September 2008

On Wednesday kernel developer and Novell fellow Greg KH opened the first annual Linux Plumbers Conference with a keynote aimed squarely at the team behind Ubuntu, Canonical. I think Greg could have used the opportunity to inspire more than attack, but Greg obviously feels strongly about the necessity for upstream development. It’s also Greg being Greg: I believe he carries around a spoon just in case he encounters a hornets’ nest.

Does he have a point?

Greg’s contention is that Canonical is a passive member of the Linux community since they do not contribute code upstream. The operative point, that Greg left out, is they do not contribute much to the projects that Greg defines as “Linux” (namely the kernel, GCC, binutils, etc.) I found his focus solely on this projects, at the exclusion of such key upstream projects as Gnome, mis-leading. The kernel is the core of a Linux distribution, so I understand why Greg feels kernel patches are the most valuable, but you can’t discount Canonical’s support of Gnome and other desktop projects.

Why Pick on Canonical?

Perhaps Greg likes to pick on Canonical because they are so successful. A lot of people use Ubuntu (me included.) This isn’t sour grapes. As Greg pointed out it’s precisely because Ubuntu has so many users that their lack of upstream support is so important. The patches/bugs those users report to Ubuntu could be extremely useful to all Linux users, if they were contributed upstream. Greg’s contention is that because Ubuntu merely consumes the kernel and other elements of Debian, (effectively pushing it even farther removed from upstream development) they are not fully participating in Linux. His point is that if you are a developer you should use your skills with those companies who actively support upstream development. By all reasonable estimates, this is a lost opportunity.

But I also wonder why Greg doesn’t pick on other “free riders?” For instance, could Amazon have built their Kindle business without the $1 billion in free software in the Linux kernel? I don’t see them on the list in Greg’s presentation or the paper we did earlier this year. Why Canonical and Ubuntu? Is it because they much more actively market “Linux” as powering their product? They highlight it much more than Amazon, or Google with Android, or Tivo, or Motorola. Should they be punished for choosing to do something I consider to be a huge service to the Linux community?

So do I think Greg should go after Amazon in a similar fashion? Not at all. Linux is free (as in freedom.) That means you can use it without giving back. That means you are free to become a free rider if you want to. Companies who give a lot back upstream should be rewarded. These are companies like IBM, Intel, Red Hat, Oracle and Novell, among others. But if we get into endless bickering about free riders, this makes the Linux community appear vindictive and petty. Just using Linux makes the ecosystem so much bigger for the rest of  us. This is why it’s so important that anyone can use it as they wish. Vote with your dollars, vote with your technology choices, vote by sending letters to companies. Reward good behavior.

So What Constitutes a Contribution?

In open source communities “code talks” and rightfully so. Canonical saw a need for a software project/company to contribute to Linux in a unique way. They focus on building a usable, more polished, more designed, better branded and better supported Linux distribution for the consumer market. By any one’s measure they have been successful in that endeavor. By my measure that is a very valuable contribution to the greater Linux movement.
Does that mean they shouldn’t contribute upstream? Absolutely not. If they can push more code/bug fixes upstream all users, not just Ubuntu users, will benefit. But Linux is also all about choice, and they have decided to fit in where they saw a need not being addressed. That is their choice. Do I agree with all of their choices? No. For instance I think including closed modules with their distribution is the wrong decision yet understand their reasons for doing so.

Ubuntu founder (and benefactor) Mark Shuttleworth recently addressed Greg’s criticism in his blog:

In Ubuntu we have in general considered upstream to be “our ROCK”, by which we mean that we want upstream to be happy with the way we express their ideas and their work. More than happy - we want upstream to be delighted! We focus most of our effort on integration. Our competitors turn that into “Canonical doesn’t contribute” but it’s more accurate to say we measure our contribution in the effectiveness with which we get the latest stable work of upstream, with security maintenance, to the widest possible audience for testing and love. To my mind, that’s a huge contribution.

I happen to agree. Open source is all about “scratching your itch.” About 200 companies contributed to the last kernel release. Mark saw a need for integration and distribution, which builds on the contributions of those 200 companies (and countless others outside the kernel.) It’s also interesting to note that Ubuntu itself is a “victim” of free riders. Many of the new Mobile Internet Devices gaining in popularity take Ubuntu and customize it to fit those machines. They slap a new brand on it and it’s no longer Ubuntu. This is the way Linux works.

I think it’s unfortunate when we get into such public Linux community “us vs. us” debates. (I feel conflicted about adding my voice to it.) Greg’s points do have merit, but they are too simplified. While I find Greg to be a very effective and strategic communicator about kernel issues, I wish he would use those considerable skills to inspire rather than castigate. Then again, he’s very good at stirring up those hornets’ nests.

Popularity: 46% [?]

Earlier this year, at the urging of the Linux Foundation Technical Advisory Board, I decided to create a new event: the Linux Foundation End User Summit. The intent is to combine a small group of large Linux users (generally on the server side) with core community Linux developers. The result will hopefully be technical innovation and knowledge sharing between those who use the software and those who develop it.

The reason this is important is clear to those familiar with the open source development process. Users need to connect with the community, especially those users who need cutting edge features and improvements in their software. But these users aren’t going to post questions to the Linux kernel mailing list: their companies will not allow them to publish this information so publicly. Unfortunately this means a lot of users aren’t getting the most out of Linux. Maybe even more importantly the Linux community isn’t getting the most out of Linux since they aren’t getting input from users.

I hope this event will change this. We’ve had overwhelming response to this event. We have had to limit this to a small number of people to keep collaboration possible. We’ve also limited it to large and sophisticated users of Linux. Is this just because we enjoy being exclusive? No, it’s because we feel these customers will get the most out of interaction with such developers as Andrew Morton, James Bottomley, Ted Tso, Chris Wright and the many more who are speaking and participating in the Summit. (We’ll have a general purpose LinuxCon for everyone else next year.)

You’ll notice a lot of the participants are financial services companies. Those companies derive competitive advantage from their technology so it’s not surprising they have the most to gain from direct interaction with community developers. We’ve got an impressive list of companies signed up: Credit Suisse, CME, AIG, Merrill Lynch, Dreamworks, NYSE, Fidelity, UBS, NYPD, US NAVY, Metlife, Morgon Stanley, JPMorgan Chase, Aetna, NAVTEQ, Goldman Sachs, Citigroup, Mitsubishi UFJ Financial Group (MUFJ).

Looking at that list you’d think primarily Wall Street firms are attending this event. Actually I was surprised at the incredible distribution of users from countries around the globe. The countries represented by end user participants include: Turkey, Ghana, Brazil, Norway, Israel, Japan, South Africa, Yemen, Germany, Canada, Argentina, Czech Republic, India, Russia, USA, Netherlands. So for a small event we’ll be a mini-UN for Linux users.

You can check out the latest program here.

Popularity: 17% [?]

Mark Shuttleworth, founder of Ubuntu, recently wrote a post detailing Ubuntu and Canonical’s contributions to the upstream projects that make up their distributions. There he mentioned a challenge he recently issued to the Linux and free software community: build a Linux-based UI and computing experience on par with Apple’s within two years.

This is the free software community’s version of JFK’s Space Challenge that resulted in the Apollo program. In 1961, JFK issued a challenge to the United State’s Congress, our space program and our academic and scientific communities: put a man on the moon by the end of the decade. (It’s also a serendipitous comparison given Mark’s penchant for space travel.) Minus the cold war rhetoric, I was struck by similarities to the open source movement when I re-read JKF’s speech:

I believe we possess all the resources and talents necessary. But the facts of the matter are that we have never made the national decisions or marshaled the national resources required for such leadership. We have never specified long-range goals on an urgent time schedule, or managed our resources and our time so as to insure their fulfillment.

Recognizing the head start obtained by the Soviets with their large rocket engines, which gives them many months of lead-time, and recognizing the likelihood that they will exploit this lead for some time to come in still more impressive successes, we nevertheless are required to make new efforts on our own. For while we cannot guarantee that we shall one day be first, we can guarantee that any failure to make this effort will be our last. We take an additional risk by making it in full view of the world, but as shown by the feat of astronaut Shepherd, this very risk enhances our stature when we are successful. But this is not merely a race. Space is open to us now; and our eagerness to share its meaning is not governed by the efforts of others. We go into space because whatever mankind must undertake, free men must fully share.

Every leader knows that the most important step is defining the goal. Good leaders challenge and push us out of our comfort zone. Mark has spoken much about the importance of leadership in open source projects. I believe this challenge is an example of that leadership. Are there issues with a collective group designing effective UI and user experience? Probably. Does that mean it’s impossible? Certainly not. Just look at the amazing technical achievements by free software communities to see what is possible.

Eight years later, on July 20, 1969, JFK’s vision was made a reality when Apollo 11 commander Neil Armstrong took a small step for himself and one giant leap for human-kind.

So can we do it? Mark certainly believes it is possible, as I’m sure JFK did before he made that speech. Do you?

Popularity: 26% [?]

This morning my “ambient awareness” (meaning the passive awareness I have from my news feeds and contacts on Twitter and Facebook) is buzzing with Clive Thompson’s excellent article in yesterday’s New York Times. The article, “I’m so Totally, Digitally Close to You” discusses the state of our relationships in a wired world and raises the question of just how close we are to the circles of acquaintance that live within our online communities.

Thompson and others in his articles make the point that by scanning these feeds we can cultivate a type of ESP about those we follow: what they’re interested in, how they are feeling that day, what the weather is like where they live, their travel patterns. I have found Twitter to be effective in those areas, but even more so in tracking “community” reaction to news in technology. It’s an effective and more personal Zeitgeist than the Google Zeitgeist since the people I follow and “know” are generally people whose opinions I value.  I am applying my own algorithm much like Google does. If I follow them, that means their opinions should matter, so when I see their reaction to the Vista Seinfeld ads (not good) or to Chrome (good), I pay attention more than if I just read one blog post or have one or two conversations about those topics. The wisdom of crowds in real time.

But how well do I know the people I “follow” online and should I be trusting these relationships? Enter the Dunbar number:

In 1998, the anthropologist Robin Dunbar argued that each human has a hard-wired upper limit on the number of people he or she can personally know at one time. Dunbar noticed that humans and apes both develop social bonds by engaging in some sort of grooming; apes do it by picking at and smoothing one another’s fur, and humans do it with conversation. He theorized that ape and human brains could manage only a finite number of grooming relationships: unless we spend enough time doing social grooming — chitchatting, trading gossip or, for apes, picking lice — we won’t really feel that we “know” someone well enough to call him a friend. Dunbar noticed that ape groups tended to top out at 55 members. Since human brains were proportionally bigger, Dunbar figured that our maximum number of social connections would be similarly larger: about 150 on average. Sure enough, psychological studies have confirmed that human groupings naturally tail off at around 150 people: the “Dunbar number,” as it is known. Are people who use Facebook and Twitter increasing their Dunbar number, because they can so easily keep track of so many more people?

So what’s my Dunbar number? And are my online friends displacing my “real” relationships? I follow 86 people on Twitter. (127 follow me.) On Facebook I have 90 friends (with some overlap with Twitter followings.) That’s already over the 150. See you later family. Obviously it doesn’t take much to follow someone on these online feeds. That’s what’s so great about it: it’s ambient and voluntary. I didn’t log into Twitter at all from a conference I attended last week and it was fine. Had I not called or texted to my significant other for a week, there would be a problem.

But I wonder how this affects those of us in the middle of open source communities? Because of the collaborative development model, open source people should know far more people than typical participants in non-collaborative industries. For instance, I know quite well all of our board members (20), all of our vendor members (a lot), all of our technical advisory board members plus others in the kernel and other open communities (the most). These contacts take more interaction than my online ones. Perhaps there can be a formula where you can take all your online contacts (.10), your work contacts (.25) and your friends and family (100%) to come up with your own Dunbar number. You can then weight the time and attention to give to these communities accordingly.

Regardless, I have these new tools extremely powerful. I can throw a request out on Twitter and get multiple calls to help. (Keep an eye out for an upcoming paper that proves this.) When I’m in a new town, I can find friends I hadn’t met before or restaurants to try. Everyone is really so helpful in my online circles. It’s the power of acquaintance and open communities.

Popularity: 18% [?]