Open Source is dead, long live Sustainable Software

I'm clocking in eleven years of working with open source this year. Despite coming a long way in those years, it still appears to defy reason that the (often and not always) astronomically higher price/performance proposition of open source has not made it ubiquitous and the default choice. I don't believe in an open source software monoculture, but there is still too many examples of organisations acquiring expensive and inferior non-open source software when the technical and financial characteristics of the open source option were superior. The vast majority of explanations for this continued phenomenon run along the lines of "vendor/format lock-in", "FUD campaign", "back room dealing", and such. All of these explanations may not be wrong, but they share the interesting characteristic of being external.

They are external to open source. External factors are ones that are outside one's direct control. If I miss my bus, I may externalise and say that the red light on the preceding pedestrian crossing held me up. If I were to internalise, I would say I left home too late. By internalising, I give myself a chance to control the situation. When we identify situations where for example vendor lockin prevents open source from competing, we are giving someone else control of rectifying the situation; the vendor administering the lock-in perhaps, or regulators.

If we introspect a little and look for internal factors to the open source movement which might hinder the greater penetration of open source software, a clear factor emerges; at least as far as the enterprise market is concerned (public and private alike). We have been promoting a non-functional characteristic instead of an end benefit. It turns out that open source doesn't actually translate immediately to any perceived benefit in the mind of the users, and software freedom is a cause when people just want to hear about effects.

It's clear that there are markets for products whose primary differentiating feature is non-functional. There is a market for organic produce just as there is an explosion of consumer activity when Apple decides it's time for the same product, only in white. Neither market is very big as a proportion of the whole however, and in both cases the demand is reliant on buyer motivations other than the tried-and-tested self-interest.

And promoting software as "open source" does not speak very clearly or directly to self-interest because it is a description of the software rather than an explanation of how this characteristic benefits the buyer. There's some scattered evidence of disorganised and directionless marketing of open source by its benefits, but it's quite ineffective and talks about "less bugs through peer review" and so on. I know, I've been guilty of it myself. It isn't good enough and it usually comes defensively after the primary "choose it because it's open source" pitch has inevitably failed.

In the enterprise context, we need to stop talking about how hot open source is and start explaining what she knows how to do in bed. And when we do that, we need to stay away from the technical differentiators between open source and its non-open alternatives. Not only because any significant technical advantage held by open source can be - given tie - replicated in non-open software, but more importantly because open source software has one enormous business benefit which is intrinsic to the difference between open and not open.

Open source systems are far more sustainable.

The way IT is today, any non-trivial component of an enterprise architecture will be obsolete and therefore a candidate for replacement within 5 to 7 or 8 years after its launch. This fact is not something non-open source vendors draw attention to since it is a foundation of their business model; the upgrade cycle generates license sales. Open source is not immune to the churn of technology either; languages and platforms change, and open source systems fall in and out of favor. Reinvestment in the same software functionality is inevitable regardless of the type of license attached to the software.

Since reinvestment is inevitable, we might as well make it as cheap as possible. Reinvestment in the same functionality here is equivalent to the cost of switching, and it is here that open source has a "by definition" advantage. There are no proprietary interfaces or protocols so chances of competitive bidding on the components is higher. The initial investment in open source is lower, making the decision to throw away the old software and acquire replacement software at a similarly low cost of acquisition is a comparatively easier decision to make.

I'm only scratching the surface here; there are other factors clearly demonstrating that choosing open source improves the long term flexibility, economics, and viability of the overall IT architecture. That is something we can call "sustainability". Software which poses lower long-term switching risk, both technical and economical is sustainable software.

And software sustainability is a message that segues very easily into business benefits since it uses an economic term "sustainability" instead of talking about something as cryptic as "source code".

We might have been using the wrong key message.