What is the Value of Open Source Software?

A recent study came out from Harvard Business School that valued open source software at $9 trillion! They reached this amount by assuming that each company would have to recreate its own version of the open source software. One could argue that if this were the case, there would likely be companies selling licensed versions of the software which would result in some cost savings. But regardless of this quantitative value of open source software, there is a qualitative component that wasn’t accounted for in this study and is much more impactful than simply saving money: innovation. 

Without open source software, many of the incredible technological advances we have witnessed over the past 25 years would not have happened. I myself am constantly amazed by the reach of open source software. The AI boom we are seeing has not been fueled by closed source, proprietary software but by open source tools like Python, PyTorch, and linear algebra software. Almost anything with a computer in it uses open source libraries: your car, phone, tablet, the infotainment system on airplanes. Even my Garmin watch contains open source libraries. Two of the most popular operating systems, Microsoft Windows and Apple OS, incorporate open source tools. The list is endless. Give it a try next time you get in your car – search around the menus on the screen and look for the licenses. Open source software is everywhere and it fuels innovation – its value is priceless. 

Taking the Plunge: Adopting Open Source Technology

When my colleagues and I founded Kitware in 1998, we built the company around the open source Visualization Toolkit (VTK) library. The original business plan was to build licensed software on top of VTK. However, after a few years, we discovered that our licensed software competed with our open source offerings. We had customers who were paying us to expand the open source visualization software but were not eager to pay a license fee. We had a choice to make. We could either embrace the open source model and abandon the software licensing or we would need to license all of our software at the risk of losing the relationships we already established with our customers. We saw the value of open source software and weren’t willing to walk away from that. So, we embraced the open source model and haven’t looked back since! It’s amazing to see after 26 years how that one decision made us who we are today. I am proud that Kitware continues to provide quality, state-of-the-art open source technology to the benefit of our customers.

A Glimpse Into How It Happened

One of the very first contracts that Kitware landed was a government-funded program to create a new C++-based toolkit for medical image segmentation and registration. Kitware was part of a team that included three commercial companies and three universities, in a multi-year, large-scale collaboration. Together, we developed the Insight Toolkit (ITK), and the impact was widespread.

A New Way to Advance Science and Research

Having released the Visible Human dataset in 1994, the National Library of Medicine realized that giving researchers access to this data was only the first step towards accelerating innovation. Rather than the standard academic approach of a professor handing a paper to a grad student and having them reimplement an algorithm, the idea was to have both the data and a core software framework and implementation made publicly accessible, enabling researchers to rapidly build on each other’s work. It became clear that open source software would be the crucial component, and that’s when our team was formed. The development of ITK meant that researchers could focus on the next big ideas instead of reimplementation. It meant that students could delve into these complex medical algorithms to better understand the code and could now reproduce results, confirm validity, and ensure scientific reproducibility.

Powerful Coding Tools

The ITK project also became the birthplace of CMake. Our team was tasked with creating portable C++ that would work on Unix, Linux, Windows, and Mac. Back then, the state-of-the-art build system was autotools for Unix/Linux, checked in Visual Studio 6 project files, and then hoping someone on the team could get it to work on a Mac. CMake was developed to bridge these gaps and give grad students and others an easier way to access the ITK C++ library. CMake allowed them to build small applications that linked to the robust tools in ITK. Little did we know that over the next 24 years, CMake would become the de-facto standard for building C++. Kitware’s servers provide approximately 100 million CMake downloads a year, and this doesn’t count the distribution of CMake through other channels (e.g. Linux distributions). It is so popular that “CMake experience” is listed as a requirement for many developer job postings, with more than 900 positions listed on the recruiting website, Indeed, at Tesla Motors, DCS Corp, Mindsource, Quanergy, etc., and more than 600 positions posted on LinkedIn by Samsung, Johnson Controls, Apple, Uber, Toyota, Microsoft, and others. 

Kitware’s Open Source DNA

Kitware was built on open source software to support critical scientific projects, and it remains at our core to this day. Much like trying to figure out the monetary value of open source software, it’s difficult to measure the magnitude of the scientific contributions we’ve made through our open source technology. But I imagine it’s quite sizable. Folks from Kitware attend large conferences like Super Computing, RSNA, and CVPR, we are always approached by thankful researchers who have been using our open source tools, including ParaView, VTK, and 3DSlicer to perform Ph.D. research or develop deployable products. When Kitware first started, open source technology was a fairly new concept, but it has grown in impact and size over the past 30 years. And we couldn’t be more proud to be part of that journey!