Let me start by saying how much I love Open Source software, peer production, the tide that raises all ships, Wikipedia, all things “open”. It is part of how I define myself. I love what happens when people share expertise, resources, their spare time. It makes me feel like I am part of something larger. It makes me feel powerful and creative, only my effort and imagination can hold me back. Yet, for some time, I have felt a growing unease with the “open” movement.

I think it started back in 2006 when the South African government established a policy directing the use of Open Source software within government departments “unless proprietary software is demonstrated to be significantly superior”.

This policy did not achieve its aim of converting government departments to the use of Open Source. If anything it probably alienated civil servants more than it made them converts to Open Source. It made them feel like FOSS was some kind of second class solution they were obliged to use because they couldn’t afford the best. I knew I didn’t like this policy at the time but I couldn’t really put my finger on why except for the basic awareness that nobody likes to be forced to do something, even if its good for them.

Fast forward a year to 2007 and we have Tony Blair saying:

“Open v closed” is as important today in politics as “left v right”. Nations do best when they are prepared to be open to the world. This means open in their economies, eschewing protectionism, welcoming foreign investment, running flexible labour markets. It means also open to the benefit of controlled immigration.

Once again, I respond positively to openness and indeed open vs closed does seem a more practical unit of political analysis than left vs right these days, at least on some levels. However, again I experience a trace of unease and this time I can’t put my finger on it. And in the ensuing 5-6 years we see openness emerging as a western agenda with the implicit assumption that open = good and by extension more open = more good.

I expressed some of my concerns when I wrote about this in 2009. I likened open vs. closed as a kind of yin and yang, two parts that can’t live without each other.  That was close to what I want to say here but I think the problem is bigger now and that analogy misses some critical nuance

More recently I’ve been thinking about “open” in the context of Open Data and how that relates to personal privacy. Clearly more open cannot always equate to more good in this context. If we acknowledge that the need for privacy is contextual, then it is axiomatic that the need for openness is also contextual. The problem with making a virtue of “open” is that it tends to steamroller nuance and context.

I am reminded, as I often am, of the Taoist parable of a farmer and his “good” fortune.  Nothing is inherently good or bad but is defined by the context in which we understand.  The more I think about it, the more I think the crux of the problem lies in an essentially manichean worldview where open is now equated with virtue, where we must fight the forces of closed-ness wherever we find it.

This is wrong in the same way that the Golden Rule is wrong. Do unto others as you would have them do unto you. What could be simpler? Yet, the angry drunk who enjoys getting into fights in bars has no problem with this rule. It helps him get into more fights. Absolute rules tend not to fair very well in complex environments.

So, let’s consider some other “good” words.  How about “kindness” and “cleanliness”? Any parent or anyone who has been in the wrong relationship knows that you have to be “cruel to be kind” sometimes.  Doing the right thing might actually involve not being “kind”.  So perhaps kindness is not a very good goal.  And cleanliness.  We know it is next to godliness and on the surface of things how could one argue with it.  But a quest for cleanliness actually has led to some surprisingly negative outcomes such as the growth of allergies.  The more we understand about our bodies, the more we realise that what we previously thought of as “unclean” is actually a part of what makes us human.

So, would a “kindness” movement serve us well?  Or a “cleanliness” movement? Well, the answer is not yes or no, it is “mostly”.  Mostly being kind is a good idea and mostly being clean is a good idea but they are bad when turned into doctrine and orthodoxy. The rationale for orthodoxy is that if you don’t keep things pure enough, then it is a slippery slope to the increasing adulteration of all you hold dear.

The problem with purity is that is that it leads to fragility.  In Anti-Fragility, Nassim Taleb argues that all complex systems need to be stressed in order to grow stronger, to reduce fragility. Perhaps open works need proprietary works to stress them into improving and evolving.  As I wrote previously, the evidence from multi-party prisoner’s dilemma simulations would seem to support this, namely that “open” strategies succeed very well in a very closed ecologies and “closed” strategies succeed very well in a very open environments.

So what’s an Open Source advocate to do?  Well, if you were Evgeny Morozov you could rubbish the entire open movement but that doesn’t work for me because I really do see and live the benefits of open all around me.   I think what is needed is a new concept, that of “Right Openness”.  In Buddhist philosophy, one of the principal teachings is the Noble Eightfold Path, which describes the “path” to enlightenment.  Each path begins with the word “right”, Right Intention, Right Speech, Right Action, Right Livelihood, etc.  What is notable about this is that there is no prescribed behaviour.  There is the overall goal of reducing suffering and being compassionate but the way you achieve that is not specified.  Mostly kindness is a great way of being compassionate but not always.  Mostly openness is a great way of achieving good outcomes in the growth of knowledge  in good governance, etc but not always. One need only look at a for-profit 3-D gun printing initiative to see how “open” as orthodoxy can lead to the wrong sort of outcome.

Let’s take the example of Right Speech. In the west, we value truth and freedom when it comes to speech. Yet, anyone who has uttered the words “Let’s be honest” in a relationship, knows that there are truths and there are truths. Time and geography matter when it comes to truth. Learning the truth about a murder that happened hundreds of years ago, half-way around the world is not the same as learning the truth about a murder that happened an hour ago, next door. The same with freedom of speech. Encouraging someone to help their neighbour is not the same as encouraging someone to kill their neighbour. We know this, yet we still defend freedom of speech when I think what we really mean is Right Speech, speech that does not harm others, that is timely, etc.

What then would Right Openness look like? It would recognise that “openness” is not an inherent virtue but rather a contextual good. What would that look like in practice? Well it would always hark back to the question of the larger goal, whether more equitable sharing of knowledge, good governance, etc.  It would then ask what right openness looks like in that context. It would lead by example, not by doctrine. In the Open Data world, it would embrace privacy issues as being fundamental to effectiveness openness without getting hung up on privacy as a violation of openness.

As we struggle to understand the complexity of the world we live, we look for simple rules to help guide us through the storm.  That’s great as look as we treat them as rules of thumb.  To paraphrase George Box, all rules are flawed, but some are useful.

So let’s hear it for Right Openness and remember kids, “Only a Sith deals in absolutes

Photo courtesy Stephanie Davidson 2008 CC BY SA



Posted by Steve Song

@stevesong local telco policy activist. social entrepreneur. founder of @villagetelco #africa #telecoms #opensource #privacy #wireless #spectrum #data

  • Mikel

    Well put. It is great to have reminders like this pop up like this.

  • GastArbeiter

    Well with software there’s relaxingly open (X/MIT, BSD, perl, 3 clause licenses meant for the world to use and build upon) and then there’s rantingly open (FSF, GPL v.3, and the long fight to free the world from domination). One seems real the other more absolute.

  • stevesong

    I think neither is right or wrong per se. Depends on what you’re building. For some things, like say a compiler, I can see the desire to entrench openness FSF-style because it is such an important tool for building other things. But for other projects, a more relaxed approach offers more possibilities.

  • thesunnyk

    You’re potentially going a bit too far with this zen thing. Unfortunately, “open” as a word is often used as advertising, so the definition gets twisted and turned until everyone is left confused. People use closed solutions because they have no option. Put simply, they put a price on openness and sometimes the “premium” you pay for an open something is too great.

    The problem is, people are terrible at gauging future risk. Today, you might be well happy with your proprietary stack. Tomorrow, you’ll cry bloody murder when the owners take advantage of the rights you gave to them. What I’m saying here is, people price openness far too low.

    GastArbeiter talks about “relaxed” open vs “strict” open, but fails to realise that the “relaxed” open simply guarantees the rights of the programmers, but doesn’t guarantee the rights of the users. BSD style licenses are actually “strict” in the sense that the programmers are well protected. Even more so because they grant the programmers the rights to bind their users. BSD style licenses might be more popular, but that does not equate to better.

    In conclusion, I think you do want to maximise openness (which you’re calling “right openness” but that would be like calling a table a “right table” just because some post modern artist look a urinal and labelled it a table, so let’s just call it “open”). But you have to look carefully because simply looking at a spec sheet and seeing “OPEN” with a tick box next to it is insufficient. Open means something — it has a principle and ethos attached to it. We need to evaluate that ethos before we even call it “open”.

  • Pingback: This week in Transparency/Accountability | open up()

  • stevesong

    You may be right about the Buddhist metaphor. At the moment, it is more of an exploration than a manifesto. I am still searching for the right way to express myself about this.

    In my experience, people are quite happy to use closed systems as long as they are affordable and they work. Google is a great example of this. I think you are right about gauging risk not just because we don’t weight the longer outcome heavily enough but also because we simply don’t understand the long term implications of what we’re doing e.g. giving up privacy in exchange for free services.

    Taleb says something insightful in Antifragile. That risk is basically unquantifiable but that fragility is quantifiable. In other words, we can’t predict the future but we can look at what we have and determine how vulnerable it is to systemic shock. Monopoly is more of an enemy than closed is. It is just that closed is a great way to get to monopoly.

    What I really mean to say is that there is a purpose behind all openness which is not openness. We don’t open for open’s sake. We open for more equity, more choice, more accountability, more participation, more collaboration, more fun, but not for open itself.

  • thesunnyk

    I agree with you until the final paragraph. On the one hand, when you say: Openness does not a-priori mean good, I agree. However, you’re putting the dichotomy in an imbalanced way, like saying “by default, everything is closed and that’s fine, but sometimes openness yields good results and then openness is worth it. But why? Why not just say “by default everything should be open, but sometimes… closed is better?”

    The problem is I can’t think of an example where you would take something open, then close it, and that would result in something better. You can only really say “opening something that’s closed might not offer any benefits”.

  • stevesong

    I miscommunicated. I DO mean to say that open should be the default, however, the default strategy to achieve whatever end goal you seek. I DO think that in terms of the current economic/social ecosystem the balance is still too far in favour of closed.

    That said, I think there a plenty of examples where closed has succeeded in ways that open hasn’t. Steve Jobs took BSD and built OSX on top of it. The FOSS community struggled for years to build a decent desktop. He did and Ubuntu has largely benefited from that closed decision by emulating a lot of their success.

    Interesting too that the most successful FOSS projects are a little closed in that they often have benign dictator leading them. Mark Shuttleworth’s self-designation as SABDFL is only partially tongue-in-cheek.

  • thesunnyk

    I don’t want to sound overly critical — I mean this respectfully, but you’re using very touchy feely word association here, and I’m trying to bash it into my left-brained head, which is not working at all. Let’s do this in reverse:

    I don’t understand how a SABDFL (even if taken seriously) counts as “closed”. Torvalds or Shuttleworth still necessarily make all their decisions out in the open, and their decisions are always accompanied by a rationale. They hold power, but it’s entirely soft power. Anyone is free to fork their stuff. In Shuttleworth’s case, you can see where this has happened, with Linux Mint, for example. In any case, none of this is “closed” in any sense — there’s no DRM, no licensing agreements, no proprietary software, no hidden decisions. If anything, it’s just a community’s consensus to defer to someone’s opinion as a matter of course.

    You seem to be equating “success” with “popularity” again. OSX isn’t “better” than BSD inherently. Maybe it wins the survival game (except technically BSD survives in OSX so that’s still a BSD win) or maybe it wins a popularity contest or maybe it has better brand recognition, but OSX isn’t “better” in any comparable way to BSD. People write apps for OSX, is that “better”? People pay money for OSX, but they don’t / can’t for BSD. Is that “better”? Maybe you find more Max zealots than BSD zealots. Would that make OSX “better”? Or even “right”?

    Your first paragraph has thrown me for a loop. I really don’t understand what you’re trying to say. Are you saying that, for problems you want to solve, you sometimes need closed source software? I suppose that’s correct. However, in today’s environment, you’d often be placed in a position where you pretty much have no option but use open source software in certain situations. In fact, those circumstances are becoming more common, not less. How many things do you own that run Linux? Over time, do you expect that number to increase or decrease? Even if it does (or doesn’t), how does that make one solution “better” or more “right” than the other? Isn’t this what I was saying earlier: most people just try and get whatever they want done right now, and worry about the consequences later?

    I still don’t understand the “yin/yang” thing where closed somehow ends up being a good thing, ever.

  • stevesong

    No worries. I’ll try to be a little more left-brain in my response. So, first of all, mea culpa, for playing fast and loose with the word “open”, the very thing I critique in my post. The only point I meant in the SABDFL comment is that the open source world is not entirely as egalitarian as it appears. You are right that it is still not opaque.

    In my comment about OSX, I wasn’t thinking about popularity. I didn’t mean to compare OSX with BSD. I was referring specifically to the desktop UI that Apple built on top of BSD. A desktop that set new standards for UI design, obviously my subjective assessment. And again subjectively, I think that part of what enabled them to achieve that was having absolute control over the UI. The desktop in the FOSS world has been bedevilled by factionalism and legacy which has led to a pretty awful user experience of the Linux desktop. Canonical were able to learn from Apple’s design to bring something much better to the Linux desktop. Again all my highly biased opinion. We see this all over the FOSS world where proprietary initiatives are emulated by FOSS initiatives. It happens in reverse too of course.

    Which brings me to economics and business models. You say that you can’t see how closed ends up being a good thing ever. There is only one real motivation for proprietary software. Money. A hard reality that everyone has to face. For myself, as the founder of an Open Source, Open Hardware social enterprise (http://villagetelco.org), I struggle with how to make a successful business out of a purely open model. I know most of my colleagues/peers struggling to developing open businesses face the same challenge. That’s why we see the evolution of mixed closed/open models like ‘freemium’ software. Many FOSS initiatives survive because proprietary software companies donate developer time to them. Maybe a more left-brain expression of yin and yang would be to simply say that open and closed models exist in symbiosis with each other.

    I think there are some exciting new ideas out there like crowd-funding that is helping to improve the odds for FOSS projects but the challenge of the economics of open is never too far away.

  • thesunnyk

    Thanks for the responses. This conversation really got me thinking. You’re right, of course, in that money decides what gets done. The classic case is the GNU FSF which, I guess, is so starved of cash that even its high priority projects are languishing, and if doesn’t get done, it isn’t doing any good.

    I wish you best of luck with villagetelco. One thing I read about RedHat is why they weren’t becoming a massive company — because they were making things orders of magnitudes more efficient. Where a company might pay $100,000 to Microsoft, they’d only pay $10,000 to RedHat, so really if their share price goes up 10% they’re actually displacing a huge amount out of Microsoft’s hand. It’s a net benefit to society, but it doesn’t mean massive profits to RedHat. In a way, this is what OSS brings. Again, I hope you find your way through.

  • stevesong

    Thanks on my side too. I leave you with a beautiful comment that I heard from JP Rangaswami (https://twitter.com/jobsworth) at a LIFT conference. He said “In Open Source work, losses are privatised and wins are socialised which is the opposite of the financial world.” Struck me as pretty profound.

  • Pingback: Designing for trust | WorkOpen.org()

  • Pingback: Going Open – Resources For All | Open Coalition()