Slashdot Mirror
ESR: 0.75 billion Linux users 5 years from now
Mike_Miller writes "In a recent interview with ESR in BusinessWeek, ESR calculates that, in 5 years, there will
be 750 million Linux users (a conservative estimate). Wouldn't that be great!?!
" I think one of my favorite parts is the title: "I want to live in a world where software doesn't stink." But this number should be fairly easy to hit-my cloning project has been doing quite well.
Personally, I prefer truth to wild speculations and fanciful predictions.
What is the truth? We live in a world of change. Linux could easily lose massive amounts to market share (not mindshare) to things like FreeBSD or even BeOS over the next few years.
As much as the truth makes a lot of you recoil, it still stands. For all we know, Linux could die out next year - unlikely, but there is still a chance.
You'll have to make it much better than MacOS X too before I switch.
The easy terrain has already been captured, or is well on its way. The hard stuff, where Linux is weak and MS strong, is still ahead of us.
Linux growth has centered around servers and networked computers, something which makes its multi-platform support an asset and its price very appealing. Its support for older and slower machines also is appealing in non-USA venues where prices for hardware and software are still higher.
But the home user doesn't care in general about ease of code porting or 100 megabit throughput. What they want is for their computer that they don't understand to work.
As long as Linux lacks support for winmodems, it is locked out of the retail channel (60-70% of retail computers have winmodems now). I expect MS to try and cripple Linux by extending this concept to other peripherals and to the PC99, PC00, PC01, etc. specifications.
So don't count those 750000000 users yet.
Linux is totally the opposite of a non hacker OS.
I'm using linux, I've used linux, it is difficult
and time consuming to use. It is not easy to
set up, it does not work with many sound
cards unless you buy the OSS for $30.00, after
you find out you can by poking around. It
does not work well with many newer graphics
cards, has no native open gl support. To get
the support requires tweaking with a 3dfx
I got it to work eventually, then could not
get quake II to work with it right.
Linux is a hobby OS, or a OS you run a server
on. It is not just a home user OS, unless
someone else sets it up for them and all
they want to do is use netscape and email
people. Linux needs installation programs
that set every single piece of hardware up
like 95 or 98 does, holding your hand all
the way. It needs a way to set a printer up fast,
like in 95. It is just not there yet.
I'm using Linux with Kde and Window Maker right now. In that
environment, it is easy to use. I can change most things about
my user interface setup and the apps on my desktop without
ever having to edit a config file.
On the other hand, I'm also a programmer. I can pull up a terminal
window and create scripts, work on programs, etc., in this same
environment. There is no inherent incompatibility.
You are right that there is not the best support for some hardware,
but most users don't have to buy the commercial OSS to get
decent sound for mp3's etc., with most sound cards these days
- you do for professional midi work, but then professional musicians
spend much more the $30 on other equipment. But, does
Windoze support non-PC hardware, Mac hardware, Amiga
hardware? Linux supports a lot more hardware than Doze.
Linux is VERY easy to install unless you also want Windows
on the same computer, and then it is moderately easy. Where
have been? I hope you are not still running Slackware version
1.0. Not to knock Slack, Slack's ZipSlack lets Windoze users
install a fairly complete distribution of Linux in a dos directory.
Just unzip and that's it. No partitioning. To add X, just download
the X files from Slackware ftp and install with Slack's simple
package utility and then add qt and kde the same way. Presto.
Compared to this desktop environemt in which I can play the
role of Joe User, Windows is just plain ugly and awkward.
And Linux is getting prettier every week.
Yes, India will hit 1G people soon, and China
:(
has reached that already. But don't forget, hundreds
of millions of people living there struggle every day to
feed their children and themselves.
Clean water for 750M people more than today in
five years is a much more interesting goal than
having 750M Linux users. Unfortunally, it's
probably even less likely.
--- Abigail
Actually he is dead on. I was making under U$3,000 per year when I bought my second PC ( An IBM AT in 1994. The Coleco Adam I had before was a gift ).
I am considered Middle class in Jamaica. I run my own business and I make well below what Entry level Tech workers take home in the US. In developing countries U$900 a year is poverty.
Best thing that could happen, would be people
starting to judge their tools by quality and suitability to the task, and ignore the rest.
Something being Linux doesn't automatically mean
it's the best thing on earth, for everybody, nor
there couldn't be something even better around
the corner.
There are good non-GPL apps, believe me. Even
commercial apps!
750 million in 5 years, that's about 1 out of every
10 people on earth. Considering that 1 out of 2 never
have made a phone call, and 2 out of 3 don't have access
to clean water, I'd say it's a bit optimistic.
--- Abigail
Oh, you'll see Linux on PDAs, but THAT's not where the growth will be.
A Pentium level machine is going to be embeddable for pocket change pretty soon, if it isn't already (StrongArm??). So let's say you can the equivalent of P166 with 24 megs a memory, a modest persistent memory and a network interface onto a $10 asembly in 2004. What could you do with that? You can make your VCRs smart enough so they can sport a really nice user interface; your home thermostat can get a lot smarter too. My pet project would be to slave all the applicance clocks in my house to a master using the power lines as a network.
Have you ever tried to figure out the temp controls on your refrigerator? How many photocopy machines are there out there that could be smarter and easier to use? If you could embed a reasonable computer for about $10 into an appliance, there's hundreds of places it would be justified on user interface issues alone, not to mention places like the automotive market where it's already done to acheive mechanical simplicity (God I hated working on my 76 Pinto -- literally dozens of tiny vacuum lines).
Think of the applications in sporting goods. Would you buy a pair of skis that adjusted their mechanical properties to snow conditions and speed, if they cost maybe 100 dollars more (mainly for exotic active materials, of course)? I could even see tennis rackets and golf clubs with computers in them. Maybe a golf club with an accelerometer and a network interface for downloading telemetry on your swing. The specific applications are farfetched, but applications _like_ these are inevitable.
Now there's things like routers that it makes no sense to run proprietary OSs on; for that matter if you're designing a printer why buy a proprietary postcript interpreterand some weird embedded OS when you can run Ghostscript and Linux with no licensing? How about alarm systems that can recognize your face or voice. Lab instruments like oscilloscopes which probably already use microprocessors to do things FFTs, but in some oddball proprietary language -- those could all be running Linux. Medical instrumentation like EEGs and robotic drug dispensers is another important nice where reliability and performance is crucial.
Home woodworking tools can get the capabilities of high end CNC milling machines, and the milling machines will probably run open sourced OSs too.
How about advertising displays? Are you going to sit a $50 license out where somebody can pick it up and walk off with it, or it can be struck by lightning? How many billboards are there in your town? How many cardboard foldout kiosks at the local book store?
Then there's the things that we don't ever think about but use all the time like avionics and telephone switching equipment.
If I sat down for an hour, I bet I can think of thousand places where linux would run, none of which are a "computer" per se.
Now, if you're paying $10 for the computer, how much do you want to pay for the OS per unit shipped? Nobody knows what MS gets for CE, but I'd be willing to bet it's north of $25, if not $50 or more.
Also, if you get into trouble in the middle of the project with an OS feature that's broken, are you going to want to wait for Microsoft to fix it?
So, I don't know about 750 million _users_, but I wouldn't be surprised if the average American or European has four or five copies of linux embedded in his home appliances and a like number in the office, and comes into contact with a huge number of anonymous Linux node through various networks he participates in.
Sooo -- Linux is a good thing to get in on the ground floor of -- kind of like becoming a LAN expert in the late 80s, only a lot more so.
I understand your point, but that doesn't really change things.
If somebody's video card is not supported, they are not going to run Linux. Whether it is the OS's fault or not is of little consequence - what matters is that it cannot support their video card, and thus they are unable to run Linux. Whose fault it is doesn't really matter - the fact that the hardware support is not there is what matters.
10 PRINT CHR$(205.5+RND(1)); : GOTO 10
Posted by DarkSecret:
;)
So this would be like 2 linux boxes per person? Kewl.
Posted by My_Favorite_Anonymous_Coward:
/. and I don't know a lot about it.) Can Linux handle Chinese/Japanese characters as good as freeBSD?) So there you go.
All I want to say is that a lot of third world country people don't pay for their OS software. So if Linux want to beat out winblows, it will solely depend on techical superiority. I heard that BSD 's very good at handling 2-byte chacaters (chinese, Japanese Character; I read this on
Posted by My_Favorite_Anonymous_Coward:
In other words there's no good way of estimating the number of Linux users out there?
Sure there's.... charge 90 dollars for every OEM copy for example.
CY
I like the assumption that Windows has some localization advantage over Linux- considering it's lead developer isn't a native english speak (he does speak english fluently, but it's _not_ his native language!). Not to mention a huge developer precesnce from Germany (SuSE, Star Division) and a growing Pacific contingent (Pacific Hightech).
The next thing to remember is that the threshold to spend on luxury goods (including computers) is relative to the local economy, not an absolute. How much does a home cost? A home in small town Iowa, for example, costs a heck of a lot less than a home in San Fransico, or New York. Most of the cost of an item is _labor_- in areas where labor is cheaper, cost of living is cheaper. Supply and demand.
What would they want a computer for? Perhaps they want to join the 21st Century with the rest of us, and plug into the international dialog. Or is the internet and computing only the purview of the very rich?
As for electricity, most of the major cities on the planet have it. No, the Kalahari bushmen don't, but there are a lot people who do. Note that my estimate is based off the assumption that only about 1 person in 4 is economically well off enough to even consider buying a computer, and about 50% of those who could, would. And that the price of the computer would rule out _by_ _necessity_ a Microsoft operating system.
The cold, hard facts are that Microsoft has killed this market because it cannot afford the margins Microsoft wants to maintain. You can't charge $100 for an OS an expect to sell it to someone for which that represents two weeks salary. Why do you think software piracy is so rife in China and other third world countries? Because for most of the customers, it's a choice of pirating the software, or not using it at all.
On the other hand, huge profits await the first company to do this. Consider- if each machine makes $10 profit (a mere 5%, a fairly thin margin) times 100 million machines is a billion dollars.
Consider: the "middle class" of this planet- probably numbering a billion people or more- are the people in developing third world countries earning about $3,000 USD/year. These people constitute a _huge_ market for computers and OSs- if the package costs $200 or less. That is about the equivelent of a $30K/yr person buying a $2K computer. Purchasing the standard $2,000 computer for these people would be like purchasing a $20K computer on a $30K salary.
The problem is that Microsoft is uninterested in this market- why? Because they want to charge $200 for the OS alone, sans computer. You need an OS that can sell for $10, or better yet $5 per copy, which runs well on a variariety of low-end hardware (ARM chips, low-end PPCs, 68Ks, 486's, whatever you can get on the super-cheap) and small memory foot prints, with internationalization capability, but with a decent user interface.
Linux fits this bill with flying colors. 80% of that 750 million people will probably be people who've never owned a computer before- and who aren't living in America or Europe.
Everything I've said here also applies to BSD, BTW.
You are certainly right, but as usual, it's pretty different with Linux. :) Just try to count someone who bought ten CD's and made five of his friends use Linux; also those who bought 3 different Red Hat distributions, and installed it on 50 PC's in a student lab, and made 50 more to netboot from one server; and someone who didn't even buy a single CD since a few years, only downloads updates/different packages, and installed a bunch of Linux webservers throughout the country.
"Ten years from now, they could do it in a few seconds." -- The Racketeer of the Hellfire Club, 1993, Phrack 42
Ok..... So he's forgetting the point of inflexion and the plateau that Linux must eventually reach.
Has anyone who's measuring the growth seen the point of inflexion yet?
Deleted
Unless something radically new comes along (always a possibility) I think Linux usage is going to go on increasing for some time yet.
100,000 million? huh???? Anyway, I'd say 100 million linux users in 5 years is pretty good
---
Kinda like WebTV, huh? I don't think they're selling 150 million WebTV units a year... This kind of thing has very limited appeal, due mostly to the limitations of the TV itself. I for one gave up using 640x480 a LONG time ago... Once HDTV becomes ubiquitous, maybe. But that's years away, too.
How so? You could have an SGI Onyx hooked up to a TV, and it would still suck, because you're limited by the resolution of the display. It doesn't matter if the device has the power to access the web, wordprocess, run Blender, and compile its own kernel in 10 seconds, all at the same time. *I* wouldn't want to do any of those things in fuzzy, not-even-VGA resolution, unless I had a Super-Size bottle of Excedrin handy. Until TV drops its 1950's roots and moves into the digital era, the ubiquitous "set-top appliance" computer will be very limited in its appeal.
There were four articles on Linux put out by BusinessWeek. This is one of the more positive of the bunch. Now if we could get Jerry Pournelle to retire (or at least stop visiting Redmond so often) we would get more people converted sooner.
See what I am talking about at
http://www.jerrypournelle.com/
Been there, Done that, Sold the t-shirt to the next idiot in line
Negroponte and his students have done an awful lot of wearables work using Linux, so I wouldn't be surprised if Linux ends up dominating the wearables market five years from now. After all, it's already got an early lead........
Finding God in a Dog
That would be $350 million people who are middle class by American standards.
These people have VCRs, telephones, cell phones, and yes PCs. They would love to do business on a network environment.
In addition, the Indian government has opened the ISP market up for competition.
I know I'm scrambling to think of ways of cashing in on this huge market. Yes India is poor, but with a middle class population equivalent to the entire US, that's a lot of growth. If only 12% Indians purchase computers that would be about 120,000,000 new computer users who would choose linux over windows.
I don't know stats about China, but I bet they are not much different.
Did anyone notice this?:
snip
---
What I really want is a machine that unifies my communications at a high level. That pushes us back to something that's more like a small portable or wearable PC. These appliances tend to grow functions and grow extensions over time, and eventually they end up being full-fledged computers even if they don't look like them on the outside. The appliances in the future are going to be like very small, very lightweight, and very carryable PCs that just happen to have a simple interface wrapped around them. And yes, I think that Linux will dominate them.
---
snip
.. and cheaper will bring computers into many more homes. All networked, and soon, all running Linux - it's much easier to create a network 'appliance' mainly for net access using Linux than using MS anything. Set up scripts to automate everything, GUI desktops to run internet access programs, and it does become a family PC.
Imagine a future Netwinder type thing, with TV output, pre-configured to connect to the internet by your local dealer, etc., and you have an ideal internet access appliance you can pop into a shop any buy.
Now just sell 750 million similar things in the next 5 years and there you have it.
India's population will hit the 1 billion mark soon. China's already has. In both countries, computer usage is growing tremendoushly, so don't discount ESR's figure entirely.
It's always hard to believe at the bottom of the hill that the snowball's gonna take down the whole mountain. In five years computers will cost as much as telephones do today, will have built in always on internet access and a roll up screen. Only illiterattes won't use them and only some people will still be using proprietary OS'. This is probably as hard to believe as believing in 1994 that most businesses would have web pages and that M$ network would be free of charge.
Here's a clue folks, exponential growth only happens in an unexploited market.
And here's two clues in exchange: (1) there are still huge unexploited markets for desktop machines. India, China, etc. Mexico is putting Linux in all the schools for the kids. And so on.
(2) Even in the US, the whole market becomes "unexploited" every couple of years or so as new apps and fancier, faster machines become available. New hardware means new OS sales. It's no longer a slam dunk that Windows will be the OS that gets installed on the new hardware -- especially with price margins dropping on the commodity machines to where WIndows represents 20% or more of the price.
-- Alastair
You know, I am sick to death of reading this. This very argument, or strains of it, keep popping up in message-places around the web (most prevalent on ZD*). "Linux isn't done yet, it won't support my foo card!"
:P
Why does everyone assume this is Linux's fault!??!?!? ARRRRRGH!!! Get a freakin' clue, people!! Does Microsoft write every single device driver for every piece of hardware for every version of Windows? NO! The hardware vendors do it! But they don't do it for Linux, because Linux doesn't have enough market share (or market presence) for them to feel it's worth their time to do so.
To date, the majority of the device drivers for Linux have been written by users who thought, "How the heck can I get this stupid thing to work!?!?" FYI, that is a lot harder than a company saying, "Here is the Microsoft driver for our foo card. We now have supplied drivers for 90% of the home computer users. That's Good Enough for us."
Did you ever stop to think at the sheer NUMBER of different cards their are for the Intel platform? It's mind numbing! Now YOU go try to keep up with every Soundblaster rip-off, every bleeding edge graphics card, without any support from the manufacturers (except 3dfx)! Daunting, ain't it?
For those of you prone to misread things, I am NOT arguing that people are stupid because the drivers DO exist. A lot of drivers are missing from the Linux Computing Experience. I'm saying people are stupid because they assume the OS is responsible for providing their drivers. Don't blame the OS, and do not blame the fine folks that have taken the trouble to write the drivers that you already have access to! These people have coded their asses off, and as a result non-coding clods like me (and a lot of other current Linux users) can actually use an OS that works. For them, I have nothing but thanks and heartfelt gratitude.
Blame the hardware vendors, Microsoft, the market, or your grandma's dog. Whatever. Just realize (now say it with me, kids) it's not the fault of the OS.
*whew*
There's that peeve vented.
--
Okay, I got Linux installed. So where's the free beer everyone keeps talking about??
How about a Poll with our own estimates for the next five years?
This is probably the same group that said the
Network Computer Market would be US$5 Billion
by 2001. This is the NC as defined by Sun
and Oracle a couple of years ago. I don't think
either company makes or sells these anymore.
This is just a hype wave that Linux is riding.
The down side of this, is that if Linux doesn't
grow to a 750 Million base when they predict,
it will be deemed a failure.
I read somewhere that US publishing houses have been shipping computer books to China as fast as they can print them so something is definitly happening in China.
I agree, partitioning can be daunting for a new user. I found RH 5.2 to have a workable compromise between easy (Workstation and Server options) and complete (custom option). Disk Druid needs to have better on-line help, as does the install process in general. The written manual is great, but people are used to hand holding.
That being said, a windows 98 install on a bare computer (unpartioned, unformatted hard drive) is much harder. If you try to run the install without a formatted partion in place you will receive a cryptic message about free space which stops halfway through and freezes the computer hard. You then must format and partition using command line tools fdisk and format which are worse than diskdruid in user friendlyness. All this with no online docs I could find from MS documenting the experience. The actual install is "pretty" and fairly understandable, an experience Linux installs could strive towards. But it does not allow power users to do what needs to be done, requires multiple reboots, and had drivers for less hardware (Matrox Millenium G200, Etherlink 3900) than my linux install.
Linux at 750 mil? That's quite an estimate. But I wonder if it's not a bit optimistic.
I have trouble believing that Microsoft will be so easily overthrown. Unless MS puts up
next to no resistance it seems obvious that they'll be able to maintain market share.
Unfortunately, I think coverage like CNN's comparison of Windows and
X could lead to problems. People don't necessarilly *want* a different experience.
Microsft still has many supporters, and many will not be willing to give up what they've gotten
used to over the last 10 years or so.
Someone has to give absolutely compelling reasons for someone to simply switch over
to Linux. Stability doesn't cut it. My Win box has been up for 15 days at a time.
Doing what ESR predicts necessarilly involves converting
incredible numbers of home users who won't really see all the benefits of using Linux.
ESR is right about one thing. Linux will grow. But not enough to surpass Windows.
That's why ESR has the job he has. Whenever you want to say something "sucks," you have to translate it to "stinks" for the press. :-)
100,000 million might be more reasonable
:)
You're right, 100,000 million is more reasonable. Better start having more children people, then develop a computer a 5 year old can use.
Maybe ESR hasn't noticed that lately Apple has been making major changes to the ROM architecture in their machines.
The Toolbox ROM is practically a thing of the past at this point. Mac OS 8 shipped with a full image of the ROM on the CD, beginning Apple's strategy of "ROM-in-RAM". Mac OS 8 (and later) will boot on some late-model IBM Power Series (800-series) machines.
Apple's latest version of this strategy is known as the "New World ROM", which is a ROM that only contains the essential boot information. The new G3's all read the Toolbox from the boot drive. Most of Apple's new PCI machines are fully Open Firmware compliant.
Should Apple open the Mac OS source, it will be useless without the Toolbox, since for all intents and purposes the Toolbox *is* the Mac OS.
Personally, I find it highly unlikely that Apple will open Mac OS. The primary indicator is Apple's Darwin strategy. Apple is not opening the source for the UI or other higher-order functions; these things represent Apple's technology advantage.
The point of this is that Apple has, for many years now, been steadily moving toward a more open hardware platform. The prevailing opinion in the Mac community is that we will soon see new efforts by Apple to support compatible hardware platforms under conditions that are far less detrimental to Apple as a company than their previous attempts at cloning.
Remember, Apple isn't a software company, nor are they truly a hardware company. What Apple *is* is a *design* company. The fact that Apple frequently has to invent new technologies to support its design vision is the proof of this statement.
Not that we wouldn't all like to see numbers like that, but come on. That's 13% of everybody on the planet, give or take. I don't think even Windows has managed to get that far.
Unless, of course, ESR is counting in this total:
Anyone who is running Linux on their desktop
Anyone who is connecting to a network with a Linux server
Anyone who is on the Internet (sooner or later they'll hit a Linux server)
Anyone who knows somebody connected to the Internet
20 million, I might believe. 30 million would be fantastic. But ~3 times the population of the United States? Sorry, but I think not.
caw caw
With the prerequisite critical mass levels attained [as per Sam Jaffe's comments], the snowball is only just starting to roll down that ol'hill. I wouldn't want to be Bill standing at the bottom. Clone or no.
BLAMMO shaken not stirred
I would prefer to also have one or all of the following choices:
- 42
- +INF
- -INF
- 3.14159265359
- 45i + 42
I remember, when I was a Warp user, talking to an IBM guy from Boca who said that their estimated 20 million OS/2 Warp users were counted by raw unit sales.
"So", says I, "what about people like me, who bought 2.0, 2.1, and 3.0 when they came out?"
"Well", says he, "you'd be counted as three people..."
Now, let's see, I've bought: Slackware 1.3, RedHat 5.0, 5.1, 5.2, and SuSe 5.2. Woo-hooo! I'm 5 of that 750 million!!!
--Quidquid latine dictum sit, altum sonatur.