Open Ballot: what would you change about Linux?


We're gearing up to record a new episode of our great Linux podcast, and you - yes, you! - get to have a say. The question we're asking is this: if you had the resources, what single thing would you change? Would you merge KDE and Gnome? Would you introduce a new package manager? (eek!) Would you find all mentions of "Linux" and replace it with GNU/Linux?

If you'd like your views read out on our podcast, please post your answer below. Make sure you include a name, and please avoid running off a large list - pick one thing and one thing only!

You should follow us on or Twitter

Your comments


I have two desktops, three laptops, and a netbook. All (except one) are running some flavor of Ubuntu or Fedora and one is running #!.

Unfortunately the highest spec desktop is still running Windows because I need it for Steam. This is a pain in the posterior as I feel the machine is wasted on Windows and I couldn't be bothered dual booting it to be honest.

If Steam do release a port for Linux it'll probably have a subset of games available. Like the Mac port which only has HL2, CSS, DoDS, and a handful of others available.

Not good enough.

@ Grimly Penguin "~ when he

@ Grimly Penguin "~ when he could hack an international credit card company running Windows? "

They use Solaris in the data centres.

@ both Grimly Penguin & Tracy Holz,

I disagree, it's one thing that annoys me, go to launch an app / open a file and offers to rename it for you. It much better as it is, F2 or Right click to rename.

Rolling release of the mainstream distros

It's been talked about many times before, but I would like rolling release for the mainstream distros (ubuntu, fedora, et al) with stable LTS.

Oh yeah, working sound ;)

Every distro to use .deb as

Every distro to use .deb as there package format

usability and polish

I love Linux and would not willingly use anything else -- but it still has a way to go as a user experience if we are to attract more 'non-technical' desktop users.

For example, with things like file chooser dialogues, the devil's in the detail. We should look to identify simple yet significant things that Win and Mac do better than us. For example, when opening a graphic file in Gimp or Firefox (or other GTK application) why cannot users choose the file that they wish to open or upload from an array of previews? At present you must click on individual image files in a list to view individual thumbnails. Also, the different ways in which the file dialogues in GTK and Qt applications behave can be a problem to inexperienced computer users.

These are things that many acclimatised, technically confident Linux users may not think twice about. However, if you put my parents or parents-in-law in front of a Linux desktop, they will raise their heads quite quickly, as the UI does not have the consistency of the one they are used to in Windows.

1. Please can we have more co-operative working between teams working on Gnome/GTK and KDE apps, so that users of one DE or another have as unified a desktop experience as possible?

2. Please can developers pay great attention to usability analysis -- watching, studying and learning from the ways in which ordinary computer users (i.e. not tecchies and computer enthusiasts) experience the Linux desktop?


I've thought about it and there's nothing I'd change about Linux itself.

The freedom to modify Linux in whatever way you see fit is it's strongest asset, making Linux a breeding ground for innovation.

Part of the attraction for me has always been the reward of learning how to do it for yourself. I learned more about computers in 2 years mucking around with Linux, than 14 years of using Windows & MacOS as professional tools.

Having said that package standardization would be a fantastic step, meaning no matter what distro you're messing with, you all have access to the best software available and I presume it would save a lot of time and energy for developers.

The lack of a professional quality video editing program is the one thing that ensures I need to keep commercial operating systems around. But as a Linux user, I find booting my windows box up to be a frustrating and painful experience.

Pointing Out

I'd like to point out that a lot of these features that people are pointing out already exist in some way or another. Also perhaps instead of whinging about a problem why don't you, the user, try to learn how to fix the problem (if it really is one) instead of relying on developers that you say are "slow".


If I had the resources to change anything in the GNU/Linux world, it would be to tell the entire computing world that it exists.

You do have a choice in what operating system you run on any computer you own.
No, it's not illegal to remove the OS that the manufacturer stuck you with.
You do not have to suffer with the constant assualt of malicious code from the Internet.

That's it, simply tell the world GNU/Linux exists and is ready for them.

"Linux" and replace it with GNU/Linux?

I'd personally get rid of all mention of "Free software", "Linux", "GNU slash Linux" and any other term that seem only to serve old political causes and factionalism.

I'm not sure those terms "Going forward" are going to convince the average Windows user who is already suffering from Stockholm syndrome to switch to another syndrome slash operating system slash faction slash...

lack of single file installs

there really should be a single file install!! who wants to compile software just to use a simple program?

I Hope...

a) To have a graphical installer (sometimes you have to be a semi-god to install some distros)

b) A unified install/remove application manager (I really dislike yast)

Edit instructions, HOWTOs, and tutorials, and certify them

There should be a group of volunteer editors who read and improve tips and suggestions and instructions available to Linux users. The result could be a kind of certification of rationality attesting that the prose is genuinely helpful -- that the HOWTO or guide meets an editorial standard.

That standard would exclude ambiguity, vague statements, assumptions that the reader has knowledge that is advanced far beyond the understanding of those who will be using the information, and so on. As things stand, sloppy grammar, confusing sentences and just plain poor writing make portions of many instructions puzzling, confusing, or even unintelligible. The reader may be sent chasing irrelevant facts, when he should have been kept on the shortest path to the proper procedure.

Informative texts should be first of all task-oriented; then they should be clear, straightforward, make few assumptions, and avoid all ambiguities.

Asides can deal with interesting details; when the text is written on several levels of importance, the reader's confusion is likely. Complex issues can be made clear if they are dealt with discretely, rather than as an amalgam.

A text written for the relatively inexperienced can be understood by advanced users and experts, while the reverse is not true. That means that "dumbing down" the material by making it accessible to all harms no one. Many guides should include glossaries with references; very few do.

Writing instructions that can be understood only by those who do not need them is silly.

I have to say it: if we had paid attention to good language usage, we would not use the word "parameter" in describing commands, nor would we say things like, "The data is corrupted." Those are examples of ignorant usage. But let that pedantry go. Errors become correct because they are used and become agreed upon.

If writers voluntarily submitted their work for editing and requested that the finished product be granted a certification of comprehensibility, I can't see that anyone should or would be offended. (Don't want your precious text changed? Don't submit it.) If Linux is worth the trouble to learn, it's worth the trouble to write its instructions clearly and rationally. We remove the bugs from our software; why not do the same to our documentation, guides, HOWTOs and yes, even man pages?

Steam has been ported to

Steam has been ported to Linux. Check it.


I have myself a home studio, in which I record, mix and edit music. At present I have to use Winblows as I have a specialised soundcard that I can't find any fully functioning drivers for in Linux. I would dearly love to move away from that old OS, but even with drivers that would make my hardware work, I'm not sure I would be able to fathom out Jack.

So for me the most important thing I would change is the sound system. It needs to be a combination of simple (plug in and play) and complicated (in that it will handle what ever inputs and outputs I throw at it!).

I hope in time it will come, but for me that time can't come quick enough!!


Finally a proper Open Ballot question.
Please stop with all those "Do you like a Banana more than an Orange?" Questions. Thank you.
On topic:
Nothing. Linus and all the other guys are doing a heck of a job. Everything else is not in their hands, and it never should.
Maybe i would force Canoncial to actually do proper work and contribute, but that is all.

Most important for potential new converts

Whenever I get an opportunity to motivate a potential new user I offer the move to Linux. Let me try to focus on a few flaws I continously notice:

1. Scanning
This is continuously an issue.

a) Simple Scan in german environment adds extra space to an A4 page when scanned by multi-page feeder.

b) gscan2pdf scans A4 but with some scanners (HP MFP for instance) continues to scan empty pages when the last page from multipage feeder already scanned.

c) XSane does not support multipage feeder anyway (supports multi-page documents but you must scan page by page - one click per page).

2. Confirmation question on deleting files
The appropriate option in nautilus is ignored since ages.

3. Hardware support
I still face plenty of unsupported or badly supported hardware (notebooks, WLAN cards, sound issues and gadgets people bought somewhere). I try to force people before trying Linux to buy the right hardware but it is not very easy to find hardware where I can be really sure that everything works - just look at Netbooks - in many cases you have to fiddle around with the WIFI drivers. Similar applies for printers. Even, if Samsung for example is usually very good, there is sometimes a little manual work required. Or now on 10.04 I experienced a problem with my docking station (display irritation), ...

4. Many applications for the same task available but most have serious flaws.
E.g. the new Thunderbird 3.0.x is crap.
A lot of changes have been introduced in Thunderbird 3.0.x that make things worse (display only one recipient by default, additional buttons in header area that are continuously moved out of sight for longer addresses, ...).
Bring back the old version! Tried to switch to Evolution, but I noticed major flaws there also (not efficiently notifying about mails in different folders than inbox etc).

OK, this is not really a Linux issue, but it is similar to scanning - it would be better to focus on one or two applications for a single task and make that stable and good.

Note: The most potential new users I face are home(-office) users and very small companies (~3 people).


Collect all the Linux (GNU/Linux - who cares?) users guilty of using the terms M$/Windoze/Winblows you know the sort. Tie them up in a big sack and drop them in the sea. It's childish and pathetic playground name calling. I came to Linux in my 40's and find all that childish finger pointing at Microsoft to be tedious and off putting.

If Linux is to be taken seriously, then its advocates need to f-ing grow up.



Fix the overcomplicated mess that is Linux audio. Replace the complexity with some simple straightforward solution that doesn't require a PHD to comprehend.

Of course, being Linux, the complexity should be left as an option for those who actually need it.

Easier network integration

I agree it can sit alongside win and is os x. It doesn't need to fight. It does need to integrate into networks other than windows easier. Right now you have to jump through hoops to integrate into a Mac network, which is an os closer to Linux than win. But all the effort seems to be in samba.

First of all we need to

First of all we need to realize there is no Linux, there is only open source community that develop code.

Second, Linux need some female touch to get nicer controls and skins.

Third, KDE must simplify main menu, it's still feels like a driving a train or something when using it.

Fourth, Linux should learn for a lot from first generations of MAC X, so get busy.

native graphics editing software for Linux!

I am not writing as a Photoshop fangirl - but the lack of any native alternative to the Gimp for image editing will embarrass Linux if it ever starts to achieve greater desktop penetration. Very sadly, many non-specialists used to tools like Photoshop Elements will be bewildered by the Gimp's interface, while power users will find it many years behind proprietary packages (and not just Photoshop) in terms of features.

Greater desktop user numbers could encourage Corel and Adobe to port their products to Linux. And if there was a really good modern Open Source graphics app I am sure that many of us would be willing to pay to use it!

Change nothing!

As far as I'm concerned as a SimplyMepis user, I'd change nothing that I can't already easily change if I want/need to....

I would however change perceptions of GNU/Linux. In fact I already try as best I can to inform colleagues/friends/family etc. of the alternatives to the proprietary cr@p they currently use. Unlike a previous commenter, I don't think increased 'market share' is a bad thing and will result in malware problems. The number of users doesn't result in malware issues, inherent security features (Or lack of...) does. Having said that, ill advised users that try to circumvent sensible security practices (By running as root all the time for example!) would be a problem!

Linux is the Public Library for the Programming Arts!

Instead of marketing Linux as an alternative to Microsoft and Apple operating systems, Linux distributions need to be touted as essential operating systems for creative development of our future leaders in the field of software development and emerging technologies.

Open source operating systems and software enable the next generation of "Bill Gates and Steve Jobs" to develop future innovations unthinkable without access to code.

In this respect, Linux is the link between innovation, efficiency, and education - who wants a nation full of students who know nothing about software code or hardware outside of plug-and-play devices? Linux allows a new generation of computer literate leaders to emerge and transform the landscape of computing for the good of all people. If Gates and Jobs couldn't experiment in their garage with computer parts and code, then what would have happened to their creativity and development of the personal computer?

Linux is the public library for the programming arts and its success is intimately linked with its ability to give our students access to creating emerging technologies.

Common knowledge

Not so much about Linux, the kernel, or GNU/Linux, the operating system, but more about the community.
It would be really nice if every Linux user built a LFS system once, so that they had a taste of what really is going on underneath. That would be really nice.
Also, people should learn about *why* there are different distros/package managers, rather than just complaining that it doesn't "just work". It's an ecosystem; there's a niche, so it's filled.

PS - why so much hate for the devs? They're people too, no matter how irritating broken releases are. File a bug report!

Standardization )stop the tribalism / distro wars)

The strength of apple and ms is their standards. The strength of GNU-Linux is the verity of inputs, but only with cooperation. Cooperation means standards. Standards means interoperability, and strength in unity.

Robust Applications

Programs that do not break when the kernel changes or conversely, kernel changes that do not break applications. I am thinking specifically of Skype video that worked fine under one Ubuntu release and quit under the next. And yes, thank you Linux community for providing a work around.

For everyone wanting an

For everyone wanting an inline replace function in the file manager ... its time to stop using dumbed-down GUI -- aka GNOME. KDE has had this feature for a long long time.

The changes I'd make to Desktop Linux


For me the priorities at the moment are:

Linux Audio
Better Video4Linux

If Linux is to truely become a genuine alternative for Windows or MacOS we need commercial games.
Now I know that the choice of the developer but at the moment there is no incentive. Linux Audio is crap, although 3D accelerated drivers are now becoming acceptable.

Steam4Linux may well be the start but say if Ubuntu chose to develop an Android flavour which has games we might see the gap bridged.

Just a thought.

Too confusing for first-timers

There are too many distributions to choose from, this is a good thing in way as you have so much choice but there can be too much choice. This for me as a Linux newbie was quite overwhelming and a big issue in whether or not to use Linux in the first place. I think Linux seems to be creeping out of the geek playground and is in a sufficiently robust and user-friendly state it can become more mainstream. However a lot of forums suggest fixes using the command line which seems archaic and makes you realise how windows and Macs have progressed so much over the years.
I agree with comments about hardware/driver problems, OpenSuse didn't work for me and I couldn't figure it out after several hours on forums regarding video drivers but Ubuntu worked out of the box and runs fine. Maybe this is my lack of understanding but there seems to be a need for a significant amount of inherent knowledge before you can use Linux almost at all.

Legislate Linux

Have the EU legislate that microsoft may not have a monopoly of OS on computers. This would oblige hardware manufacturers to offer alternatives which would make the world a fun place. This would also disolve the contractural stitch-up that MS has with manufacturers. (It could be a double edged sword for Steve Jobs).

Id like to see..

A nice ready made replacement for MS Access. Yes, I know the DB Admin folk will jump and down and say that mySQL is the way to go, but honestly, most of the time people don't need that level of complexity, especially if you are required to muck around with a LAMP configuration - that'll scare off the meekest of Windows converts!

Also, a fully functioning MP3/OGG ripper, converter and Librarian all in the one package. Similar to Media Monkey in Windows. By having all these separate applications in Linux makes it too cumbersome. One package to do the lot.

dual boot all PCs with linux

Make all PC sellers put a Linux distro on the PC alongside Windows (or just Linux) -- Linux should come pre-installed on all PCs so that people really do have a choice of what OS to use. If they do not like it then they do not have to use it, either install a different Linux distro or just use the pre-installed Windows.

Get Adobe, Corel and Quark on board

Get Adobe, Corel and Quark on board to make their software work with Linux. Even if it is just a version that runs in WINE/CrossOver, so long as it works. Currently it is not possible to run the latest Adobe Photoshop, CorelDraw or Quark XPress on Linux, but these programs are vital to creative types who do a lot of graphics/DTP work. Also the same for Sony Vegas.
If we get these software makers on the side of Linux, then Linux would be taken seriously, especially by creative people. We want to use Photoshop, Dreamweaver, CorelDraw, Illustrator, InDesign, Quark XPress and Sony Vegas, but without having to run Windows or buy Apple hardware.
Of course we would have to pay for the software still, but the linux versions could also be cheaper than the Windows or Mac versions.


Corel used to make a version of Linux, which ended up as Xandros. If Corel were supportive of Linux in the past, maybe they could still be.

Re: Get Adobe, Corel and Quark on board

You could say cheaper if it was some kind of hardware with linux as the OS but in software terms the price would be higher because all the linux distro's combined wont get you a double digit in market share.


I wish all the distros and factions in the community (KDE vs. GNOME, FOSS/FSF etc.) would stop fighting each other, and focus on nuking MSFT.

And Ubuntu still won't make me a fudgeing sandwhich...

Get Adobe, Corel and Quark on board

I agree with getting big name software makers to do Linux versions, or else make the programs run easily with WINE. The Mac platform has always been popular with designers because of the great applications available. Eventually they were ported to Windows too, and so now designers have a choice of Mac or Windows. Get the design software on Linux, and get Linux to be more attractive to designers in general, and that will help a lot.

MultiCore Support

I would like to see all linux software take advantage of every cpu core on the system. Maybe through a kernel option or a separate application that could detect the number of cpu cores that are available, 2,4,8,24, or even 240 cores of some gpu's and use all of them to run the application at a speed that the system is capable of, instead of seeing one core pegged at 100% and the rest of the cores sitting at 3%.
Multiple cores are not a new thing but no one seems to know how to make use of all of them at the same time.

Bang on! Myles Wakeham!

Myles Wakeham's response is exactly what I beleive to be the most important issues that need changing.

Myles, you're observations are bang on!

K-Alec :)

Multicore seconded

I second goodrench's suggestion. What's the point of multicores when the compiler can only deal with one. Multicores only work with multiple processes coexisting.


OpenMP looks very cool but is that something that the software creators will have to use or is that something that I need to use to compile everything I install to utilize more than one core?
I would like to see something that is just built into the distro. Works out of the box. No configuration(lots of single command lines to figure out how many cores you have).
I just want them all to work. Am I being unreasonable?
OpenMP looks like it is for the software developer. If we leave it up to them to write the code for multi-cores, it just won't get done. Dev's already have the skills and the opportunity to fix this but they just don't do it.
That's why I propose something to run at boot time.


My netbook has an intel atom cpu and it shows up as two cpus.
It is very painful to watch my netbook chug along when only one of these cores is being used. Don't do it for me. Do it for my netbook.


I'd like Linux to be used by more than a tiny handful of geeks.

1. Fewer distros - having so many serves the egos of the developers more than the needs of users.
2. Better hardware support for peripherals - perhaps if we had few distros, manufacturers would pay more notice to our pleas for proper linux drivers


Many of the previous posters have valid points and some of them are the reasons (problem descriptions) why FOSS will stay where it is.

1. There is a saying that says 2 dogs fight over a bone and the third takes it. GNU/Linux and various applications would be far better than Propriety code if they stopped there ego-centric agendas. M$ has there fights, but they need users to make money and if they solve the users problem the Micro$oft way they have the users money.

2.It's a nightmare to standardise on Gnu/Linux. As a business decision maker I have had to write my own standards. The file system is a mess, the inode tables are a mess, Documentation is a mess, the source code is a very big mess.

Want companies to rewrite code to Linux? Give them a stable API or standardise up to a GUI, simple actually if you remove your emotions from the situation. Companies should actually write for wine and wine should move down to kernel integration. Then they have 1 code base to maintain and users can move to GNU/Linux. I'ĺl probably then start to think about writing something for GNU/Linux and document it.

Why port a app from 1 distro to another fixing bugs that doesn't exist in the main branch? 2 things happen in my business when duplication occurs, or originator or the duplicator gets fired or we solve the problem and stop the duplication. It costs money to duplicate work and Governments are experts at duplicating and misappropriating tax funds.

3. Why should I give a good word, financial support and bug info when something that worked in the previous release, is broken and on inspecting the bugzilla find a moderator/maintainer/developer that gives a (stinky thing) about the current users and says they should wait for the next release. O'ja I am talking about Ubuntu and launchpad here.
Should I roll back to the previous release, pick another distro or move to a OS that I know has fixed those bugs many years ago?

4. Firefox, mySQL, PHP, Apache if I think fast, I can exclude from the problem description:
I will rollout Gnu/Linux with FOSS tomorrow if I know the users will have the same or better user experience then vintage XP with 3rd party software.

My rugby couch always said the line when all our fancy tricks started to cost us points "get back to basics".
Fix the bugs, fix the irritating bugs that has been there since the start, add the expected basic features, then only add the fancy. Then and then only can we as a community start thinking about calling desktops and FOSS software replacements and alternatives to proprietary software.

Image viewers that cannot play animated gifs are photo viewers in the state of getting there just like Pulse and Alsa that are still think about playing the click sound 5 clicks back.


The void left by the absence of a QuickBooks replacement in the Linux ecosystem is *immense* to small businesses. The vast majority of small businesses use QB, and because QB tightly integrates with M$-Office they then also continue to use M$-Windows.

Many in the small business community are fed up with Microsoft products, but can't wean themselves off them because of all the dependencies on Microsoft products. I see QB is the linchpin to nearly everything else. If a suitable replacement for QB could be written for small businesses, one that can import QB files and is extremely easy to use, then I think an avalanche will begin.


Better 3+ monitor support, More mature replacements for Dreamwaever + fireworks on linux.

Pour a lot more money into OOfice, WINE and Virtaulbox.

Ask oracle to relase ZFS on a compatible license.


There are a few Linux programs that I install through the package manager that do not give a menu entry, even though the programs have a GUI. All Linux distros should be able to detect when a program is installed and give it a menu entry, even if the programmer did not specify where it should go.

Geeks or Suits?

Every year we get the question of "Is this the year Linux goes mainstream?"

The problem is that Geeks want to be associated with cool projects, but if we want a Linux on company desktops for more than just Open Office aps then it need some quality business software.

Where is the Linux equivalent of Sage accounting software?

I remember that in the early days of Linux Format many of the reviews were of commercial Linux software, now that is very rare.

A few thoughts on a few of your posts

Games. Ok, what about bundled, specialized wine to run the games on? What if game makers only had to adjust their games with wine in mind, instead of 10 or 15 different distributions and conditions? What if they were coded that way to some degree from the ground up, but stuck to Windows-only methods or resources when it was beneficial to gameplay to do so, and pulled some workarounds to do equivalents in wine-run copies? Cheaper and easier and probably easier than you'd think, in fact. Game makers could throw a little money at wine developers and so on, get special versions built for generations of games, etc. This would motivate the next group some of you commented on...

Hardware makers. They would be pressured by game makers if the above were so, and would start releasing specs and at least a little code or some libraries and proof of concept drivers. Enabling gaming in Linux like above would create far more of a market, and markets push innovation and openness in most cases. Also, with the shift to hardware as a commodity (RAM always has been, but not so much with other things) and the move to some cloud and server computing, hardware makers will have to squeeze every penny out of their transactions, which dictates an invitation to hobbyists and the curious.

Integrated/standard interface. NO! Never. Seriously. That's the charm of Linux to non-geeks. That each distro looks different, feels different, just like OSX and Windows look and feel different. That's the standard, beyond basic functionality, that the users adhere to. Let them find a range of things and pick what they like. Keep e17 (e16 for me personally) and KDE and Gnome separate. Maybe find a way to use the same themes in more than one DE though. We really do have to try to see through the average consumer's eyes and realize as long as it works, the next thing they care about is how does it look? There's no right answer to that.

What else? OOBE. Out of box experience. There was a deep focus on that with the early 486 and Pentium PC's. Intro scripts and basic tutorials and some fanfare and what have you. We don't need to go to that extreme, but a couple ideas for ways to do something other than dumping a newly installed user to a blank desktop with their 'start button' in the 'wrong place' could be worked up. I have no idea what, though, I'm not a UI designer or anything.

Also, while doing it our way or ways is nice, we might work up a distro that does it their way, exactly like they expect, similar names and mechanisms and so on up-front. I don't like this idea, but there it is anyway, we claim to be a democracy verging on a little meritocracy, so no point hiding bad ideas.

This really is a marketing driven world. Linux needs evangelists to the non-geeks as was said by others above. Linux needs less prophets and more salesmen and street preachers.


So many people here don't understand anything about what makes GNU/Linux what it is.

Linux has its end user salvation - Ubuntu. What Linux has needed for a very long time was polish. It's fine having amazing features such as package managers, compositing window managers, good drivers coming in the kernel; but what was needed was an Apple like approach of keeping everything simple and beautiful.

As soon as a standard end-user (not you and me) is given a painless, polished, consistent and elegant user interface, building on already fully understood paradigms, Linux will flourish.

GNOME was the desktop environment to do this, Ubuntu is the distribution. All that they need to keep doing is carrying on the exact same route and we'll see more and more commercial applications starting to distribute debian packages alongside the windows executables.

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.

Username:   Password: