I think, is “who the users of OSS are." Who will actually use the published OSS to build systems? If a government agency has a strong in-house development team, OSS can function directly as a means of sharing code within the government. However, many Japanese government organizations lack internal development capabilities.

If there are no in-house engineers, there’s a concern that even if the software is released, no one will use it.

Exactly. That is precisely why, in Japan’s case, I believe we need to design with the possibility in mind that vendors contracted for system development will use it. Vendors use the OSS released by the government, and systems incorporating that OSS are delivered as government systems. Only when this cycle is in motion does the true value of utilizing OSS emerge. The question is whether we can look beyond simply saying, “We’ve released it,” and envision the vendor ecosystem that follows.

As Mr. Yoshida pointed out, many government agencies in Japan do not have powerful in-house engineering teams numbering in the hundreds, as is common in Europe and the United States (particularly the U.S. and the U.K.). Given Japan’s overwhelming lack of resources, how should we scale OSS?

Given the current situation in Japan, it is not realistic for the government to develop all systems in-house. One possible strategy would be for the government to focus on providing a common foundation—such as data models and core computational logic—while entrusting the implementation and widespread adoption (scaling) of these as actual services to an ecosystem of private-sector vendors.

I see. So, in other words, the government sows the “seeds,” and the private sector takes on the role of nurturing them and reaping the harvest.

Yes. For example, the government provides the “seed” (OSS) by stating, “This is the standard data model and calculation logic for procedures.” Multiple vendors then utilize this to create operational systems for local governments or SaaS platforms for applications. If the base is the same “seed,” data integration becomes easier even between local governments using Vendor A’s system and those using Vendor B’s system. Since a GovTech startup ecosystem is also growing in Japan, wouldn’t a strategy of approaching that ecosystem to achieve scale indirectly be appropriate?

That ties in with the discussion of “native varieties” from Part 1. Just as with vegetable breeding, if the government creates and releases good “seeds (varieties),” farmers (vendors) can cultivate them to produce delicious crops (services) and distribute them to the market. Conversely, breeding improvements (suggestions for enhancement) come from the farmers on the ground in a bottom-up manner. Creating this cycle may be the practical solution for resource-constrained Japan.

Exactly. That’s precisely why OSS released by the government doesn’t need to be a “finished application.” In fact, “components” such as user-friendly libraries, validation (input check) logic, and data schemas are often easier for private companies to integrate and provide greater value. As private-sector development workflows are changing, making libraries available to access government-defined standard data formats is crucial for accelerating data circulation across the entire industry.

Another major challenge is “sustainability.” Since government employees typically rotate every two to three years, there is always the risk of “person-dependent” issues—where a project grinds to a halt or loses momentum the moment the person in charge changes. How can we overcome this?

To ensure a project doesn’t “die out” even when the person in charge changes, we definitely need a system in place. However, while the EU and others are able to publish and maintain hundreds of repositories, why does the fear that “it’s over once the person in charge leaves” take precedence in Japan? This is partly a matter of mindset, but it’s also a matter of institutional design.

I believe the solution lies in integrating the system with the “institution” itself. For example, systems related to e-Tax or My Number don’t just disappear because the person in charge has been transferred, do they? That’s because those systems are embedded as an inseparable part of the social infrastructure. I think it’s possible to position future OSS as a means of “Rules as Code.”

What exactly does “Rules as Code” mean?

Until now, regulations and specifications have been written in natural language. However, natural language is open to interpretation depending on the reader. As a result, different vendors interpreted specifications differently, leading to problems where systems couldn’t interoperate. Therefore, we define the “correct answer” (standard model) for administrative procedures as unambiguous programming code and release it as OSS. By doing so, OSS becomes more than just software; it functions as a “regulation free from interpretive ambiguity.”

I see. If systems become part of the social infrastructure, maintenance will no longer be a “personal hobby” but an “administrative obligation.” This means we can move beyond individual dependency and achieve sustainability.

Yes. In a society facing population decline, creating an efficient digital society with limited resources will require an approach that reduces the coordination costs associated with natural languages such as aligning interpretations and handling inquiries—and instead describes society through a code base. This concept is similar to the “Techno-legal framework” (the integration of technology and legal systems) advocated by countries like India. Shouldn't Japan aim for that as well?

In reality, the calculation logic for social security is extremely complex, and it’s an area where vendors struggle with requirements definition every time. If the government were to provide this as a “rule engine” in code, it would dramatically improve efficiency and eliminate errors. If open-source software (OSS) were released to realize “Rules as Code,” it would ensure sustainability and provide the greatest support to the private sector. Furthermore, to ensure sustainability, investment in the “system” itself is essential. While it has become common for private companies to establish specialized organizations called “OSPO (Open Source Program Office),” we should also establish a “Government OSPO” within the public sector.

So, the idea is to give the organization the function of managing and supporting OSS.

Yes. For example, Germany has created a mechanism called the “Sovereign Tech Fund,” through which the government provides funding to support important OSS projects. I believe Japan also needs to reach a consensus to establish specialized organizations and funding mechanisms—even if it means using tax revenue—to support OSS that forms the foundation of our society, rather than simply “releasing it and calling it a day.”

Finally, this survey report is merely the first step in “periodic monitoring.” How would the data and situation in Japan need to change in a year, or a few years, for us to call it a “success”?

I’d like to see changes in “quality,” not just simple numbers. For example, specialized organizations like an OSPO being established within government agencies, or “release as OSS” being standardly included in procurement specifications. I hope to see a “robust ecosystem” taking shape, where these mechanisms become firmly established and private companies utilize them to expand their businesses.

I’d like to aim for a future where the “Rules as Code” concept we discussed today becomes widespread, and OSS is routinely utilized in national and local government systems. Of course, I believe human interpretation will still be necessary for parts of the system that require ethical judgment. However, considering that a world is coming where AI agents will handle procedures on behalf of humans, I think the importance of this way of thinking will only continue to grow. It may take time, but I believe the landscape will change as pioneering players like the Digital Agency, the Ministry of Land, Infrastructure, Transport and Tourism, and the Geospatial Information Authority of Japan take the lead and accumulate successful case studies (firsthand experience).

This survey revealed that Japan possesses a unique strength in its “native species.” We must nurture this without letting it wither, while leveraging the power of the private sector to scale it up. We intend to continue our efforts with the goal of seeing this “Japanese model” take a clearer form and be reflected in the data by 2027. Thank you very much for your time today.