Skip to main content

What Ubuntu should mainly do for 11.04

Posted by fabriziogiudici on October 26, 2010 at 7:15 AM PDT

Reading around, many people are expressing their disapproval and worry for the latest directions of Apple: developers for the deprecation of Java and even some regular users in Mac communities are a bit worried about the honesty of Steve Jobs assuring that the new Mac OS X App Store will always be "just another way" to install applications.

In a way or the other, it's clear that a number of people are evaluating alternatives to Mac OS X, at least in perspective. One alternative is Windows 7; the other is Linux. Speaking of regular users, Ubuntu is the principal option, thanks to the particular attention to usability that it demonstrates, right from the installation.

Usability which, unfortunately, is still far from enough. While the installation process is smooth, there are still lots of tuning configurations that should be done manually, including (on my MacBook Pro) the activation of "closed" drivers, which include the WiFi connectivity and the graphic card. The sensitivity of the touchpad out-of-the-box is bad, miles beyond that of Mac OS X; it can probably fixed by tweaking the config files, but again it's not something that a normal use expects.

Sometimes installing a software is still an engineer job. For instance, if you try VMWare Fusion on Mac OS X, it installs and runs with a number of clicks. On Linux, VMWare Workstation needs compiling some drivers... erg, in 2010? Yes, the process is nicely hidden behind a graphic progress window, but unfortunately it miserably fails. And you discover that there are syntax errors due to changes of the kernel, and you have to apply a patch. Of course, five minutes for an engineer, but a total showstopper for a regular user (and don't think that VMWare is a stuff for engineers only: especially in these periods in which people are trying for alternatives, nothing better than a virtual machine could help transitioning to a different operating system with an incremental approach - I mean, some non-engineers have done that in Mac OS X for transitioning from Windows).

Last but not least, the look and feel is still geeky. Surely, it has been improved, but not enough. The alternatives are even more geeky, and it's appalling that Canonical wasn't able to hire some good graphic designer for this job (with "good" I mean somebody who has got in mind the end user expectations).

No surprise that Linux failed to take advantage of the Windows Vista fiasco and is still relegated to a little niche in the desktop world (talking in general, of course, including end users). Now there's another chance: will be able Canonical to catch up with the next release? Note that these are things that require small amounts of work, so they would fit with the deadline. It's a matter of awareness and political decisions.

Related Topics >>


What Ubuntu should mainly do

The first thing I learnt in my dealing with Linux (I have been Linux-only for ten years) is that you are in it for nobody else than for yourself. Do you want to use it? Fine. Do you want to look / work / behave differently? Fix it. Do you want somebody else to fix it for you? Ask very nicely, but don't demand them to do something for you and don't rely they will do it. Do you hate what I just said? Switch to Windows/Mac OS X immediately, but never say what should Mark Shuttleworth/ Canonical / Fedora / Red Hat / community / some-particular-developer should do. They won't. And don't pretend you are speaking for somebody else ("ordinary user", "general consumer", "my mom"). Unless you have tons of numbers and data to prove, you don't know what you are talking about.
Second, don't try to stuff your Windows/Mac programs to Linux. They usually don't work that well even when they pretend to do so. OK, I am biased (being a Red Hat employee), but my experience is that 3rd party binary blobs just don't work well.
vmWare IMHO works well only for highly stable server installations (RHEL or SLES; and of course, my employer suggests that KVM would work there better). I think you are really much better with native solutions. I use kvm/virtual-manager, but many people are pretty happy with VirtualBox.
On the same note, 3rd party binary-only drivers (graphical cards or wifi) are PITA. Unless you need some REALLY high-end stuff, you are much better with native drivers (it feels to me that wifi cards are more less solved problem now that even Broadcom caved in and supports at least some of its cards natively; if your card is not supported, then really ... miniPCI cards are couple of bucks and they can be replaced, but majority should be supported already). Concerning graphics card the situation is more problematic, but nouveau driver seems to work pretty well for most of my users (somehow, maybe not with compiz and 3D stuff) and radeon driver is even better.

It may well be that Ubuntu

It may well be that Ubuntu will never be a consumer operating system like Mac OS X or (never mind, that's not a consumer operating system either). But why do you care? Is Ubuntu a good operating system for a Java programmer? Does it run on the hardware that you like to use for your job? Does it run the tools that you need speedily and effectively? 

In my case, it has delivered quite well. I use a Thinkpad that puts my Macbook to shame in terms of weight/performance/keyboard quality. Ubuntu runs on it just fine. And it supports all the tools I need, most of them better than Windows or Mac OS X. Sure, there is the odd moment when I need to recompile something or edit a config file, but so what? I don't care if my mom would freak out at the prospect--my mom isn't a coder.

In a perfect world, Mark Shuttleworth would hire better graphic designers, and all Linux developers would be gifted UI engineers. But we don't live in that world. Get over it and pick a platform that gets your job done :-)

The desktop wars are over.

The desktop wars are over. Actually, the desktop was the battlefield of the 1990s, and the clear winner was Microsoft. Now it's the year 2010, the battle is about mobile Internet and so far this race is being led by Apple which appears to be the only company with a very clear vision and concept for the future. If you are spending developer resources on desktop platforms, you are investing in the past. Let's face it: Except for maybe some insurance companies and banks, nobody ever cared for Java on the desktop, and also no user or consumer ever cared for a Linux desktop. If you want to do something for end-users on the foundation of Linux and Java, you better join Google and work with and on and for their Android platform - it's the only thing remotely resembling a Linux/Java desktop that has a chance of succeeding. In a fistful of years from now, desktop computers as we know them will become an endangered species whose only niche will be developer desks. The next major evolutionary step is within reach, and the Linux and Java hackers should think about what people will be doing in two to five years from now instead of what they have been doing ten years ago. The folks in Cupertino and Mountain View are pushing this evolution while the folks in Redmond and at Oracle still don't have a concept for what's coming. And sadly, the Linux folks still believe that Redmond is "the enemy" while in fact they've become a toothless paper tiger. Maybe they were a powerful force ten years ago, when Bill Gates was still running the company, but nowadays they are just a poor shadow of themselves and, like Big Blue IBM, they're not a threat to anybody anymore and are just limping behind the development. Oh, and maybe one more thing: A normal user won't setup a computer. They either buy pre-installed systems or have an administrator who configures the machine for them. It doesn't matter if it's hard to install a proprietary driver on a system, because they will never do it. They care for APPLICATIONS - and whether the platform has them. You want Ubuntu to succeed? Make sure that there are enough hardware vendors that bundle their machines with Ubuntu. That's the way to succeed. Easy to install graphics drivers are not a priority - that, once again, is last decade's lost battle.

@theothersteve You're

@theothersteve You're referring to the typical mistake the Linux community does - which can be summarized as "ignoring the marketing". First, most of end users don't ever know what FLOSS is, and BTW this is a big problem we have to deal with (especially with the mobile platforms). Second, you can teach people what's the FLOSS advantage once you've estabilished a communication channel, and the best way would be to have them pick your product. But to have them pick your product you need to "sell" it and people will decide for the look and feel and the "coolness" factor. That's the first and foremost; all the remainder could come later.

@winfriedmaus Java has got an excellent penetration in the industrial desktop segment. Just have a look at the showcases for NetBeans and Eclipse RCPs. I don't think the iPad will change a lot things in the industry in the short term, while of course its huge impact is with end users and there the scenario is going to be revolutionized.

So there are two different

So there are two different markets here:
- developers: they really shouldn't care what the platform looks like, as long as it is functional (and linux is perfectly functional for a Java developer, I use it every day), they should be aware of the reasons why some hardware can be difficult, and more broadly, why openness should be valued over looks (Apple's recent shenanigans as a case in point)
- everyone else: I'm sure Canonical are more aware than most of what the issues are with linux adoption in the broader community, you really don't think they're spending money on UX? The Ubuntu UI (and linux in general) has improved over the last few years, and will continue to improve into the future. The real issue is not how pretty the interface is (Windows has looked like ar$e for decades, hasn't hurt it's adoption), it's interface familiarity/expectation and support by hardware vendors.
And in the end the desktop is not really where it matters, have a look at the smart phone and tablet markets in a few years and we'll see how well linux is doing.

Linux won't gain a

Linux won't gain a significant desktop market share as long as people can't appreciate the subtleties of the free software ecosystem (in short, probably never). Until you appreciate exactly why closed drivers require a couple of extra clicks, and why the integration of proprietary software is sometimes awkward, you're not going to be able to get past the initial hurdles and appreciate the benefits of an open system (and shouldn't really waste your time).

So it's not really about when linux is ready, it's about when you are. Until then there are alternatives.... alternatives that sometimes decide to deprecate software you depend on. Perhaps that's a greater usability issue?

Well, a really good

Well, a really good alternative for VMWare is VirtualBox. I have been using it for a long time now and works perfectly here.