Opinion: On Canonical, Red Hat, and their communities
When I can, I try to participate in The Linux Link Tech Show when it is streaming LIVE... but even when I can't I often listen to the archived recordings. When I find something interesting I'll sometimes shoot Dann Washko an email with my thoughts. This morning I found myself writing a long email to him on a subject they covered on their June 15 episode (#407). I thought I'd post it here too.
It just so happens that several of TLLTS regulars had attended the Southeast Linuxfest the weekend prior and one of the conversations that Dann encountered there was about Canonical and Ubuntu. Dann spoke about the questions and opinions he heard raised and asked for everyone else's opinions but he didn't get a whole lot of feedback so I thought I'd provide him with some.
I'll admit yet again... I'm a big Red Hat and Fedora fan and I am biased... and I sometimes even serve as an apologist for them. While I think everything I say below is "fair and balanced"... I'm sure there are plenty of folks who disagree with me... and maybe one or two who agree... I do encourage feedback and comments from all sides. Read on at your own peril. :)
I missed the live show because I had a father/son(s) event to attend where we built a planter out of wood pieces and ate pizza. Anyway... some comments about the show... particularly about your initiated discussion about Canonical and Ubuntu.
I think Ubuntu is about in the same situation as Red Hat was right before dropping Red Hat Linux and focusing on Red Hat Enterprise Linux... they were the most popular distro at the time... with a significant chunk of the desktop userbase (such that it is)... with the vast majority of users not really contributing back in any helpful way. To me if Ubuntu has a ton of users who aren't giving Canonical any money they are just an albatross around their neck... dead weight that is just a reminder that LOTS of deadbeats are still deadbeats.
When I say contributing, I don't necessarily mean financially but that would be what Canonical would want most. At least at the time Red Hat had Red Hat Network going and your average user could pay $60 / year for priority access to updates and Red Hat actually had some users buying that. I was one of them.
Canonical also offers paid support for desktop users at $105 or $165 per year. I wonder how many subscribers they have. Their server support costs seem to be very similar to Red Hat's although I haven't compared them recently. Since Canonical isn't a public company, they don't have to report any of their financials... but the thought is that if they were doing well, they'd be talking about it in public.
While I like the single CD install concept (which Canonical popularized) I'm not fond of the different spins that everyone is producing now (including Fedora)... where it gives new users the idea that they have to pick a single desktop environment and if they want something different, they have to download another install media and start from scratch again. I don't like how it has lead distros to focus on one desktop environment as their "default" becoming the one they give most attention to... and their developers are now split up into camps focusing on different spins. Everything comes from the same package repos and their isn't any reason to keep them separate. Who doesn't have a DVD reader these days? Put more on there. While many distros do still produce "install only" media where you can actually pick what packages you want and don't want... the vast majority of users get a small set of specific packages that are supposedly "best in category" when it has a lot to do with filespace availability on a single CD. I like to choose for myself and often that means installing multiple applications for any given application category because all apps have their strengths and weaknesses so why be stuck with just one? Of course I do realize that I can install any and everything after the fact but so many users seem to think, for whatever reason, that that isn't the case.
One idea is that Canonical and Ubuntu have increased the visibility of Linux among non-traditional users and the general public... and as a result they have increased Linux's desktop market-share. To the best of my knowledge there is no evidence of that. Linux still has dismal desktop market-share numbers (which I don't really have a problem with) and all Ubuntu has done is poach off users from other distros. That isn't to say that there aren't any new users because there are but they are only at the expected rate of growth. Some believe that Canonical has taken away what could have become over time users what would have contributed more in different ways... only to coddle them into staying lazy, non-contributing users who are only visible when they pop their heads up to complain about changes they don't like.
I will spend some time comparing Red Had and Canonical.
Red Hat is a significant code contributor in a number of large projects and every survey that is done, and there have been many of them, shows that Red Hat is on top in every measurable way. They are the biggest code contributors to the Linux kernel, gcc, gnome, and a number of other core components... and they financially sponsor a large number of open source projects.
Red Hat tries to release everything they do under an open source license because they seem to genuinely believe that it is the superior development model. They aren't always able to immediately do that and sometimes it takes years. Red Hat has actually purchased a number of commercial software products and worked hard to make them open source. Take Netscape's Directory Server for example. Now they are working on what was Qumranet's SolidICE product (now Red Hat Enterprise Virtualization). SolidICE was a Microsoft .Net based product requiring Microsoft Windows 2003/2008 Server with IIS as the management platform... also requiring Internet Explorer / ActiveX on the client side for management. Red Hat is in the process of reimplementing everything in Java and the fruits of that will eventually be open sourced in some fashion. Red Hat was criticized for a long time because the server side of Red Hat Network was closed. It took a few years but they finally open sourced it as Spacewalk. They have page after page on their website explaining what Open Source is and why the believe in it.
Canonical seems to not believe in the open source development model as strongly. A number of their newer efforts contain closed software components. Do I have to mention them? I think they are pretty well known. One example would be the server side of Ubuntu One (should have been named Canonical One but I digress).
Mr. Shuttleworth gave a speech at the most recent Ubuntu Dev conference that I'm not sure what to think about. He seemed to be making the case that the open source development model is broken because it can only get 90% of the way there and that we are left with some really bad desktop applications... and as a result we should somehow cater to the commercial software companies who will not open up their source... because we need the high quality that only the proprietary development model can make. If you don't know what I'm talking about, watch the video yourself and try to make sense of it.
LWN did a feature article on it but it seems to have been largely ignored by everyone else and I'm not sure why.
Mr. Shuttleworth even mentions selling the house so one has to wonder if his goal is to do with Canonical what he did with his previous company (Thawte)... which was get it into a favorable financial position so he can cash out.
Canonical seems to be trying to do a lot of things all at the same time (See: Dear Ubuntu: The netbook is toast or Is Canonical’s Ubuntu Focus Too Darn Diverse?). The thought is by throwing a lot of stuff at the wall hopefully something will stick. While it is assumed they have been and will always be focusing on the desktop, if some other area of the business started showing the potential to be a profit center, is there any question that their focus as a company would change to suit?
The very popularity of Ubuntu seems to stem from three key factors:
- A lot of users were mad at Red Hat with the RHL to RHEL switch and the cost of Red Hat Network and Canonical made it clear that "Ubuntu would always be freely available". The message was that Canonical would never do to users what Red Hat did to theirs.
- Ubuntu is based on Debian and some Debian supporters thought that Debian was just too big and bureaucratic of a project supporting too many architectures with too long of a development cycle... and without enough focus on desktop end-user friendliness for the Debian project to ever capitalize on its enormous strengths. Ubuntu was a way to take the strengths of Debian... speed up the development cycle... put focus on improving the ease of use and the desktop experience. Contrast that to Red Hat's very public statements that they had no desire to work on a commercial Linux desktop because there simply was no market for that. It has lead to a bit of bad blood between Debian and Ubuntu... and as a result Debian has sped up their development cycle some... which has lead to a shorter support cycle for Debian.
- Canonical started by producing fancy CD media in a fancy slip cover and shipping that out in bulk for free to anyone who wanted it. You didn't even have to download it and burn it yourself... you could just fill out a web-form, wait a week or two... and you had your own media copy and lots of additional copies to give out at installfests, LUG meetings, whatever. Who could ask for anything more? They don't seem to be doing that anymore though.
Red Hat seems to focus on making Linux better for everyone by working on things everyone can use whereas Canonical wants to make their own stuff that makes them stand out as different than everyone else. As one of many examples... Fedora produced PackageKit which is a somewhat generic GUI software/package manager that is backend agnostic. It can be used by both .deb and .rpm based distros as well as a few others. Canonical has focused on their Software Center with app store type extensions trying to make it attractive to commercial software vendors so Canonical can get a cut of the sales like Apple does on their app stores.
Red Hat through the Fedora Project... is the major advancer in new technologies and doesn't seem to mind getting a little egg on their faces from time to time. In fact, I'll be crazy for a second here and state that if a distro isn't occasionally breaking a few things, they probably aren't trying hard enough. What does that mean exactly? Well as an exercise... look up a few of the following dates. When did Fedora start using SELinux? When did any of the other major distros? How about Xen? How about KVM? How about Pulse Audio? How about KDE4? How about systemd? Sometimes there is as much as two years between the adoption by Fedora and by others. Canonical generally plays it safe... although I'll grant that Unity was quite a risk for them to take.
Wow... that was a lot of text from me. Sorry for the length.
Now to address your question about is Ubuntu harmful to the Linux community as a whole? I don't really think it is. It might have lead to or be leading to... dumbing down the average Linux user. Personally I'd prefer to have less Linux users if they are more competent technical users (being more likely to contribute back in some way) than having a lot of Windows and Mac refugees who are "twelve o'clock flashers". Perhaps that's just me though?!?