Unix was ostensibly forged on the philosophy that every entity be designed to do only one thing, and do it well. Has Canonical, which develops one of the most popular Unix-like OS’s around today, thrown this philosophy out the window when it comes to business strategy? Former Canonical COO Matt Asay thinks so. Here’s why he may be wrong about Canonical and Ubuntu Linux.
Arguing recently that Canonical missed the bus long ago on the netbook market, Asay, who served as chief operating officer of the company for ten months before resigning last December, also made the case that Canonical’s endeavors are currently too diverse:
Canonical needs more focus. The company does well whatever it does, but it’s organizationally impossible to be exceptional in so many different areas.
The key to achieving major success, Asay suggested, is to focus on a single market. Canonical’s investment in disparate areas, from the desktop to the server to mobile computing to the cloud, leaves it unable to emerge as the very best contender on any of these fronts.
Diversity and Canonical
As former COO, Asay certainly knows a thing or two about Canonical–including, possibly, information that the company, which remains private, has yet to divulge publicly. No one on the outside can be sure how profitable the organization currently is in any area, or how much it still relies on the charity of Ubuntu founder Mark Shuttleworth (it is worth noting, however, that the company continues to expand its workforce, a healthy sign).
From my perspective as an outsider to Canonical, though, it seems like the organization’s diversity is key to its success. Ubuntu has gained huge prominence and influence within the open source ecosystem thanks to its focus not just on either the desktop or the server, as other distributions have tended to do, but on different and diverse segments of the channel. Ubuntu’s momentum stems from the fusion of passionate desktop users with server administrators willing to pay for Ubuntu based solutions, many of them involving not only traditional infrastructure but also the cloud.
It’s hard to see how focusing on only one part of the market could even be feasible for Canonical. If it dumps all of its resources into advancing Unity and the Linux desktop, it will lose important revenue streams in the server and cloud markets and probably go the way of predecessors like Mandriva, Inc. If it shifts exclusively to servers, it will deny itself the enthusiasm of the Linux community which has been key to the success enjoyed by Ubuntu thus far. And if it abandons the cloud and mobile computing worlds, it risks being shut out of emerging opportunities.
Emulating Apple?
Although he didn’t say so explicitly, Asay seemed to want to suggest that Canonical needs to become more like Apple, which he called a “premium innovator,” and less like Microsoft, a company with a broad focus that, in Asay’s view, is now “struggling to compete” with Apple. I’m not sure this makes much sense, and not only because Microsoft’s performance over the long term deserves more credit than Asay gave it.
Apple didn’t achieve what it has by putting all its eggs in one basket and getting lucky. The company’s most popular products–iPad, iPod, iPhone and PCs running OS X–constitute a range of different platforms and devices. And even though most of its products are directed at individual consumers, OS X server has yet to go away.
If Canonical is able to maintain a broad development front for long enough, its break may well come, just as Apple’s came when the company was on the verge of ceasing to exist. Scaling back the scope of topical investment is not the right move for Canonical.
Besides, Apple’s success is founded on jacking an operating system from computer scientists at Berkeley and building a culture of commodity-hungry consumers trained to associate overpriced electronics with personal sophistication and prestige. The last thing the world needs is another Apple. But we’d best not go there.
Sign up for The VAR Guy’s Weekly Newsletter, Webcasts and Resource Center. Follow The VAR Guy via RSS, Facebook and Twitter. Follow experts at VARtweet. Read The VAR Guy’s editorial disclosures here.
Read More About This Topic
Share This Post
Tags: Apple | Matt Asay | Microsoft
Interact: Add a Comment | Trackback Link | Permalink
Subscribe: RSS Feed
Don't miss Charlene O'Hanlon's weekly columns...
For the last year and a half, I’ve been harping that Canonical needs to make development for Ubuntu much easier by having an officially “knighted” set of tools (IDE, toolkit, language, base libraries) and copious tutorials for it. I think they are finally moving in that direction now that they’ve got the Ubuntu Software Center in decent shape.
Where Canonical needs to go next is to work harder to hook Ubuntu One SSO into the desktop. They could use the U1 SSO to host services like XMPP chat, StatusNet microblogging, and other communication technologies which are set up and ready right after installation. Pushing PSAs and offering support would be a simple matter, then. Right now, new users need to find Ubuntu Forums or AskUbuntu, sign up (these should be U1 SSO), ask a question, and go back to check for answers.
Daeng Bo: true. And I don’t think having an official set of tools is at odds with retaining influence across disparate channels.
There is a reason why OS companys (OSS or not) tend to focus on one space, the requirements of each ‘flavor’ of computing are completely different. Focusing on enterprise server functionality hampers desktop innovation (Red Hat, Solaris, *NIX). Focusing on desktop requirements hampers enterprise server stability and reliability (Ubuntu, Windows, Mac OS).
In order to serve several masters, there would need to be a different team building each ‘flavor’ with some crossover. M$ showed this was possible, but cross-pollination has caused desktop instability to creep into the server lines which they are just now recovering from.
Server customers do not want an app store and nifty XMPP chat support channels. They want minimal footprint, high performance, stable, reliable workhorse platforms. Desktop users want visual splashiness, their favorite websites to work, intuitive GUIs.
Asay is right, focus or fail.
I think Ubuntu’s diversity can definitely be a strength. *Because* it’s Linux, you would think that keeping the server side solid would be really doable. I have high hopes for Unity, but I’ll probably wait until 12.04. Still, I think it’s the right way to go. Ubuntu One makes me curious, but I’m syncing everything through Google for now.
What Ubuntu and all Linux distros need are apps. Yes, there are some good programs, but it’s really not enough.
Cultivate a developer community that allows people to make a buck or two from writing their code, and I think you’ll have a winner. With my Android phone, I have no problem spending $1 to $2 for something cool or even $10 to $20 on something that gives me a solid productivity tool.
I love FOSS, but let’s not pretend that people don’t need money. Almost anything Canonical can do to expand their potential user base is a positive and provides more draw for developers.
diversification = more developers = more apps = more wins
@Jay – Don’t forget that a good part of the Mac OS X, and NeXT’s before it, kernel was derived from Mach at Carnegie-Mellon. Plus, the BSD license is very permissive. As another example, Netezza, the data warehouse appliance company recently acquired by IBM, uses a database management system based on PostgreSQL which also has a permissive license. There’s nothing unusual or wrong with this practice involving permissive open-source licenses.
Also, I can’t see Canonical turning into another Apple as it is derived from Debian which is squarely in the GPL camp (minus the Debian k-freebsd experiment currently underway which btw is very cool).
@Christopher Tozzi – Like you, without some insight into Canonical’s finances, it is difficult to speak with any confidence on whether or not they are spread too thin. But, I do like the direction they are going with regarding the Unity desktop and Wayland replacement of X.Org. Imagine, desktop Linux supporting multi-touch with a consistent user interface on tablets, netbooks, laptops and desktops. I sincerely hope they make it to the finish line as I am not a big fan of either Apple or Google.
I don’t think Ubuntu has lost any trains or busses on any markets. Tablets are not replacing netbooks, HTML5 is not replacing desktop apps, ARM is not replacing x86… I could go on. It is very easy to forget that we’re just starting. And I’m talking about the entire computing industry, which hasn’t even grabbed 15% of the global market yet. We’re in the very early stages of something new. We’ll see a lot more diversity than we have so far and many of the form factors in use today are going to stay. People are not going to stop working at their desks, so we are going to see desktops. Mechanical keyboards are not going to be replaced by on-screen keyboards any time soon. A hammer does not replace a saw.
Ubuntu is available on several different kinds of computing devices, and with Kubuntu Mobile, it’ll reach yet another market. It may seem like going into all these markets at the same time dilutes the effort, but in reality, it concentrates it. It brings a common platform to ARM, Intel, servers, mobile, tablets, desktops and the cloud, which in turn makes knowledge more reusable. This is key to making Ubuntu an attracting platform for new developers, which in turn makes Ubuntu more interesting to end users.
It also makes it easier to narrow the gap between different technologies. As an example, have a look at the new indicators in Ubuntu. Because Ubuntu comes in different flavours, like Ubuntu, Kubuntu, Xubuntu and Lubuntu, it makes sense to make these things work everywhere. I was very surprised at how easy it was to write an appindicator, and knowing that it would automatically work and fit perfectly into the other environments as well, makes it very gratifying to write one. Is it likely that this seamless interoperability would work as well if Ubuntu was only focused on Gnome and Unity? I don’t think so.
I wrote a small database application a while ago, based on CouchDB and DesktopCouch. It ran on my desktop, synced with my Ubuntu One account and then to my laptops and my mobile phone. It was an eye-opening experience.
I believe that a broader perspective improves focus, because the big picture defines the small tasks. If Canonical and Ubuntu can keep narrowing the gaps, then it will soon result in a developers heaven.