DEAL OF THE DAY at TigerDirect


What Would You Do to Improve Linux?

The Linuxsphere is a wild, wonderful and adventurous place. By its nature alone, Linux is often considered a maverick. It’s been my observation that Linux users tend to explore and take more risks when using their computers. For better or worse, it appears that Linux users have been associated with the hacker elite. I’ve even thought of it as being the wild west of tech. I’ve referred to using Linux as computing without a safety net.

Linux magic wandThere were days past when I let my idealism and fervor dictate my thoughts and actions. I saw magnificent potential in Linux as a desktop operating system. With all my heart and soul, I evangelized for a competitive presence of Linux on the desktop. Like many idealists, I ignored the facts and barreled ahead as if by sheer will I could blow obstacles to dust that would dissipate in my wake.

That didn’t happen. Pure intentions or not.

You and I have talked about Linux in the marketplace a good deal over the years. We’ve examined why it can’t be profitably marketed and we’ve discussed how that lack of everyday exposure to the everyday computer user has kept Linux on the desktop from succeeding. To be honest , back when Microsoft was twisting arms and making deals with the OEMs at large, Linux wasn’t ready for the desktop.

I’ve spent a good share of my time asking myself what would have to change in order to make Linux on the desktop a viable choice for the mainstream user. I became curious enough to ask you a question: if you could wave your magic wand and change only one thing about Linux or even the Linuxsphere in general, what would it be? Let’s take a look at what some of you had to say.

Thomas King, server meister extraordinaire gave some serious thought before answering:

“My opinion is definitely going to be a bit unpopular: the linuxsphere needs depth in many different areas. Stop reinventing the same applications (I read a couple weeks ago of a brand new file system manager. Seriously, stop it!). Improve what we have and find ways to stretch their features and integrate into other parts of the os without creating dependencies or regressions.

“Demand more out of your desktop than pretty themes, a package manager, a browser, and a working wireless.

“Linux is pretty rock solid (i.e. years of uptime). Its services are solid as well. Now make the services work smartly with each other but not dependent on each other. Whether that means services become more aware of each other and can automatically work with each other if allowed, that’s up to other visionaries. Increase the depth not the width, and definitely don’t create more silos.

“We need more moments of ‘Now why didn’t I think of that’ in Linux.”

My buddy Tom touches upon something that has been true ever since Linux became functional. A unified package/software management situation would at least diminish one of the barriers to new users. When the new user decides to try another distro, he’s often faced with a different way to handle packages than he’s used to, be that .rpm, .deb, YaST, Portage or Synaptic. To the new user, this is chaos and it’s not difficult to see how this thing in particular can scare a new user off.

Spencer Hunley brings up a topic that falls into my field of vision, so to speak. I am dealing more and more with older folks. Some of them have physical conditions that make it hard to use a computer the way you and I do.

I can relate personally and here’s why; chemo and radiation treatments left me with some challenges in sight and focus. I can’t look at a monitor with a backlight for more than five minutes without becoming horribly ill. Someone in the Linux community stepped up and altered a 32 inch Samsung monitor so that it uses side lighting instead of the normal backlight. That’s the only way I am able to sit in front of my monitor and share this with you now.

Spencer voices his concern here:

“We need more accessibility and assistive technology software to include and incorporate people with disabilities into Linux. While there are fantastic examples available today (orca, etc.), we need to keep improving and offer an alternative for those who may not be able to afford and/or want to avoid being locked-in to certain devices and companies. That, and as +Paul Bucalo stated, possibly more marketing.”

Let me add that text to speech is being improved by great leaps in the Android development world. Spencer noted Orca, a GNOME screen reader. I appreciate the work and I kneel before the absolute genius of Mark Mulcahy, the original programmer for Orca. Sun Microsoystems took the project over when he left, but then Oracle found it in their laps. I wouldn’t count on Oracle to do anything to improve Linux-at-large. I simply hope that this work can get done without the corporation hindering the development of great tools like this.

I’m not the only one that thinks the corporate world can do more harm than good. Fellow Google Plus guy Alessandro Ebersol feels that corporate control can be a bad thing for the Linuxsphere:

“I would love that corporate influence be kept off GNU/Linux, since their influence is never good for the communities, but for themselves (the corporations)”

I’m not saying that all corporate influence is bad, but with more and more businesses entering the realm of free and open source software, abuse of open source code is bound to occur. Overt actions such as removing open source code or using code outside of its license isn’t rare in our realm.

Something I’ve really never thought of was also mentioned when I asked your opinion on the one thing that could change the status of Linux on the desktop. Ryan Gallagher put it this way:

“Leadership. The kind that inspires loyalty and unites people in pursuit of a cause. I’m not thinking exclusively in terms of an individual human either. Google has (arguably) made Linux the most widely used consumer mobile OS. They’ve done such a good job of streamlining the user experience; most people don’t know they’re running Linux. But implicit in the responses above, people would like to see desktop Linux enjoy the same successes as Android. Setting aside the argument that the desktop is dead, the only way the PC market will be won is if the battle is led with the same focus, vision, drive and cold hard cash as the phone wars.”

Others offered their short but accurate suggestions.

“WAY more Microsoft Windows compatibility” — Ben Vrooman

“I would love to have a good, competitive video-editor. There simply is none in GNU/Linux.” — Axel D.

“Less dependence on CLI tools for software management.” — Bojan Landekić

“Proprietary graphics drivers need to be improved. In addition, +NVIDIA need to stop holding back features in Linux just because the same feature doesn’t work in Windows.” — Claire Farron

And these three guys are men after my own heart…

“I would go back in time to the days when IBM had Linux commercials and steadily build from there. It was a sad day when they stopped promoting Linux and just started marketing their services. They might have given Microsoft some competition over the years.” — Bob Pianka

“I think the best thing that could happen to Linux is early adoption of the ideas behind open source and lowering the barriers that seem to exist in tech. Get the kids involved, make them aware of what the challanges are and see if they are willing (with guidance) to meet them head on. I think we’d be surprised at the outcome.” — Chuck Green

“My personal preference is to have Linux install apps as a self-contained unit in its own directory. Also, I’d like to see apps be able to be installed in a user’s home directory if it’s for their own personal use. That way there isn’t a need to have root access to run certain programs.” — Chuck Cave

So yeah, if wishes would be fishes…

Still, the things talked about or offered here are not without value. If I wanted to brag, I could point out a couple of changes that took place after we talked about it here, but I won’t. This isn’t about any particular writer or website. This is about affecting change for the better.

Affecting change for those kids coming up behind us. They’ll be here before we know it.

The following two tabs change content below.
Ken Starks writes and publishes The Blog of Helios, a finalist in our Best FOSS or Linux Blog competition. In addition, he's the person behind the Reglue project that refurbishes older computers and gives them to disadvantaged school kids in the Austin, Texas area.

46 comments to What Would You Do to Improve Linux?

  • This is a short essay on Linux desktop failures, from my web site.

  • Cory

    I have Four.

    First: Unified Packaging System. That way people that write software only have to package for “Linux” and not for .rpm vs .deb vs .pkg vs however many others. It is too much work for every distro, to have their own repositories with their own version of an installer. It is a nightmare for companies to write software for Linux because of that.

    Second: Make Gnome admit they had no idea what the average desktop user wanted. Admit that Gnome 3 was better built as a tablet interface and adopt Cinnamon as the Desktop GUI and keep Gnome 3 as a tablet GUI I’ve seen other people with this same viewpoint.

    Third: Somehow make packages that can be installed on any distro. Have one single repository container that is controlled by, qualifies and oversees the security and quality of the software packages and then categorizes the repos. If you want a non-free game, you install that repo, or allow that one program from that repo to get updates. But something needs to be done to allow 3rd party apps and drivers to be installed effortlessly. If the qualified package doesn’t work, it is the distro’s job to make the distro compliant to the software packages.

    Fourth: Get OEMs to install Linux… which one you say? Well if my first three issues were solved, then it doesn’t matter! Just as long as it isn’t Ubuntu (I don’t like their views on everything). I would accept CentOS, Mint, Fedora, or Arch.

  • CFWhitman

    I think that Thomas King definitely has a point. Sometimes in certain areas I do find myself thinking it would be nice to have three really good programs in a category rather than eight mediocre ones.

    There are areas of development where some fairly amazing progress has been made, like raw photo develpment. Darktable and Raw Therapee have become very competent programs in a relatively short time, and Raw Studio provided some competition, though it doesn’t seem to have as much life as it once did.

    Video editors are just starting to really get interesting, but they are not there yet. Again you might feel that there are too many splitting developer time.

    I’m not sure, though, whether the idea of several applications splitting developer time is really the difference between the good categories and the mediocre ones. It may be a fallacy to think, ‘There are so many developers that will work on programs of “A” type, so if there are ten “A” type projects those developers will be split too many ways.’ I’m not sure it really works that way in real life. Part of the strength and the weakness of Free Software is that developers do whatever is interesting to them. I think sometimes that means you will find a lot of different projects doing a similar thing.

    One thing that I do think about Linux on the desktop right now is that it’s got a foothold on the desktop because more and more really young people seem to have used it and are not so intimidated by it. I see more people who have heard of Linux and who have used it than I’ve ever seen in the past. As desktop computing becomes more of a commodity and less of a money-making platform I expect to continue to see a greater presence of Linux than ever before.

  • Guys, this has all been done before. Almost every OEM on the planet is shipping some GNU/Linux legacy PCs. Anyone can ship a .tar.gz archive and have it install and run on a GNU/Linux PC, even in the user’s directory (tar strips the leading “/”). There’s scarcely any need for extra applications if you use one of the big distros, like Debian. They have about 40K packages. You can choose from a whole bunch of desktops to suit what you want to do how you want to do it.

    Businesses do need to get involved with FLOSS. They can afford to hire programmers and salesmen.

    The last barrier to adoption in a wider way is space on retail shelves. In my nearest city, almost all retailers stock that other OS even if they can’t sell it. M$ pays them to do that. On the other hand there is one chain of retail stores that does carry GNU/Linux PCs and they allow consumers to configure and see the price of that other OS or install GNU/Linux distro of choice. Go out and tell retailers you want GNU/Linux or No OS PCs. Problem solved.

    Once OEMs and retailers ship/sell GNU/Linux in a big way all of the stupid little problems listed above will fade. OEMs and retailers need products that work and GNU/Linux will with a lot less cost, malware, re-re-reboots and lock-in. Ken, people do make money selling GNU/Linux: all the major OEMs, all the retailers who provide retail space and yes, the developers who get employment either producing the code or providing services related to it. Android/Linux has succeeded in getting space on retail shelves and no one grumbles about these little things any longer. Get GNU/Linux on to retail shelves and the same will happen. Folks will appreciate the faster performance of GNU/Linux (native code) over Android/Linux (interpreted code). Folks will appreciate the advantages of Free Software on lower cost ARMed hardware. Price may be less important here but in emerging markets price is key and FLOSS is winning.

  • Ok, read everything before getting that huffy look on your face. The FIRST thing that has to happen is this. Make ALL mainstream distributions have the same look and feel out of the box. Dont’ take away the wonderful variety after that, let them hack it to pieces. But make the initial install consistent. It makes it more palatable to support at the OEM level, and easier to help the NEW user. Nothing harder than trying to help someone on the phone who installed some random “Ultra Pokemon Minimalistic ZooTree Edition” of Linux. Now the 5 years I worked to get them to try it out is lost and they never will come back. Bottom Line: Make it START OUT the same. Until that happens, we’ll be called a “Hobby” OS for the desktop market.

  • archuser

    First i would remove all the effects , animations in ubuntu(like XP)
    Double click package install
    Make the desktop more sleek and fast.
    Better compatiblity with MS Office.
    Run option to quickly launch apps.
    Make games more easily available
    Better support for resolving linux issues.

  • commu

    I dont think people will switch to linux when they see no incentive in doing so, they are already familiar with windows and they dont see any advantage in switching to linux.

    So linux foundation should provide incentives to people who encourage linux use among their community. like providing certificates , awards etc.

  • msos

    We should follow MS Policy in order to increase linux market share, since canonical has the resources it can supply a copy of ubuntu in a pen drive with every laptop sold.

    Try Ubuntu and you have a chance to win exciting prices.

  • Hunkah

    Quit calling it GNU/Linux for frig sakes!

    GNU is NOT Linux. They are two totally separate projects created by two totally different people, with different agendas and belief systems. RMS himself says this. He’s just pissed off that Linus is getting all the glory for something he tried to do first.

    Then there’s the argument that GNU was used in Linux… well YEAH! That is what RMS was trying to do! Get everyone to use his stuff FOR FREE! Freedom to use, amend, change, create… Where’s his freedom now? He should really practice what he preaches. He promotes this freedom, except when you’ve used anything he created, then you have to stick his big fat GNU in front of it. Well I disagree!

    GNU means Gnu is not Unix. That project is a full out operating system created by RMS and the Free Software Foundation. It has nothing to do with the philosophy of Linux besides the GPL. A licence that RMS promotes people to release software in. Where does the line end? Should everything GPL called GNU? No! That would be stupid! Because everything isn’t GNU, INCLUDING LINUX!

    So knock it off for frig sakes!

  • Bernardo Verda

    I think that the biggest, easy-to-fix issue for Windows refugees would be some consistent attention to making sure it’s easy for the Linux newbie to add an additional harddrive to their system — and that they can easily discover this.

    And I’m not talking about mounting expansion or back-up drives on external USB or eSATA, nor about mounting under /media, but rather, as an integrated drive or partition that mounts automatically at boot, at a user selectable, designated mount point.

    For example, a user might want a separate “storage” drive devoted to their music and video files, and/or games, and/or a bittorrent partition, accessible to all accounts on the box, or just designated users. Or perhaps they might need a drive devoted to serious, “must-be-kept-separate”, work data, mounted somewhere under that user’s own /home partition.

    A nice, standard, GUI tool for easily adding and managing such volumes, that works on — and is easily found on — any desktop environment (eg. Gnome, KDE, Xfce, LXDE, and plain X) would have made a huge difference to several fairly computer-savvy Linux newbies I know, who ended up going back to Windows.

  • 2 options: Lobby to phase out printers or simplify Linux printing management.
    Personally, I am in favor of all those little things were the Linux distributions just make it better.
    One of the things that always annoy me on a Windows machine is that I do not see a miniature of the image when I am attaching an image file in the browser. This is a small thing, but an example of better polish.
    Thunar custom menus and the KDE context menus are powertools in their own right. There is a lot of stuff just executed right.
    I think there would be an excellent potential for something like the Chromebooks with strong tools to tie into something like Exchange or ownCloud. I am not talking about the hardware itself – but a lightweight OS that would go on anything from netbook to powerhouse and make use of the cloudy stuff. OCOS, an ownCloud-connecting OS, is a wet dream.

  • Mike

    @Robert Pogson

    > “Anyone can ship a .tar.gz archive and have it install and run on a GNU/Linux PC, even in the user’s directory (tar strips the leading “/”).”

    There are very real problems due to differences between Linux distros which you are glossing over here due to differing file locations and init systems, among other things. Your example only holds for trivial applications. Try running something like MythTV like that.

    >”There’s scarcely any need for extra applications if you use one of the big distros, like Debian. They have about 40K packages. You can choose from a whole bunch of desktops to suit what you want to do how you want to do it.”

    This is partly true, until you need something specific or newer than what is in the repo, then the problems resurface. I myself run Debian. It’s one of the only distros I trust to separate out the non-free junk I don’t want on my system. Even with all 40K applications, I still find the need to go outside the repos for specific things. That’s when the library versioning game/nightmare begins.

    I would like to see a further strengthening and modernizing of the filesystem hierarchy standard. The fact that Debian and Fedora will be using the same init system in the future is a positive step (even though I dislike systemd personally). If they could do the same with packaging systems it’d be another huge step. While I don’t want to see a loss in variety of distros, there are a few basics that it’d be great to see all the major ones agree on.

  • Mike

    I’d also like to remark that even a unified packaging system wouldn’t solve all (or even most) compatibility issues. I’ve seen people completely hose their Debian systems by installing an Ubuntu package.

  • jymm

    I also agree, a unified package and management system would be best. Not only from a Linux point of view,but also for hardware manufactures. You could then buy a printer, scanner, camera, router or other peripheral with a CD to install drivers and extra dedicated programs, as one CD would work for all.

  • Here is my response from the thread on It relates very closely to what Thomas King wrote. Linux needs depth, not endless new versions, reinventing the wheel, and fast upgrade cycles:


    The endless upgrade cycle every three minutes or so is great for tech writers and those who constantly like to tinker and play with the latest. For the other 99% of us who actually have work to do, it’s somewhere between a nuisance and impossible.

    One of the nice things about not writing about Linux for a couple of years and concentrating on real work for real businesses and government is that I could completely ignore the upgrade treadmill and keeping up with the geeky version of the Joneses. Nobody cares. No sane business manager will run an OS that only has nine months of support. Heck, nine years seems short. A once every six week browser upgrade? That’s funny.

    So… yeah, it’s been nice living in Red Hat/CentOS and SUSE Linux Enterprise land. Things are quiet, they are sane, and they mostly just work. If a business has a compelling reason to need the latest something or other (PHP was a good example for one of my clients) then, yeah, I find a trusted source with updated packages or I get to make some myself.

    Upgrade to the latest and greatest anything? Let the kids play. It’s not for me.


    As far as consumer space is concerned, Robert Pogson hit the nail on the head. Until Linux is in big box stores next to the other guy it might as well not exist for most consumers. Where it is available (Android, ChromeOS) it does very well indeed.

  • Mike

    I don’t think even getting Linux into stores will help much. It certainly wouldn’t hurt but I think there is something else which holds Linux back…expectations.

    Most people with PC’s barely know what version of Windows they run (if they know at all). People buying a PC expect it will work with whatever software they’ve already used. That means Windows software almost exclusively. It doesn’t matter if Linux has equivalent or better software in almost all cases…it isn’t the EXACT SAME software, which is what they expect. People expect all PC’s to run all PC software (i.e. Windows software). To them, a different operating system might as well be a different wallpaper.

    Android and Chrome are less encumbered because they tend to be found on devices that people do not associate with the term ‘PC’ and so the same expectation isn’t there. People expect them to act different and run different software, even if they dont exactly understand why.

  • Eddie G.

    My “take” on it is simple and might seem a bit unrealistic but since we live in a free country I think I can voice my opinions without fear of being persecuted! LoL! And in a nutshell it’s this: Linnux is fine, yes, there IS room for improvement, and yes, they COULD do things a bit more “Windows-ish” (because not for nothing, that’s what a LOT of this sounds like, make this more compatible with that, make this as “easy’ as it is with other OS’es…etc.) But for the life of me I cannot figure out WHY anyone would want things to be that way, while even I have to admint that there are more choices within Linux than I will ever be able to use, that doesn’t mean I want to see those choices go away! I would prefer a dozen “Not-So-Nice” desktops than to have 6 stellar ones. And the reason why is simply this. What works for one doesn’t necessarily work for someone else! All those wou would like to see things go one way with Linux….there’s a distro out there that suits you to a tee….and for those who march to the beat of their own drum? there’s a distro out there for you as well. I think Linux is fine, and doesn’t need to be altered to meet the “demands” of any, hence it being so appealing in the first place, and while some distros have their “favorites” when it comes to desktop environment, web browser, email client, programming app…IDE….text editor…etc..etc..etc. The real “beauty” of Linux is…you cna uninstall and install YOUR preferences instead! Mind you I don’t downplay or negate anty of the opinions that have gone on before, but being a Linux user since 2004 I can say I’ve seen it evolve, and I’ve seen what can hapen sometimes when the developers try to make EVERYONE happy, and living on a planet with BILLIONS of people on it?…that will NEVER happen. This also pertains to the whole “Year Of The Linux Desktop”…listen people, there’s no REASON for Luinux to be on a desktop in a company or corporation unless the company WANTS it there! And don’t be fooled there are a LOT of them that have Linux on their desktops, they might not choose t9o broadcast that info, as:
    1 – It gives them a competitive edge from the start by allowing them to hold on to more money to use for developing their product/services or market their clients products/services and
    2 – It allows them to focus more on what to DO instead of spending frivolously on different applications and add-ons from various organizations and companies.
    Now realize, I’m not just whistling Dixie here, I happen to work in Manhattan NYC…(that big ‘ol apple!..LoL!) and I can name some pretty well-known companies that use SUSE/openSuSE…and a few companies that have CEntOS on their desktops and of course a LOT of companies that have RHEL / Ubuntu / CEntOS and SUSE servers in-house. So for all those who keep searching for this Year Of The Linux Desktop, you need to stop looking, there might not ever come a time when EVERY desktop in the corporate or private sector runs Linux on it. There’s just too much “MiApple” out there with too much of an advertising budget to have that happen, but you know what? I’m good with that. Those that know….”Know”…and those that don’t…DON’T! For those who would like to see Linux on more machines, just continue to spread the word, show those that “don’t know”…and let them see just how food it is! For those who are able to…start LUG (Linux User Groups) in your neighborhood. Don’t wait for some company to advertise Linux, YOU ARE the advertisement, take your laptops and tablets to the local Starbucks / Panera / coffeehouse and let others see your OS. I for one have been chastised today and had to undo some work I did recently. A friend of mine had an aging laptop, and when it continued to freeze, and get hung on apps and programs he asked me to “do whatever I could” for it. Immediately I began talks of Linux Mint or Xubuntu with him….

  • Eddie G.

    …cont’d – He took to it like a fish in water, he was onboard all the way, and after installing Linux Mint on his machine and watching him enjoy it for the better part of a year, he only recently called me to tell me I had to put Windows back on the laptop. Apparently he has upgraded to something nicer, and he wants me to put Linux Mint on that, but his old laptop is going to his girlfriend, and no matter how he tried, she just couldn’t use Linux Mint, she somehow coudln’t grasp concepts of Thunderbird accessing all of her three email accounts in one interface, nor could she understand the update process and why she had to always enter the admin password to have it run. So I grudginly re-installed Windows 7 on the laptop. This just goes to show you that some people CAN and WILL use Linux if prompted and directed, shown how things work and how things interact, they’re open enough to be willing to try something other than Microsoft or Apple, other’s unfortunately are just locked into their mindset of what a computer OS is…and what its supposed to do. I use this as an example, that the Year Of The Linux Desktop won’t come abou8t until something “tragic” happens to either Apple of Microsoft, and since there are too many businesses that rely on them for their operation….this just won’t happen…..maybe in our children’s lifetime it will…but for right now…..most Linux users are content with the way things are, yes there could be SO MUCH MORE that could make Linux the OS of the Century, but a LOT of things would have to happen first….and those things are a long way off. I have a 13 year old son who uses Fedora Linux daily, he won’t use Windows…he knows about the LibreOffice suite, and loves the various desktoips he can install with his Linux, so I’m not too worried about him…nor the fact that he will be carrying the torch for Linux in the future. To me the most important thing Linux can do to make themselves more appealing to the masses?…is just educate the Next Generation because they will be the ones who carry Linux into the future/ Ok there…I’ve said my piece.,…I’ll step off my soapbox now.

  • Albin

    I started using Ubuntu / Mint about four years ago anticipating replacement of XP, and have gotten “end-user” familiar without being much of a techie. I quite like the diversity of DEs and have no problem with general OS operation or occasional use of the Terminal (normally cut and paste commands from reliable online sources.) I’d say the biggest obstacles to mainstream adoption are:

    1) requirement to reinstall at least annually to keep repositories up to date, that requires reconfiguring everything from the OS itself to all the major third party software settings. “LTS” may (or may not) keep OS versions going a few years but the repos go out of date much more quickly.

    2) failure of Samba as a local network interface. Linux would be much more attractive if someone would build an MS-like “wizard” front-end to recognize and configure Samba. I frankly gave up and limit Linux “networking” to cloud services like Dropbox.

    3) my experience of peripheral / driver support has been spotty.

    Nothing is going to cause Windows users or businesses reliant on MS Office or specific Windows-only software to switch, but Linux could be more appealing if it would stop lurching on to one “cutting edge” after another, and less glamorously support and build the best of current versions to maturity.

  • Albin wrote, “peripheral / driver support has been spotty”.

    Nvidia used to be a pain. It’s much better now. Most popular devices are well-supported by the kernel. No need for special installations.

    Albin wrote, “failure of Samba as a local network interface. Linux would be much more attractive if someone would build an MS-like “wizard” front-end to recognize and configure Samba.”

    Samba is not a failure. M$ deliberately varied its networking for years just to confuse the Samba group. Now, under government orders, they have published the specs so Samba works much better. The absurd complexity of CIFS/SMB networking is the result of years of abuse by M$. M$ itself has had many bugs and vulnerabilities in it. If you are using GNU/Linux, there is no need to use Samba at all. NFS works very well. So does SSHFS. I use SSHFS on my systems and the little woman can access her files from anywhere as if they were on her device.

    Albin wrote, “1) requirement to reinstall at least annually to keep repositories up to date”.

    Use Debian GNU/Linux. Their releases are about every two years and the repositories are current for a bit longer. You need to edit a couple of lines in a file to access the newest release and issue a few commands as root:
    apt-get update; apt-get install apt;apt-get dist-upgrade;shutdown -r now
    or use the GUI provided, synaptic. “Synaptic is a graphical package management tool based on GTK+ and APT. Synaptic enables you to install, upgrade and remove software packages in a user friendly way.”

    Debian has been around a long time, has hundreds of repositories around the world, supports many languages, has tens of thousands of software packages and has been very reliable for me.

  • Tom Tucker

    A relatively easy way to create a small home/business network of linux machines, ala active directory. It is available, but setting up the workstations to have something like login scripts, mapped drives, etc. takes quite a bit of work, and the setup is complex. There are products like zentyal, but it is commercial.

    A decent desktop publishing application. Scribus doesn’t quite have it yet.

    A decent greeting card maker would help on the home front.. Several of my friend and family have switched, and that is their major complaint.

  • Implied Consent

    There’s one constant that M$ and Crapple have: $$. $$ creates Marketing. *nix community has (IMO) genius in the point logic of interchangeability, innovation, problem solving – but lack – marketing skills in their projects.

    I can hear the eyes rolling, it takes money. No, it takes existing technologies and a bit of the skills the community already have … and a little imagination. A little Marketing refresher from your undergrad days:

    * Marketing research: Who is your market? What do they want?
    * Product Design: What does your product need to do to fit the needs of the market
    * Advertising: making the world aware of your product
    customer feedback: using feedback from your customers (and potential customers) to improve the first three.
    * Who is your market?
    * What is your project about?
    * A little crowdfunding for possible video media commercials?

    What I see with most *nix rolls when pointed to their project: a link to “xyz” That’s pretty intimidating if your trying to capture a desktop user.

  • […] Fortunately, the Linux community includes clear-eyed observers and thinkers such as Ken Starks, who recently penned a piece entitled, “What Would You Do to Improve Linux?” […]

  • e8hffff

    Obviously Linux is a freedom based community so no one can control it, but if I could I would stop software makers from releasing Windows versions of software. Example Gimp for Windows, stopped. VLC/ffmpeg for Windows, stopped. Why support Microsoft in spite of your own free community. People will simply not move to Linux if all the treasures are already on Windows. Making Windows versions is desperate and detrimental.

  • YetAnotherBob

    What does Linux need to capture the ‘desktop’? only one thing, Patience.
    “The Year of the Linux Desktop” that most of the press loves to either trumpet or bemoan isn’t some future date. It was the year 2000. That was the time when the desktop became ‘good enough’ for an average user.
    Since then, Linux has just kept getting better.
    Growth, however is a slow and organic thing.
    Back then, total share of Linux Desktop Systems was very low, around .01%.
    Prior to then, You had to have a technical bent to run Linux. I did, but I only ran Linux to learn the Unix Way. I am an Engineer. I knew that the Windows PC’s I had been using were limiting me for some things, and I was trying to find a system that would let me expand.
    I have had Linux installed on some computer or other since 1995. However, use was rare. That use has slowly climbed over the years.
    For Linux to take over, the Linux development community just needs to keep on doing what they are doing.
    When a group like Gnome takes a wrong turn, there will be others who will just come up and replace them. Mint has done that with Ubuntu.
    I now routinely see figures for Linux use on the desktop that range from 1% to 10%. The lower figures are based on PC purchases. The higher figures are based on web site statistics. Since I believe that most Linux Desktop Users are like I have long been, dual boot users, I think that the latter figure is closer to the truth than the lower one, however, that means that Windows plus Linux use is higher than the 90% that is non-apple. The 9% overlap is almost as large as the total Apple use.
    Most users don’t just jump into a different OS. They dabble for a while. That while may just be years, even decades.
    So, Ken, don’t worry. There will be no sudden shifts without something drastic happening. Instead, there is a slow drift in the market. That drift is continuing.
    Linus used to joke about ‘Linux World Domination’. LWD happened around 2004. We are now living in a Linux Dominated World. The High End is almost exclusively Linux. The low end is nearly so. The middle realm, where Desktop Computing lives is slowly changing.
    Personally, I expect the Linux Desktop to peak at around 20%, but that will take another ten to twenty years.
    The problem isn’t with Linux, it’s just that we have a long time. We lack patience.

  • Dan Saint-Andre

    I propose that every linux distribution come pre-installed with an “application migration” package. This package would enable an end-user to install and run any commercially available software simply by inserting the delivery media or launching the delivery download executable. As an aid to success, the package would identify any reasons why the commercial software install cannot succeed. The reasons would appear in a report that is (1) easy to understand by non-geeks, (2) suitable for printing, and (3) useful as a checklist for problem resolution by the end-user or their helpful “technicians.”

    Yes, “commercially available” is a code word for MicroSoft Windows and Apple Macintosh off-the-shelf products. However, there is no reason to stop there given the modular nature of all linux-oriented software. There is no reason why the migration package could not learn to launch any software presented simply by binding the appropriate {platform not-A} modules.

    Today we have “virtualization” from Wine(tm) and Crossover(tm) to Virtualbox, XEN and KVM. While these are great, they are too complex for most end-users. Virtualization also makes heavy demands on the end-user workstation hardware. An always available “virtualization executive” might be beneficial going forward, but my suggewstion specifically targets those with workstations that are currently productive using commercial software and who face end-of-life or other computing deployment projects.

    My career with computing began in the Sixties. Over those many years, I have worked with numerous situations where {platform A} runs applications designed and built for {platform not-A}. We accomplished this with software suites, sometimes with hardware assistance, running on {platform A}. My first ran 7090 code on 1410 mainframes. Later 1401/1410 code ran on System360. The best of these migration suites had the Digital Equipment Corporation(R) marketing name, “Application Migration Executive, or AME.”
    There VAX hardware ran any of numerous variants of PDP-11 software under VMS.

  • People don’t use operatin systems because they want an operating system, they use it because they want to run apps. Linux needs quality apps, which users *want* to use. If the apps are there, then people will endure even crapppy desktops like GNOME Shell or Unity.
    Then, the second thing needed in Linux is developers who want it to be better for users, not to booost their egos (and I am getting back to the GNOME 3 oddity).

  • George

    A common package format. We are now to a place where one has to choose RPM or DEB and commercial people will sometimes only pick Ubuntu (deb), this is going to be a problem in the future.

    Standardize this folks it’s is batshit crazy to support so many different formats.

  • Tom

    If we’re talking ordinary users here, might I suggest something that is a bit of a joke in the Windows world: automated help when something goes wrong. Specifically, when something goes wrong, the user boots a special CD/DVD/USB stick and is told to “follow the instructions”. Since the heart and soul of Linux is based on ordinary text-based configuration files (Please, Lord, save me from Windows fools who want to implement some sort of binary “Hive” on Linux!), it ought to be possible to fix just about anything that’s gone wrong. That all depends on how well you design the diagnostic tree, of course. Asking the right questions in a way the user can understand is three-quarters of the battle right there. And I’m not talking about stupid computer Q&A here. If the user is complaining about a hardware problem, the software should take a look and decide on what additional tests to run, based on a database created by experts. And when it’s all done, it ought to give the user clear advice on how to prevent the problem from happening again.

    Not everyone has a friend or relative to run to when things go sideways with the computer. Some of the fixes implemented by this AI will no doubt be obvious, but they’re only obvious to you or me, not the wives, girlfriends, and grandmothers out there to whom computers involve some sort of black magic satanic ritual. I’m not suggesting that this sort of program would be easy to write — far from it, in fact. Call it “Computer Autodoc” if you like.

  • ModernDayEskimo

    I’m not an expert in the Unix world, but am able to use Ubuntu well enough to install software, and if need be go to the forums for help, and terminal( have memorized a few commands, otherwise its all cut and paste), what I found challenging is choosing from the different desktop environments in the beginning of my migration in Ubuntu. I started in the Gnome period of Ubuntu then I watched it go to Unity which was a bit disorienting. The main problem I see with it and is totally my own opinion is devs cannot stay in one place for releasing each Distro of Ubuntu/Linux and having to re-run all the drivers and software packages all over again was a problem at the switch from Windows but I’ve been using Win7/8 and Ubuntu off and on since Ubuntu 11. The hardest part was learning how to get the OS to run anything that required restricted content a.k.a MP3 playback, DVD etc etc. Luckily there are forums and hundreds of Ubuntu/Linux savvy folks happy to help people and I say thanks to the people who posted threads to help me learn what little I know on Linux.

  • Daniel von Asmuth

    We need a distribution like Suse Enterprise Desktop that will get significant market share and make some money so they can spend it to get the quality right. Then we need to convince the application vendors to port their wares, which means we start paying for software.

  • I think that the focus on Linux’s problems is wrong, though I agree that the lipstick on the Linux “animal” doesn’t do much to improve things. (I won’t say “_ _ g”.)

    Here’s the order that the user questions should come in:
    1. What do you want to accomplish? (assume computer needed)
    2. What software will allow you to do exactly that?
    3. What computer OS has software that will totally do the job?

    So for example:
    1. I need to do business accounting that my accountant can use. My accountant uses a QuickBooks to look at client archives. It must sync fully with tax tables, credit cards and banks.
    2. Accounting software: QuickBooks (not GnuCash)
    3. I’ll need a Windows or a Mac computer workstation. (not just QuickBooks on a Linux server that Windows workstations can use.)

    I hope you get the picture. Linux business software is good, but it doesn’t cut it for an accountant with a Windows computer running a QuickBooks for CPA’s.

    Want a novel idea? Write Linux code for something like GnuCash that can make full QuickBook archives in Windows format that any accountant can use. That might just satisfy the real itch. Such an application probably would require an army of coders though.

  • […] Starks writing for Foss Force asks What Would You Do to Improve Linux?, and I’m glad he did, because I have a list. Oh my do I have a list. For me, the Linux […]

  • JL

    I would like the /home folder rationalized. Separate the OS related stuff from user related stuff. Just one mountable point for all user data.

  • […] Starks writing for Foss Force asks What Would You Do to Improve Linux?, and I’m glad he did, because I have a list. Oh my do I have a list. For me, the Linux […]


    I have already responded to Jim Lynch’s “How would you improve Linux?” of 28 March (IT World) take-off on your excellent article.

    The idiots who have seen the NEED to “improve” Linux over the last four years have made it virtually impossible to use today on older machines, one of the stated advantages of Linux.

    Want to use Linux on an older machine? The only rational choice you have is Klaus Knopper’s KNOPPIX. You can not even use MINT any more because of the need for memory extension (PAE) in your hardware.

    Perhaps the current state of affairs is not that bad, after all. Knoppix is an EXCELLENT choice as a desk OS.

    Again, I have spared you my feelings on this subject, because they’re readily available on the 28 March edition of Jim Lynch’s ITWorld offering “How would you improve Linux?”

    (I didn’t think the world could tolerate two versions)

  • terry

    - a working search in KDE.

    -better power management in laptops.

    -making the system tray icons (KDE) resizable for people with bad vision. every single thing on that desktop is adjustable EXCEPT for those icons that happen to be the most watched thing on the desktop. like many options on the DE, this doenst have to be used by everyone to be considered useful but having total configurability and then refusing to give this power to the most watched piece of screen real estate is annoying for those (elderly) who cant see.

  • Purple Library Guy

    Hmmm, waving magic wands, one thing.
    Well, my first instinct would be apparently dumb, but it would be: Have all games released on Linux co-equally with other platforms. It’s amazing how much uptake of Linux has been stalled by the lack of games; the number of people who said basically “Well, I’d try it, but I can’t play my games on it, right?” is huge. But, Valve seems to be taking care of that as we speak.

    Second might be: Have schools use it, thus saving a bundle of money. If all the kids were learning on Linux at school, mind share would follow and all the individual improvements that people wish would happen, could be tackled by the now massively larger Linux community.

  • […] Starks writing for Foss Force asks What Would You Do to Improve Linux?, and I’m glad he did, because I have a list. Oh my do I have a list. For me, the Linux […]

  • oiaohm

    Really things have moved on. tar.gz example in fact works with most applications. Docker takes this to a completely different level. Like if I am making a server class application release it as a docker package and you can have your own init system avoid all distrobution issues bar 1. What is the one remaining trouble issue. GPU acceleration.

    From my point of view complete off docker even if it involes getting Linus out to curse Nvidia and AMD to support cgroups from video card level with proper seperation between hardware particular code and what is inside a docker instance.

    Completing docker properly ends the distrobution problems. Why if a package is only released for X distrobution install it in a docker instance of that distrobution end of problem. Due to docker being cgroups the only overhead is ram usage.

    ditloml PAE limitation can be removed by custom kernel options. Not all distrobution have the PAE issue. Problem is secuirty. NX secuirty cannot be done without it. NX enters x86 CPU as of 2001. Yes PAE is even old. So ditloml are you serousally tell us that you are still using hardware older than 2001. PAE is in 1995 pent pro and the pent chips released after it. Just because windows requires a 64 bit processor to use PAE does not mean Linux does. To not have PAE we are almost talking 20 year old hardware. Most CPU that have been in use from 20 years ago are dead.

    So i386, i486 and i586 are what distrobutions are dropping by forcing PAE support. i686 and latter is supported due to containing PAE. In fact most of these chips would be out performed by a 50 dollar PI. Let alone the new clones. Yes you can software emulate on a PI to match a i486 100 mhz. The clones of the PI that have just been released can out perform the fastest i686 emulating x86. So why would you be messing with hardware that old with aged hardware issues it makes no sense.

  • Mike


    I agree that Docker may be a game changer. I haven’t used it yet, but I was under the impression the docker images themselves were distro specific. Am I wrong?

  • Purple Library Guy

    As to GPU acceleration issues, I suspect with Valve pushing hard, and with a good deal of success near as I can tell, for more gaming on Linux, those issues are going to improve rapidly.
    Consider–beyond the flood of games released for Linux lately, we also have most of the major game engines such as Unreal, CryEngine and Unity, all supporting Linux. This in turn means games developed in the future will be far more likely to support Linux, as making a game work on Linux will be relatively trivial–worth it for even a small amount of revenue.

    And we haven’t even seen yet what kind of games Valve is sure to have lined up for the launch of their Steam Machine. All in all, from a game library point of view, the Steam Machine is likely to have enough of a catalogue to enable considerable success. But none of that matters if games won’t run worth a darn on SteamOS (aka tweaked Debian). Making sure the graphics on Linux is good will be a high priority for Valve, and they have the influence–and money–to make things happen.

  • oiaohm

    Mike the internal contents of a docker image are normally distrobution tied. But docker images interfaces to the outside OS are limited how they are done so distrobution netural.

    1- The Linux kernel all the lxc stuff and device mapper stuff required to make a docker instance. So any Linux that supports docker will meet this requirement and systemd has releated requirements. The bare min lxc stuff is identical to systemd and the bar min device mapper stuff for docker are required for different forms of raid support and LVM.

    So building a modern distrobution without the feature dockers require is prity much impossilbe without screwing over users with your kernel.

    2- Hardware dependance for the likes of closed source userspace GPU drivers big problem how these are not fully multi user.

    3- is docker interfaces for logging and the like. These are highly stable and still not distrobution dependant.

    Why is docker like this its mean to be for server farm setups where you will be wanting to transfer workloads around. It is designed particularly to avoid Distrobution dependance in made up images. The little bit of kernel dependance and hardware dependance are still not avoidable yet. Yes hardware dependance include having to be built for the right cpu. But I don’t see most vendors would be upset having to release 1 image per cpu type. Lot smaller number.

    Docker itself is fairly determined to work. Like if you don’t have all cgroup features Docker would like to have it will still work just give warning messages. Older docker required a kernel with AUFS support this requirement been dropped in newer versions. AUFS using images can be migrated to the device mapper solution without rebuilding them. Reason the images never used AUFS directly in the first place.

    Think of docker being designed splitting roles cleanly between hosting OS and Contained OS. Since they are nicely split except for one or two nasty areas a lot of changes can happen without requiring the contained OS image to be rebuilt. In fact a day 1 docker image will work on latest docker installs and not even notice.

  • Mike


    Thanks for the info. I will have to check out docker.