Has Ubuntu, in fact, lost it?

Since 2011, Canonical has been pissing a lot of people off with its approach to Ubuntu development. There has been a lot of discussion on the Internet as to whether Ubuntu has completely lost its way, and this very topic was the subject of a major feature in this month’s Linux Format.

I use Ubuntu myself, so in case you care, here are my own thoughts.

Button position

In Ubuntu 11.04, the minimise, maximise, and close buttons suddenly migrated from the right to the left side of the window. A lot of people complained about this, citing it as a meaningless and arbitrary change that damaged their workflow by messing with muscle memory. Canonical made a weak defence, claiming something to do with improving creative and innovative activity in users, apparently on the basis Apple also puts its buttons on the left and artists swear by their Macs.

In the end, people quickly figured out how to get the buttons back to where they were supposed to be, and everyone did so. At the time, we all thought it was a weird blip – Canonical makes a really strange change, but it’s easy to undo, and hey, this actually expanded on the choice that Linux users rant about, as we can now choose where the buttons go. Now, if someone actually likes the buttons on the left side, they can have them there.

Little did we know that this was a prelude to the shitstorm that was…

Unity

Hoo boy. This was a big thing. Unity exploded onto Ubuntu users’ desktops in April 2011, and they’re still complaining about it.

Unity was a massive, unexpected change to the way we were used to using desktops. Gone were the sensible panels and menus; in their place was a dash menu that showed a grid of icons, including applications, folders, and files all mixed together. Applications were launched through the dash menu or through quicklaunch icons on the side menu; switching between two instances of a single application was annoyingly tricky. Furthermore, there was a single, persistent panel at the very top of the screen which contained the normal icons one might expect (clock, logout, battery, volume, and so on). In addition, when a window is maximised, its menu bar goes into the left side of the top panel rather than remaining attached to the window itself.

This was the true reason for moving the buttons to the left – when a window is maximised, its buttons go to the leftmost edge of the panel, regardless of whether the user has set them to be on the right or not. This takes some getting used to.

Additionally, since all application menus share a single panel, the menu items one gets are always those of the currently active window. This means there is considerably more clicking around involved when working with multiple windows, and the cursor has to travel a much greater distance; working on GIMP under Unity is an absolute chore. In fairness, Unity does bring the option to do a text search for a feature by using the Alt key, which would be very useful if you can get used to it, and I suppose could make a programme like GIMP or Blender, with their hundreds of options, easier to use. Most people, however, were used to clicking through menus, and so this presents a serious interference to people’s workflows.

In fact, people even left Ubuntu for other distros over this. I find this baffling – sure, I didn’t like Unity, but KDE, Xfce, and LXDE are only an apt-get away, and there are other desktops out there that can be acquired with only the additonal step of adding a repository.

One might be tempted to call these teething troubles. Indeed, Unity has come on in leaps and bounds over the last two years; I’m actually using it right now, and it’s a rather slick piece of kit, albeit one that requires a pretty decent PC to run. However, Unity is quite clearly an attempt to create an interface that can be used on both keyboard-and-mouse and touch-based computers. How does it work?

As I said, it works pretty well on a desktop. Seeing it on a tablet or smartphone is still a few months off, but from what I’ve seen, it works very well indeed in that environment.

However, the fact remains that Unity randomly put itself onto the user’s PC, replacing GNOME, and confusing and upsetting a lot of people. Shouldn’t Ubuntu have stuck with GNOME as a more PC-oriented desktop, with Unity reserved for touchscreens and as an optional download for those who want to try it?

Absolutely not, and here’s why.

A month after Ubuntu 11.04 alienated Canonical’s userbase, Fedora 15 arrive bearing GNOME 3  Holy shit was that desktop a mess. At the same time Canonical was developing Unity, Red Hat was pushing the next version of GNOME into a form that would also work with both mice and fingers. And it was terrible. Linus Torvalds famously called it an “unholy mess” and switched to Xfce, though he has since gone back to using it after the Frippery extension made it usable again by restoring GNOME 2 behaviour.

Now, in fairness to GNOME 3, it, like Unity, has improved quite a bit since its debut, but the lack of any sort of menu icon, its smooshing together of the meta and meta-W functions, and hideous, massive, intrusive, useless top panel still inhibit its use as an everyday desktop. However, it seems to work very well on tablets, which is the way things are headed nowadays.

The Unity and GNOME 3 debacles seem to have raised some bad blood between Canonical and Red Hat. Mark Shuttleworth claims he came up with the idea for ?Unity first, and Red Hat happened to independently come up with a similar design some time after. Adam Williamson, however, counters that the ideas for GNOME 3 were already available and under discussion when Canonical conceived Unity. I’m inclined to believe Williamson, as this PC World article reports Canonical splitting away from  the GNOME team over differences in opinion as to how the desktop should look.

I think Canonical made the right decision here. There was (and is) a lot wrong with Unity, but it’s far better than what GNOME 3 ended up becoming. I thnk that if Fedora 15 had appeared before Ubuntu 11.04, the rage poured out upon Unity would have been a bit less if for no other reason than that people would be able to say “Well at least it’s not GNOME 3”.

Meanwhile, the team responsible for the Ubuntu-based Linux Mint saw which way the wind was blowing and made their own highly divergent fork of GNOME 3, resulting in the absolutely awesome Cinnamon desktop. At the same time, other developers took the code of GNOME 2 and came up with MATE. In fact, the rage induced by Unity and GNOME 3 collectively seems to have spawned a great diversification in Linux desktops as users sought something they could actually use.

But there was one other factor in the Unity saga that caused major controvery. That’s right. It’s time to talk about…

The Amazon Search Lens

This is the gesture that really turned people against Ubuntu, and caused people to revise their opinions of Canonical from incompetent to malicious. Richard Stallman famously slammed this as malware  and whatever you might think of him, he was absolutely right.

The Unity Dash includes several ‘lenses’, which allow you to narrow your text search to specific categories. When it first appeared, these were All, Applications, Files, and Media. There were also other lenses available for download, suck as Wikipedia.

Then Ubuntu 12.10 appeared, bringing with it a host of new lenses. There were now seperate lenses for photos, music, videos, social, and, most troublingly, Amazon. This one lens would allow you to search for books on Amazon, and due to the way Unity works, it would also display Amazon search results in the All lens.

Understandably, people went ballistic.

According to Canonical, there was no danger of privacy here, as the results would have to pass through Canonical’s servers, which would encrypt them so that they would be presented completely anonymously to Amazon, but in such a way that Canonical would still get a commission on any product bought via Dash. Critics countered that this wasn’t good enough, and we only had Canonical’s word that data was secure; furthermore, the fact that Canonical was harvesting user data was a major cause for concern in itself. Then there’s the fact that when you’re searching for programmes or files on your own PC, you don’t want to have things you can buy cluttering up the results.

Now, it was certainly possible to turn off online results, but the option was hidden in System Settings -> Privacy. One could also remove the lens altogether (as I did), but this requires knowing that each lens is a seperate package – and since Ubuntu targets beginner users, people who might wish to do this would not be aware it was a possibility.

Really, the fact that Canonical coded the lens at all is not in itself terribly objectionable. Don’t get me wrong, Stallman would still criticise it due to this dislike of Amazon, but as long as the lens was just an optional download, nobody would have minded too much. But that’s not what Canonical did.

Canonical wrote the code, then installed it on users’ machines without asking, and then had it switched on by default, so you didn’t know you were sending them your data until ladies’ underwear started appearing when you were searching for your CD writing tool. In response to these criticisms, Shuttleworth resonded (emphasis added):

We are not telling Amazon what you are searching for. Your anonymity is preserved because we handle the query on your behalf. Don’t trust us? Erm, we have root. You do trust us with your data already. You trust us not to screw up on your machine with every update. You trust Debian, and you trust a large swathe of the open source community. And most importantly, you trust us to address it when, being human, we err.

This is not reassuring. People had serious, genuine concerns, and Shuttleworth just waved them away, saying “Dude, you already trust us, so there’s nothing to worry about.” What he failed to realise is that trust must be earned, and once earned, it can still be lost. By forcing all Ubuntu users to send them their data, Canonical acted in remarkably poor faith, and the Ubuntu project lost some major ubuntu. This would have been easy enough to fix in three steps:

  1. Apologise.
  2. Remove the Shopping Lens from Ubuntu downloads
  3. Make an official rule that any new package which passively1 transmits a user’s data will never be added without the user’s explicit permission.
But instead of this, Shuttleworth dismissed concerns, declaring himself to be right and all others to be wrong. What with this and Unity in general, he seems to be getting rather disturbingly like Steve Jobs.

But it’s not just users Canonical is pissing off. It’s also majorly upsetting developers. That’s right folks, it’s time to talk about…

Mir

Since 2008, a loose group of hackers have been working on Wayland, a display server for Linux (and, eventually, BSD) to replace the aging X. In short, Wayland would be superior to X in every important way. The GNOME and KDE developers are all up for it, and Fedora will be switching to it soon enough.

Canonical supported Wayland in the early to middle days, but then a couple of months ago, announced that they would instead be using their own display server, called Mir. This managed to annoy absolutely everybody, and points to a growing case of NIHS, as Wayland is nearly complete, works well, there is no reason Unity can’t work on it, and Canonical doesn’t seem to have anyone capable of writing a display server. The KDE and GNOME developers are insistent on using only Wayland (and possibly X), utterly disreagrding Mir.

Mir seems to have even taken some of Canonical’s employees by surprise – apparently it was developed in a secret back room and sprung on everybody without warning1. This is sleazy.

Most likely scenario, Mir will trundle in in Ubuntu 14.04, hang around for a couple of releases with nobody outside Canonical doing anything with it, and be replaced by Wayland in 15.04 or 15.10.

However, neither the Amazon deal not Mir is the worst, most shameful thing Canonical has tried to pull. No, that honour belongs to…

Landscape

Landscape is a for-fee systems management framework for managing Ubuntu servers. Fair enough so far – Ubuntu Server is surprisingly popular, so it makes sense that Canonical would offer a  paid tool to help manage a network of servers running their product.

The problem comes in the EULA. To quote:

1.1 In consideration of You agreeing to abide by the terms of this Agreement, Canonical will permit you to access and use Landscape to manage up to fifty (50) computer systems solely for the purpose of internal testing and internal evaluation.

1.2 You will not allow any unauthorised access to Landscape or tamper with or insert any data or information into Landscape which may cause any damage to Canonical or third parties, or be unlawful.

Right away, this violates freedom 0 by instilling a time limit on how long you can use the software. Yes folks, Landscape is proprietary software. I’m honestly surprised this has gotten so little attention – Canonical has up to now been on occasion stupid, but one generally felt that its heart was in the right place, and it really seemed to take seriously its goal of providing an entirely free (as in freedom) solution.

Then it turns its back on this laudable goal and develops its own proprietary system.

People have defended this move on the grounds that Canonical has to make money somehow. Fair enough. It does make some money off tech support, but I don’t see why it has to resort to proprietary software in order to do so. After all, Red Hat became worth one billion dollars last year, and not only did they do it entirely with free software, they actually buy the rights to proprietary software and release it under the GPL. There are twice as many servers running Ubuntu as RHEL, so one would think that Canonical could at least pay its own employees out of that.

Knowing that Canonical is developing and pushing proprietary software leaves a bad taste in my mouth. I don’t know if I’m ready to abandon Ubuntu just yet, but Trisquel and Mageia are looking more attractive every day.

So, having alienated a huge chunk of both its users and its developers, and betrayed its principles, does Ubuntu have anything left? Well, there’s always…

Community

When Ubuntu appeared, the community was its main selling point. In those days, Linux users could come across as somewhat hostile to newbies. Mark Shuttleworth recognised this, and saw an opportunity.

From the start, Canonical encouraged the development of a friendly, involved, welcoming online community which would patiently help newbies with their troubles. It is this that fueled its explosive growth in the early years, and which also dispelled the idea that Linux was hard. Mandriva in the 90s had made an admirable attempt to get Linux perceived as easy, but Canonical is the company that finally opened the floodgates and brought Linux into, if not the mainstream, then at least somewhere that Windows and Mac users might actually notice it.

One factor in this popularity was making Ubuntu phenomenally easy to set up and use. Choosing Debian as a base was the right move here – a bit of automation and streamlining, and Debian’s massive repositories, brilliant package management, and ability to run on any hardware under the sun gave us the most user-friendly distro out there.

The other factor was, as I just said, the community. If you were having trouble, the fora were there to lend a hand. Lots and lots of knowledgeable people gave a kind and gentle introduction to what became a steady trickle of new converts. Even if it was just a simple “Look here“, this went a long way to popularising Linux among non-sysadmins.

Moreover, I think this had a positive effect on the wider Linux culture. I don’t know if people came to expect such friendliness on other distros’ boards, or people moved from Ubuntu to other distros and brought this attitude with them, or other communities thought this was a good idea, or other communities saw the rise of Ubuntu and wondered what made it special. What I do know is that as Ubuntu got more popular, the message baords for other distros, and for Linux in general, got more friendly. Even the Arch forum is friendlier than one might have expected, and Arch is geared towards people who are already Linux nerds.

Unfortunately for Canonical, the community is starting to desert it. Not quickly (yet), but they lost a lot of people to the Unity debacle. The Ubuntu-based Mint has taken Ubuntu’s crown as the official newbie distro, and the Mandriva-derived Mageia also made some waves in recent months. Mint and Mageia are thriving by doing what Ubuntu did in the beginning – offering a solid, easy-to-use operating system coupled with a great community. And in that, I fear Canonical may lose the one thing that originally set it apart.

Can Ubuntu survive?

Perhaps. Canonical has a lot to make up for, and I think they will have to significantly reorganise how they do business.

So Canonical, here are some ways you can regain people’s good will and market share.

  1. First up, release Landscape under the AGPL. Keep charging as much as you like for it, but allow others to improve and fork it. You know, like they do with your other products.
  2. Any programme which passively passes on user data to a server can only be added by the user’s specific invocation.
  3. Drop Mir and resume contributing to Wayland.
However, if Canonical really wants to stay in business, I think they might have to leave home users behind. They should do what Red Hat did in 2003 and, instead of devloping a general-purpose distro, sponsor their successful, community derivative. Just as Red Hat moved to sponsoring Fedora, so should Canonical shift to sponsoring Mint. Then, every so often, take the code from Mint, tody it up a bit, tweak it, and make is super-special awesome, and release that as the new Ubuntu version. Needless to say, in this scenario, Mint would remain as the home user operating system, while Ubuntu would be a server- and enterprise-oriented OS.

The paid technical support that comes from the enterprise packages is Canonical’s primary source of income, so if it can divest itself of the requirement to support a home OS as well, Canonical might just be able to pull itself back from the brink.

It also wouldn’t hurt that unlike Mark Shuttleworth, Clement Lefebvre actually seems to know what he’s doing.

1. As opposed to programmes such as web browsers and email clients, which only transmit data when the user explicitly tells them to.

2. The problem for me isn’t that Canonical changed their mind, but that they didn’t (or the developers weren’t allowed) to tell me! If you know for 6 months that you’re not going to do something you said you would it’s rude not to tell people. It now sets our schedule back and that’s really really frustrating.

=-=-=-=-=
Powered by Blogilo

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s