Looking at the recent events in open source give us some interesting clues for the future. Especially as 2018 was one of the most important years for open source since its foundation.
More than a few of the peak moments are new shoots that suggest open source will grow in interesting and unpredictable ways. While also extending (and replenishing) its roots as a stable form of both enterprise and community innovation.
These are also the topics that pop up over and over again. At the start of 2019 I wrote that I was going to begin researching the future of open source. In the interviews conducted in the two months since I keep hearing the same themes.
Everyone seems deeply concerned with at least one of these major streams. So I thought it would be worthwhile to dig into the top ten themes that keep emerging — and ask what the major events of open source in 2018 might mean for the year (and years) ahead. Let’s go!
Salesforce announced in March of 2018 that it was acquiring Mulesoft for $6.5B. Then it was Adobe’s turn to announce it was acquiring Magento for $1.68B. And then SUSE got sold off to EQT for $2.5B. That’s quite a few billions right there.
If that wasn’t enough to get the mechanical keyboards clacking on Hacker News, along came the IBM to announce a $34B acquisition of Red Hat. In and around those larger numbers are acquisitions and acquihires of comparatively smaller open source companies. The news of CoreOS’s acquisition in early 2018 already feels like a millennia ago.
All of this activity creates a positive sentiment for open source — if not outright pricing the market with a premium in the process. Much like Red Hat’s record IPO kicked off way back in 1999, a rising tide floats all boats. Although we might just ignore the years 2000 to 2002 that followed.
Speaking of IPOs, there were some notable activity there too. Red Hat has held the record of largest open source IPO for nearly 20 years — with a not insignificant $3.4B. Right up until Pivotal’s IPO in 2018 set a new high of $3.9B.
What’s interesting here is that this is a whole other era. Red Hat’s cofounder Bob Young once explained to me the hysteria around their IPO back in the day. Sure it was a great company with solid numbers — but it was the first open source brand to raise venture capital and IPO. They were a symbol of an entire new vertical opening up. It’s not hard to imagine that mix of positive sentiment from within the open source world mixing with the sense of discovering new frontiers (and greed) from without. The market responded accordingly.
More capital is ostensibly a good thing. Free software being written in free time is great. But having a steady wage already aligned with the open source mission is almost immeasurably important to the overall ecosystem.
Likewise increased capital means greater ability to acquire necessary technology. While Linux gets the historic mindshare for open source, it’s impossible to imagine the open source landscape of the last decade without without middleware, virtualisation, or storage. Some of these IP investments might not quite be the revolution as intended (hello Big Data) but others such as containerisation are rewriting the rulebooks in realtime.
The acquisition of open source companies in 2018 was a big deal. And so was the acquisition of the platform that open source communities are increasingly relying on. We saw three expected things happen when Microsoft announced a $7.5B acquisition of GitHub.
The first was the immediate knee jerk from parts of the community that were concerned that evil Microsoft was getting its hands on the home of so much of open source. This is a valid concern but it didn’t really happen.
This is probably for reasons that have a lot to do with Microsoft bought them in the first place — GitHub isn’t just a repository but a workflow and a community. As we say in the SaaS ecosystem it’s a very sticky product. And with all those developers stuck fast, Microsoft swooped in to announce its first nefarious move after the acquisition was… to make private repositories free.
And finally came the investment announcements from GitLab (following the opportunity of an acquisition pricing and energising the competitive market). Which is interesting because GitLab isn’t just a competing code repository but a product aimed at the entire DevOps lifecycle. And open source. Watch this space.
This is some of the few bits of good news in the topic of sustainability in open source. An issue that Nadia Eghbal called “the internet’s biggest blind spot” back in 2016. Nadia talks articulately on the topic, and her “lemonade stand” is a handy guide to financial support for open source.
Nadia’s work stands out due to her experience across multiple points of view, ranging from VC to a part of GitHub for a time. My own view back at the SustainOSS summit, is summed up by my project submission in the report. Others like Tidelift are taking an enterprise approach to community resourcing.
Which isn’t as common as one might think. Xavier Damman of Open Collective wrote a great piece explaining the personal and brand incentives for supporting a project. And of course there’s no shortage of well-meaning tip jar platforms, patronage and bug bounties. All better than nothing.
But the example of Tidelift is one that echoes the rise of open source in the enterprise itself. If it wasn’t for Red Hat’s ability to elevate Linux into a trusted enterprise relationship… would the industry be where it is today? Is Red Hat’s example of $3B revenue from the enterprise realm a better goal than rattling three billion tip jars?
Whatever the ultimate mix of methods that maintains the open source world, there’s little question that the venture capital product line benefits from it (but doesn’t support it), and that the answer will likely be to actively engage in the market versus seeking handouts from it. The search continues.
Open source is not always a very nice place to work. Mailing lists and even conferences can be frustrating and exhausting things to encounter. Dedicated contributors even from within the culture can be burned out and turned off — let alone curious or casual explorers. Especially those from outside of the genesis cultural norms.
And a large part of this comes from the first wave of open source and free software identities that forged its image in an era (and as an act) of counter-culture. Such an emerging culture requires the forging of an identity. And such a culture built around a dogmatic battle for righteousness usually does so and against a clear villain.
And yet here we are in an era where the war is over, open source has won, and it’s just maybe a little less okay to be a jerk to one another.
Something that Linus Torvalds admitted when in September of 2018 he announced he was going to “take time off and get some assistance on how to
understand people’s emotions and respond appropriately”.
As one might imagine there were feelings about this. But the world didn’t stop and ultimately…
In the aftermath of the announcement from Torvalds emerged a new Code of Conduct. The reactions were somewhere in the range of behaviour expected. We live in interesting political and social times and the conversation reflects those extremes.
It’s a topic that deserves more careful nuance than a summary list on a blog post. But one inseparable from prevailing politics, and perhaps political fatigue, in a manner that will be influencing and affecting the open source culture for some time to come.
At face value the issue seems simple enough. Cloud vendors being accused of profiting from open source projects as they fill out their product offerings and battle for dominance. Amazon in particular.
Retold with less dramatic emphasis, it is understandable if companies based on an open source development model get annoyed when the likes of Amazon use their codebase without either paying or contributing. Especially when the increasing monopoly in the cloud space is putting pressure on those companies running the upstream communities.
Under that lens it would make sense to make changes to the open source license to stop the freeloading. Which is what Redis did in 2018. Kind of. The move wasn’t particularly popular among open source supporters. Not at all.
Another perspective offered was that this feeling has always existed with open source. Others point out that licenses applied to specifically block out a cloud vendor is just one commercial company using an open source community against another. And that the problem isn’t AWS itself — it’s open source revenue models.
Especially considering that Redis just closed a $60M Series E funding round — with the likes of Goldman Sachs Private Capital Investing and Bain Capital Ventures participating. Not organisations necessarily known for passionate open source advocacy.
All of these perspectives have an interesting point of view. The nuance of this issue makes for a fascinating case study of the open source ecosystem itself. Such as…
The interesting thing about companies like Redis adjusting their position on open source licensing isn’t just the ongoing tide of conversation around it. It’s the serious work of self-reflection at the core of open source itself.
In this case the Affirmation of the Open Source Definition released by the Open Source Initiative. It’s a fascinating read. If you’ve ever wondered what the point of difference between free software and open source is, there’s your primer. Specifically on the topics of “trust” and “commerce”.
The interesting ripples in this issue are a growing awareness that open source is evolving. There will come a time where the current licenses are updated to match. But as the OSI indicates, the very existence of open source as a commercially successful enabler of collaborative creation relies on trust and the ability to satisfy personal goals within the greater scope of the whole.
It’s not that cloud vendors are freeloaders (although they’re arguably contributing more than anyone else right now). It’s just that freeloading isn’t a bug of open source — it’s actually a feature. Although one spread unevenly.
The headlines about Microsoft open sourcing 60,000 of its patents would be pretty confusing for someone waking up from a coma. Microsoft giving it all to open source? This Microsoft being run by this guy?
Thankfully things have changed at Microsoft (and not just the share price). But not so much that they’ve given away patents. The nuance takes a little more digging to understand exactly what happened.
The implications of Microsoft joining the Open Invention Network means that Microsoft is making a public (and legal) agreement “not to use any of its 60,000 patents against other OIN participants for their use or distribution of ‘Linux System’ technologies”.
The interesting thing about Microsoft is that it actually is “listening to customers and developers and is firmly committed to Linux and other open source programs”. That’s Corporate Vice President and Chief IP Counsel Erich Andersen as quoted from the press releasing announcing the news.
And that’s a Microsoft embracing open source as it chases Amazon’s cloud leadership, refocusing after decades of singing and dancing outside of the actual major innovation shifts, and even convincingly getting behind startups. Not in the “oh wow free credits” way most programs run — but in a focus on co-selling.
Microsoft might have been the ideal villain for the open source community to find its strength and voice fighting against. But now it looks like they’re taking an impressive leadership role.
Software ate the world. And open source is the backbone is the entire software development experience. The definition of open source might have only recently celebrated its 20th birthday, but it’s long since grown up and moved out of home.
In a future post I will dive into this topic further. And acknowledge the greatest difficulty here is knowing that the need for wider discussion about the future of open source is something that will antagonise some of the open source community itself. Especially now that open source is for… everyone.
Books such as Jim Whitehurst’s wonderful The Open Organisation showcases how open source’s origins as a community model can be strategically applied as an organisational framework. Then there’s CapitalOne’s open source culture as defined by policies and programs. Or Data61, the digital arm of Australia’s CSIRO, releasing guidance on open banking. Still in Australia we see the open government partnership releasing open data, which was valued in 2014 at around AUD $24B.
These may seem like efforts drifting away from the core definition and origin or open source. But all are orbiting the same point of gravity in terms of structuring collaboration and contribution at scale. And all serve as examples of the applicability of the lessons of open source going far beyond the origins of the culture.
That’s an excellent question. What do you think?
I’m currently researching the future of open source and would love to hear your thoughts or comments… 🙌✏️💻🧨