Open Source Software

worm

There is an interesting feeling of the place where the author of this article lives. It is a Dutch themed neighborhood, right next to a bridge. The houses have slight differences in design, but they are very standard with 10 feet times 10 feet and 12 feet times 12 feet rooms.

One would imagine that the veterans of Operation Market Garden designed them to their children in the 1960s. The operation led by Dwight D. Eisenhower, Bernard Montgomery only partly succeeded. The parachute and airborne divisions “pulled” 22. They could not capture the last bridge of Arnhem running into two armory SS Panzer divisions of Tiger Panzers. The failure lengthened the war by months to an eventual victory in May 1945.

Software Engineering Managers face similar decisions. One of the most notable is whether to use or contribute to open source, or keep intellectual property confidential.

Any failure can cause losses by losing intellectual property, customers or patent lawsuits. Spending energy on new solutions may seem unnecessary compared to using an off the shelf solution.

Closed source solutions are generally quicker to build. There is less requirement to be standard, there are fewer rules.

Many advancements of Apple like the lightning connector are the result of the strength of closed source. You could plug it in from either directions compared to the original mini and micro USB connectors.

Engineers can do better decisions with fewer constraints. The solution is cheaper, and it takes less time to ship.

It is easier to change closed solutions in face of customer demand. Complex user interfaces can be closed sourced, increasing the barrier to entry for competition.

Keeping changes confidential poses a risk, a bond that is paid, if the sources are stolen. Staff training takes longer. Rigid rules may alienate talent. Teams will be stuck with engineers, who are paid to keep the codebase afloat raising fixed costs.

Open source software comes from a different direction. It allows sharing a simple codebase by the industry. This decreases the industry specific risks.

You can hire staff from the free markets, who are already mastering a toolkit. A bigger community makes solutions last longer. Linux was founded almost thirty years ago. Investing in the knowledge pays off for students and engineers.

Companies may decide to distribute open source on top of just using it. Such decisions lower risks for investors. Any hostile takeovers, bankruptcy or management decisions allow the codebase to survive.

The stability will lure more customers and a community of engineers. This will cause more scrutiny and less potential bugs. Investors will be keen on investing more with better terms.

Example:

$1000 spent on closed source may generate revenues of $300 for four years with a profit of $200 overall. Competition may copy the user interface, but they can create a smaller codebase, so that they can make it obsolete early.

$1000 spent on open source will trigger a discussion to improve the codebase. The competition may allow the company to charge only $60, but for ten years, making the company less profitable. If two more competitors use the codebase, the project will generate $1800 with $300 profits.

The companies still gain by making their support costs less, if three of them divide the initial investment. The documentation is available to the community. Users can find solutions themselves instead of paying high cost of proprietary support professionals.

It makes sense for open source companies to invest less early on as a result. Hadoop a popular clustering software was started by a single developer, Doug Cutting.

A hybrid approach is probably the less risky. Even smaller startups can write and release open source, but they should invest only a little. They can leverage the resulting transparency and buzz as marketing instead of R&D. The community gives an important feedback on profitability early on. This can help investment decisions. Open source likely improves the survival ratio of VC seed and early stage investments.

Any differentiation on the other hand can stay closed making a valuable intellectual property to compete with. The company can still hire knowledgeable engineers from the market, who can easily get familiar with the proprietary changes. Product market fit can be achieved faster by listening to clients. The company can still gain strong momentum with additional R&D.

Problem Every industry will eventually develop an open source alternative.

Solution Patents of the original inventors eventually expire. A successful product will bring attention and new graduates will eventually invest in an open source solution. An open source alternative will rise, if the closed solution uses the confidentiality to keep prices high. The closed source alternative paid off the initial investment by this time, so they will make it cheaper. The resulting similarity in pricing and value add will lure more customers that helps the industry. The open source alternative will likely be simple and cheap as well.

Problem Companies will keep their differentiation closed.

Solution It is pointless eventually to replicate open source closed. Keeping secrets is expensive. Differentiation generates additional revenue and income. Companies will retain the legal rights as a result. They do not need to deal with additional unnecessary feedback that likely comes from competitors. This helps rational decision-making.

Open source

It is available like classical music, religious books, or common science.
OS has low entry barrier having professionals that are trained already.
OS is an accepted standard.
It is lower risk to compete with an open source standard.
OS creates free markets sparking a flourishing vertical supply chain.
OS has high value, if it is kept streamlined with fewer bugs and security issues.

The most successful cloud companies like Amazon Web Services and Microsoft Azure leveraged open source Linux and Kubernetes, but retained their extra R&D. Less successful rivals like Google kept much of their Cloud Application logic closed, even if it was ahead of its time.