How can Linux get 5 percent desktop market share?

Also in today’s open source roundup: 5 web browsers for Linux that you might not know, and 2 Linux Christmas carols

How can Linux get 5% desktop market share?

Many people have been predicting the “year of the Linux desktop” for quite a while now, but it’s never happened. A redditor recently asked what it will take for Linux to actually achieve 5% desktop market share, and he got some interesting answers in the Linux subreddit.

Daniel Fore started the thread with this post:

Jack Wallen thinks that elementary OS will be instrumental in reaching 5% desktop market share in 2017. We're pretty proud of our current track record with 2/3rds of downloads coming from non-Linux systems.

But I'd like to ask all the lurkers and dual booters and part-timers in this sub: what should we focus on next year to get you to switch to an Open Source OS?

What are the barriers that keep you from using Linux full time?

More at Reddit

His fellow Linux redditors responded with their thoughts:

Jampola: “Everything needs to be working A+, it needs and polished. Put yourself in the shoes of your average users who want to perform mundane tasks such as scanning, printing, web browsing, mounting USB sticks, burning CD's, zipping, unzipping files, viewing organizing photos etc. You have 100% take yourself out of the equation and make sure that these things just work. If you want to get to the 5%, you have to understand that probably 4.8% of said users couldn't care less about the OS, they just want to work/play.

I was tasked with developing a spin based on Debian to migrate from Windows for 100's of users, and literally, nearly every simple task would have some minor issue that the user would come up with. Things like language switching not always working, inconsistent font rendering, certain PDF's not being displayed properly. These are the things that piss people off who otherwise don't care about the OS, they just want to get their shit done.

Secondly, as everybody has mentioned, work on finding a partner who's willing to sell appliances with Elementary pre-installed, then test the hell out of it until it breaks.

Good luck! :)”

NotThatTree: “For me, every single time I've installed Linux for someone, they seem to like it fine, but then a few days/weeks later they desperately need to run some Windows app or another for school or work and there's no way to get it running easily on Linux. So I gave up. The exception is my girlfriend who only ever uses a browser, but even then I left her with Windows for a year to make sure she wasn't lying.”

Ken0201: “Think is really what happens to most. They use Linux a few days, think it's awesome, blow away windows... Then they run into some problem they can't overcome... And reinstall windows. I always recommend dual booting for at least 6mo before making a decision to nuke windows. I still have windows, even though I've not booted it since initially setting up this laptop about 6mo ago…and I've used Linux going on 10yrs. I'd nuke it, but I keep it around in case someone needs help with windows, I can boot it and walk them through it.”

Octalus: “Getting laptops shipped with Linux would probably have the largest impact. A lot of casual computer users just use whatever's shipped and either are unaware of the existence of Linux, or don't care enough to actually try Linux.

Expanding the market share this way would also have the effect of more companies porting products (due to larger market share) and eventually Linux gaining the more enthusiast windows market due to more software support.”

Gondur: “We had that with the netbook bubble years ago. People hated the linux experience and returned then it masses. Netbooks were accepted when winxp became available.

Preinstallation, while helpful , is not the rootcause of linux suffering ... but the system itself with several architectural weaknesses which mskes it unsuitable for the end user PC market.”

Ixxxt: “I use it full time. But what we need is more users. Perhaps more outreach events or showing things to manufacturers proving peoples distrust and dislike of Windows 10 and trying to get more laptops made with Linux by default, getting more people downloading won’t be enough to make big percentage jumps. ”

DELETED: “Hardware support is the major one I have found. When people buy a printer/scanner/phone etc. They expect it to just work when you connect it.

Until there is near ubiquity, I do not see there being a major uptake in the home.”

Ragingbwner: “What stops it is simply put the fallacy you make in your own post. You seem to treat "Linux" as a single platform with your market share statistic. The only reason people do that is because someone once decided that you should name your OS after the kernel which really makes very little sense, if anything, naming it after the libc makes the most sense because that's where the real compatibility lies but even that doesn't make much sense.

The only reason you think all these widely divergent operating systems deserve to be grouped together just because they use the same kernel is because someone once decided to give them all the same name. If someone once decided to name operating systems after their bootloader and we called them GRUB, sys, lilo and efistub systems then people wouldn't so easily group them together any more and act like they're a platform.

What stops "Linux" from gaining a bigger market share is that by grouping a bunch of unrelated systems together people already inflate the size of the market share to begin with. Different systems that use Linux as a kernel often have about as little to do with each other as FreeBSD and Solaris do. Basically, they're not a platform and software doesn't run across them. And that's what stops it from gaining a bigger market share.

The actual market share as relevant for companies is a lot smaller. People sort of sell "Linux" as a platform to companies, then they do research into porting and they are like "wait a second, 'Linux' isn't a platform at all, if I publish binaries for Ubuntu 14 then there's not even a guarantee it'll run on Ubuntu 16, let Alone Debian Jessie and certainly not Void and Alpine, no thanks...' so they won't bother.

Steam was the only one that tried and it's actually really sht because of that. Valve had to basically write half an operating system themselves in the Steam runtime, these are things an operating system nowadays is expected to provide, which Windows provides, which all these systems that use Linux do provide, but just in a different binary way, and they had to re-invent it and bundle it themselves, which is expensive, smaller companies cannot affort this. Also note that Steam really does not run everywhere, if your system does not use the glibc libc then you can forget about it (see above) as well as a lot of other stuff. Other systems than Ubuntu often have to hack pretty hard to get Steam to run with varying degrees of success.

How did we always solve this? Free software, software whose source is public does not just typically run across "Linux" but across every Unix, even though a binary never will, because once the source is public typically a quick patch can be applied to make it work and often it's just a case of recompiling and the compiler will figure it out on it's own.

Free Software is what allowed Unix to diverge so aggressively on a binary level, software just continued to run across all Unixen because the source code was there, if it didn't run a simple patch made it run.

Now a lot of "Linux users" want to invite users of mostly proprietary software to switch and they won't switch until their proprietary software comes with them, and it can't work, you cannot write proprietary software that works "On Linux", you have to provide the source to do that.

And no, Flatpak and Snappy don't solve this, it turns out they just displace it, they themselves again don't work on many different Linux systems because the system is different there and does some of the things they need differently.

The only way for "Linux" to grow is for one of the Linux-based systems to just win so proprietary developers can target that one and that one alone. As it stands Ubuntu is becoming that, a lot of proprietary developers officially only support Ubuntu and let others figure it out for themselves. That doesn't make Linux work though, just Ubuntu. And even so, as said, Steam on Ubuntu works in a way more inefficient way than on Windows where it can offload a lot of things to the OS and even the kernel they have to do themselves now. Ubuntu is very happy to break their binary interfaces between releases. Windows never breaks them, it can't break them, too much proprietary software depends on them that would stop working.”

BashScriptsForYou: “OEM perhaps. If you can have a laptop, or some kind of computer you can sell with it, that has great support, eOS becomes a type of computer instead of fringe OS. (Like how chromeBooks are a computer, instead of crappy hardware+ weird os, to the end user)

Maybe less intermediate layers? I remember getting wifi to work was painful because it needed a newer kernel. And you're based on ubuntu which is based on debian. Seems you could just be based on plain debian”

Letoiv: “Hardware partners. The day that a tier 1 PC maker agrees to ship all of their consumer laptops with Linux pre-installed as the default OS is the day that the year of the Linux desktop has arrived. It'll never arrive without that. Someone who ships a lot of units needs to agree to make Linux the default choice.

What conversations have you guys had with PC makers about what would be required to make that happen?

The community will tell you hardware support but it's really 1-2 big hardware partners you need to hit the next level of growth. The market of people who are willing and able to install a non-Windows OS is small, and Linux already dominates it. What you need is to get one PC manufacturer on board, do an NDA and get tech specs for every peripheral they bundle, and do drivers for all of it. Then they get the machines in Best Buy and you are the biggest Linux distro overnight.

The question isn't what do we want (though we appreciate you asking). The question is what do the PC manufacturers need to make Linux appealing to them? They make money by shipping Windows because they can preload a Windows PC with apps from ISVs who pay them. But they also have to pay for a Windows license. So it is a bizdev question. You can cut costs for the PC maker by being free, but you also need to figure out how you can make them money so that they come out ahead on a Linux PC sale vs a Windows sale.

One easy place to start is to look at what comes preinstalled on Windows PCs. Some of it will be junk you don't want to deal with, but some of it maybe not. For instance if Dropbox is paying PC makers to come preinstalled with a free trial deal, you go to Dropbox and you say hey what would it take for you to do this on the Linux side too? Then you go to the PC maker and say now Dropbox is on board. You have to do this kind of footwork for the PC makers, unlikely they will put it all together themselves.

There was a time when they were all locked into contracts with Microsoft which imposed penalties if they shipped a non-Windows OS. I think that's all gone now. Which is good because that would be an expensive legal battle to overcome.

Bottom line: show the big PC makers how you can make more money for them than Windows can. Put together the partnerships for them and show them what it would look like. Figure out what it would take, start pitching and keep pitching. It only takes one.”

More at Reddit

5 browsers for Linux that you might not know

Linux is blessed with lots of options in terms of web browsers, and there are some that you might never have used on your computer. A writer at Make Tech Easier has a roundup of 5 web browsers that you might want to consider for your Linux box.

Ada Ivanova reports for Make Tech Easier:

While you might be fine sticking with the default browser for your Linux distribution, very often one browser isn’t enough. The default browser might be crashing all the time, or you simply might need more. For such cases you’d better know what the alternatives are.

There are probably a dozen or more browsers for Linux. As you can expect, not all of them are equal. Some of them can’t open properly even a moderately complex site, but they are still useful for simpler sites. In addition to Opera, ELinks and Midori that you already know about, here are five more Linux browsers that deserve attention.






More at Make Tech Easier

Linux Christmas carols

With Christmas happening in just a few days, a writer at Computerworld has a couple of Linux Christmas carols ready to help you celebrate the holiday.

Sandra Henry-Stocker reports for Computerworld:

Get ready to start caroling around the office with these Linux-centric lyrics to popular Christmas carols.

My (first) Twelve Days with Linux

To the tune of: The Twelve Days of Christmas

On my first day with Linux, my admin gave to me a password and a login ID

On my second day with Linux my admin gave to me two new commands and a password and a login ID

On my third day with Linux my admin gave to me three man pages, two new commands, and a password and a login ID

On my fourth day with Linux my admin gave to me four privileges, three man pages, two new commands and a password and a login ID

On my fifth day with Linux my admin gave to me five cron jobs, four privileges, three man pages, two new commands and a password and a login ID

More at Computerworld

Did you miss a roundup? Check the Eye On Open home page to get caught up with the latest news about open source and Linux.


Copyright © 2016 IDG Communications, Inc.

InfoWorld Technology of the Year Awards 2023. Now open for entries!