Today is the final day of my “10 Days of Ubuntu 10.10 Feature Requests” series. See the series introduction here.
Day 10 – Paper Cuts, Paper Cuts, Paper Cuts
I believe that the One Hundred Paper Cuts project will end up being the most significant thing Ubuntu contributes to the Linux ecosphere.
Yes, that’s right – more significant than PPAs, Launchpad, Upstart, or any other Ubuntu innovation.
Think that’s crazy? Rest assured that it isn’t.
Here’s why.
There is no “one flaw” that explains why Linux has yet to reach mass market adoption. People can talk about any number of individual problems – be it hardware support, marketing, stability, whatever – but at the end of the day it is impossible to argue that either Windows or OSX surpasses Linux in all of these areas. For every person who argues that hardware support is better under Windows, another person will argue that it’s better under Linux, and the same could be said for any other macro-level problem.
So if there is not a “master flaw” that keeps Linux distros from becoming mainstream desktop OSes, what are other possible explanations? In my opinion, the primary problem is not any one failure in particular, but the sum total of a number of very small, almost-consequential-on-an-individual-basis problems.
I’m talking about paper cuts.
Ask yourself, for a moment, why Apple is commonly referenced as the de facto standard of technological usability. (I realize this is up for debate, but clearly the mainstream media and general public – at least in the U.S. – believe that Apple represents a high point in usability design.) Is it because of some singular feature? Some singular design?
No. It is because Apple sells products with a singular design focus: a simple, cohesive user experience. Every program looks and acts roughly the same under OSX. An iPhone doesn’t have twenty different interfaces for twenty different system menus. Your new iPod works in a way that is logical and easy-to-understand, even if you don’t know anything about the underlying technology.
Where Apple has excelled is in selling the idea that their products are fun and simple to use. Microsoft tries to do this, and so far has failed. Linux…
…well, I’m not sure how Linux distros attempt to portray this.
At the end of the day, Apple teaches us something very interesting about technology: that most consumers really don’t care which product is more technologically advanced. They care which product is more fun and simple to use. That was the real innovation behind the iPod, and arguably the innovation behind most of Apple’s products.
Now I am not an Apple fan, and I don’t think the Apple business model is the only way to operate. But it is difficult to argue with Apple’s general marketing prowess, particularly when it comes to their carefully crafted messaging of Apple products as being both the most fun and the most usable.
So why the lengthy rant about Apple’s marketing success? Because Mark Shuttleworth, the owner of Canonical, has stated that by the end of 2010 he wants Ubuntu to be competitive with OSX in terms of user experience. Not coincidentally, this article series is about the last Ubuntu release of 2010 (Ubuntu 10.10).
Some macro-level features have already emerged in Ubuntu that will address this competition in grand fashion. (Ubuntu’s Software Center versus Apple’s AppStore, for example.) But Apple has not achieved its status in the tech world because of any one magical feature, so while the arrival of these macro-level features in Ubuntu is clearly good, such features alone are not an indicator of the project’s future success.
The real measure of Ubuntu’s competition with Apple will be in the way that minor, seemingly inconsequential interface problems are resolved.
Or said another way: the real measure of Ubuntu’s future success will be the way it deals with desktop Linux’s many, many, MANY paper cuts.
Really, I view the resolution of paper cuts as the ultimate purpose of individual Linux distros. Major feature development is often best left to specific development teams (such as kernel, KDE/GNOME, or application developers). Distros must focus more on rolling other group’s major pieces of software into a coherent whole. Ubuntu has yet to do that as successfully as they need to if they want to compete with Apple the way Shuttleworth has outlined.
As an example, consider the paper cuts being addressed in the upcoming 9.10 release (Karmic Koala). Some of my favorites include:
- F-Spot puts photos in Photos folder not Pictures folder
- Changing workspaces via scrollwheel on desktop is problematic, especially when using touchpad
- Default folders inside Home Folder (e.g. Documents, Music) should have special icons/emblems
- “Auto eth0″ confusing for most people
- The thumbnail of an image should not be bigger than the image itself
- Spellcheck in Gaim, Evolution, gedit etc doesn’t recognize “Ubuntu”
- Annoying beep on shutdown using “System -> Shut down…”
- ‘Open With’ Nautilus list is unsorted
Note that none of these paper cuts are likely to be mentioned in any of the major 9.10 release announcements or reviews. Additionally, no new user is likely to drop Ubuntu because of just one of these issues.
But when you look at the sum total of them (and this is but a sampling of the ~100 being addressed), you realize that those are some pretty damn annoying quirks, and all of them together represent a major barrier to the viability of Ubuntu as a major desktop OS.
So please take me seriously when I say that Ubuntu’s ability to deal with paper cuts will represent its single greatest success or failure. Canonical must know that specialized distros will almost always surpass Ubuntu in terms of hardware support, performance, stability, and any number of other macro-level features. As such, I hope the Ubuntu community doesn’t spend an undue amount of energy trying to compete with other distros/OSes on those merits alone.
Because if desktop Linux is ever going to truly succeed, it will not be because of some as-of-yet undiscovered “killer feature.” It will not be because of tremendous technical prowess. (If that were the case, you can bet it would have already won.)
It will be because it has successfully dealt with its many, many, MANY paper cuts.
Conclusion: The most important Ubuntu 10.10 feature will be the successful resolution of as many “paper cuts” as humanly possible.
<< Day 9 – Renewed Focus on Marketing
Related articles:
[...] Here is the original: Ubuntu 10.10 Must Fix as Many Paper Cuts as Possible [...]
[...] Read more here: Ubuntu 10.10 Must Fix as Many Paper Cuts as Possible [...]
Gnome 2/3 must address the papercuts regarding the treatment of network shares in GVFS. It’s a mess; for instance, if you try to drag a file from .gvfs to Rhythmbox, the import will fail because Gnome has translated the URL to smb:// in the background.
Although you can work with network shares the same way as local files, there are some Gnome programs and operations that will bring up an error message “Cannot perform this operation on a remote file”. It’s the old “law of leaky abstractions” – Gnome abstracts away the differences between remote and local files, but the way the system works or doesn’t work means you have to keep track of what’s remote and what’s local so YOU can treat the files differently.
That’s a most serious paper cut, especially now that home Windows users can finally create network shares.
And it’s a royal PITA for me too!
Thanks for this post. Totally agree!
Oh you can’t guess how much I agree with you!
F-spot sticking pictures in pictures makes sense. I don’t know why you need a different photo folder. Auto Eth0 is not confusing if you do just a tiny bit if research.
But that’s the very thing — Users don’t want and shouldn’t have to “do just a tiny bit of research”, they want it to ‘just work’.
I read through the entire ten days of feature requests.
Here are a few I would suggest.
1. Add a setup network section to the installer.
While the current installer works well it leaves functions such as network setup until after the install. This leaves Ubuntu/Kubuntu users at the mercy of some occasionally pretty bad network managers in Gnome and KDE. Defaulting to DHCP is taking the easy way out and not everyone uses it by default. I have had nightmare after nightmare of wasted time with nearly every release in this area which generally leaves me hand editing everything in config files after the install.
2. Setup SAMBA during the install CORRECTLY and with the correct permissions.
Ubuntu/Kubuntu is also taking the easy way out in this area. Samba is one of the most critical functions used in Linux to allow interoperability with Windows systems and networks. Its a mess in Ubuntu/Kubuntu and as with the network setup above I am generally hand editing the config files to get it working worth a damn. Do this during the install and do it right.
3. Test ALL software included in the repositories for functionality with EVERY release. You would think this would be a given but every new release of Ubuntu/Kubuntu has its share of applications that either are broken or buggy beyond belief. Recent example? GLabels. Great application. One of a kind in Linux but non functional in the 9.04 release with no update (that I know of) as of yet, even though the fix is out there. This required a hand install/compile session to get working on my system.
4. Give Kubuntu the SAME amount of attention and polish that you give Ubuntu. Personally I can’t stand Gnome. If you like it more power to you. Use what you like. My point is if the company is going to release a KDE version then by all means give it the same amount of support as the Gnome version. Don’t just splatter it against the wall as say “Help yourself!” as appears to be the case with Kubuntu.
5. Stop including unpolished software in releases.
Example?
The 8.10 version of Kubuntu with its obviously not ready for primetime version of KDE. While I was able to easily see that KDE 4.x had incredible potential in the 8.10 release I also easily could see it was no where near the functionality of of the older KDE versions.
The right way to have done this?
Release a “test” version along with the “production” version so those that wish to and have the time can test away to their hearts content. When it is actually ready to satisfy desktop users needs THEN release. (do’h, what a concept eh)
Yes, it would require more work on the part of Canonical but that IS THEIR job, not ours. I’m more than happy to help out but I need to have a system that works first, Testing is much lower in my priorities.
6. Fix printing in Wine.
Nuff said.
7. In general and in closing.
To Ubuntu/Kubuntu/Shuttleworth et al
Stop acting in such an unprofessional manner with release quality. Get your act together NOW while you still have time. These issues are inexcusable in a so called “desktop” “quality” operating system. The longer this goes on unchecked the more of an embarrassment you’ll become to Linux and the more turned off the general public will be towards Linux as a viable choice. Remember YOU have placed yourselves in the position of being the desktop Linux choice.
LIVE UP TO IT!
What are Ubuntu requests and what are GNOME requests?
I ask because I have Kubuntu running next to 2 other KDE4.2 distros and its almost identical. Seems like all distros do is fine tune a few things like icons and themse and wallpapers.
The real choice is not at the distro level but at the desktop level.
Btw, one of the reasons I never installed Kubuntu on our home machines and have gone with Mandriva and PCLinuxOS is the dual boot that Kubuntu offers in all its DOS black.white glory is pathetic.
Seems like a perfect example of NOT being user friendly.
In my experience the old Unix FHS directory structure is one of the biggest hurdles to new user acceptance of Linux. It’s often in fact a deal-breaker as folks coming from Windows and Mac expect to see all their program’s file in the one sub-folder under eg /Programs/ProgName/Version, not randomly dispersed all over the place. I know this is more than just a paper-cut, and most of us have just learned to understand it, we can’t expect to entice the less tech-interested to our cause with this train-wreck of a directory system. It may have made some sense for the time-sharing Unix of 1978, for Ubuntu 2010, it’s just plain awkward and time-wasting.
Canonical would do well to hire you…
This was a great series of posts, thanks! I was glad to see you put so much emphasis on “small” usability problems, because – as you say – they affect the user experience more than we often think. The big individual things that you brought up were also well chosen.
Here’s one I’d like to add. I’m not sure if should be called a paper cut or not, but anyway it’s a simple annoying problem. What I’m talking about are the user’s permissions regarding hard drives. As long as I have used Ubuntu (from the first day it was available for download), I have had to struggle with permissions to use drives – especially hard drives – in many situations. For example recently, when I created a new partition onto one of my hard drives, I wasn’t able to save anything into it, because the drive was now owned by root. It’s not possible to change permissions in this situation from the GUI, so I had to use the terminal (and search for the right command, which I don’t remember by heart, and scratch my head for a while to use the command right) to change the permission.
Another problem with permissions came up when I moved my torrent downloads to another hard drive. At least in Karmic (had a different drive set-up before), I have to give my password to be able to use my “secondary” hard drives. If I haven’t done this, and start Transmission, it won’t see the unfinished files, because the drive they are placed on is not mounted. When I mount the drive, Transmission checks all data that has been downloaded so far. It takes quite a while if I happen to have 30 GB of unfinished files. If I find the time, I’ll make a bug report (it’s surprisingly time consuming to make a good bug report).
Forgot to say that I too think that Canonical should hire you :-)