Essentially, there are two ways of producing software. One is the ”Silicon Valley” way, where you start up as an entrepreneur in your garage, reach out to venture capitalists (VCs) who then invest into your idea, and later scale, if you’re successful - like Facebook. The second is a community way, where numerous members inspect, modify and enhance the applications that have openly available code bases. They gradually improve and add on features as the project goes along, like Linux.
The Silicon Valley ethos
The fist way, popularised by Silicon Valley ethos, follows the logic of explosive growth. Here the startups are funded by the VCs, who make their profits on the verge of creation and disruption. They look for companies who develop a product compelling enough to make customers switch to it rapidly from existing technologies or services. This rapid conversion rate gets VCs their maximum cut. Take the example of Amazon or Skype, the adoption rate went through the roof simply because there was no other service they had to compete with. Quite naturally, the investments in Silicon Valley model are always concentrated in rather narrow industry segments that promise large enough potential markets to compensate for the risks. In other words, only those software production businesses who promise growth, scalability and rapid and large payoffs are going to have the funding. This is of course great news for the new Googles and Facebooks. An obvious downside of this ethos, however, is that some initiatives will never get any funding if this model was the only one in town.
Design by community
Small markets that cannot grow sufficiently large to ensure capital gains, and are not interesting to VCs, even though they also can be highly profitable. These “nice growth firms”, as human ecologist Martin Kenney defines them in his study of the evolution of capitalism as a social and economic system, mainly work in two areas. First is medical and scientific services, such as medical device research and production clusters. For example, Jena, a German company which works in optics, or a US-based SDP, doing cancer research. The second area is consulting and design - for example, Ideo and frog design, the two US design firms which now has gone multinational. SAS, a North Carolina State University project, the most widely used analytics software. Such companies never receive any venture capital, since they don’t demonstrate enough rapid growth and can take a long time to “incubate”, essentially, growing internally, with the rate of quality adding up much faster than scaling.
““Nice growth” firms take a long time to “incubate”, essentially, growing internally, with the rate of quality adding up much faster than scaling”
Why open source can do better
Here’s where open source comes in. Open source is software with source code that anyone can inspect, modify, and enhance. Initially, despite the bright prospects, open source has gained ill fame because of the quality of code and lack of framework. Fair enough, you could get ready-made code modules for free, but they were usually created as solutions for rather specific ends by individuals or groups who never had resources to test and maintain them. Thus, oftentimes it was easier to write your modules, rather than trying to debug and essentially re-build hairy code written by someone else. Now, however, the situation is changing. GitHub has now evolved into a well-functioning social network focussed around code reviews. The increased use of pull requests as production ceremony has increased the accountability of contributing individuals. Moreover, open source now starting to adopt regular production procedures like automated testing, containerisation, infrastructure as code and CI/CD (continuous integration and continuous development). This means that not only the source code itself but whole architectures can be designed, developed, tested and deployed as community endeavours. Communities around open source are now becoming more diverse, with IT specialists from many walks of life participating to contribute with their skills - DevOps, Q&A, as well as product owners, bringing open source closer to product vision and focussing on the delivery of value. The asynchronous nature of production pipelines promises the potential of being structured enough to be able to deliver modules or even whole features - like Elastic search, - that have no less value than those developed by “Silicon Valley” model companies.
“Open source could even do better than the proprietary model, since the inspection and debugging of it is not limited by the capacities of the team that the company hires to do the project ”
The round-up
There is a mismatch between the business logic of explosive growth or so-called “Silicon Valley ethos”, and communal approach to nurturing high technical skill and complex problem-solving. The former is driven by the capitalist logic of maximising returns on investment via rapid expansion. The latter previously muddled by the lack of tooling around asynchronous workflows, testing and strategy, now has the potential of producing comparable quality products. The “closed source” ethos is appropriating the software, making it impossible to copy, inspect and alter the code. Given the current state of the industry, this could arguably be seen as detrimental to the code itself, since these procedures are always limited by the capacities of the team that the company hires to do the project. In open source, there is no such limitation. Thus, arguably, open source could even do better.
