Slashdot Mirror
Linux Five Years Away From Mainstream
wellington wrote to mention a ZDNet blurb about a Gartner group study. Gartner indicates that 'mainstream' use of open source in IT environments may be 5 years away. From the article: "Gartner's latest Linux 'hype cycle' report shows that open source is halfway to maturity but warns the biggest test will be whether it can demonstrate the necessary performance and security to function as a data centre server for mission-critical applications. Leading-edge businesses are generally still in the early stages of Linux deployments but Gartner expects increased commercialisation and improved storage and systems management for the operating system by the end of 2005, with Linux being used primarily for WebSphere and infrastructure applications on mainframes and web services on blades and racks."
Did you know that nuclear fusion is only 20 years away? Just like it was in 1950! (No, I'm not skeptical. Not at all.)
When I wrote my article and its follow-up on directions I think a Linux Distribution could take, I expected that there would be some controversy. However, I hardly expected the shear number of responses to the effect of, "Linux is great as it is! Never change it!"
Which is surprising, because the very point of the Linux design is that different distributions were supposed to be able to explore completely different tracks. There shouldn't be any "one distro to rule them all", yet many of the respondants demanded exactly that! (Amusingly, they couldn't agree on *which* distro to rule them all.)
When I pointed this out to many responders, and mentioned the fact that I'm merely attempting to suggest a Desktop environment that would help Linux adoption, I got another surprising response: "Who said we wanted regular users? Linux is for the elite. If you're too stupid to recompile your kernel or read all the scattered HOWTOs, you're too stupid to use Linux!"
I understand that the Linux community is wide and varied, but this sort of attitude is not helping anyone. In fact, this sort of attitude causes Linux to take two steps back for every one step forward it takes in the market.
It's normal that Linux users will disagree. That's why Linux is just a kernel, KDE/GNOME are just desktop environments, and the GNU System is just a collection of Unix utilities. It's so the end distributions can build the OS necessary to meet their users. But such a design DOES NOT require that users berate each other! Rather, Linux users should understand that "idiot" users using an "idiot" distribution is okay. Gentoo users can still recompile Gentoo to their hearts content even though Ubuntu exists. Ubuntu users can still use Ubuntu workstations even though Fedora exists. Fedora users can still a have 100% "Free as in socks and gun ownership" OS even though SuSE exists.
There's no reason for this OS bigotry. It's causing confusion in the marketplace, and generally turning the public off to Linux. Just pick the distro you like, and be happy for other people who use something that works for them. K?
Javascript + Nintendo DSi = DSiCade
Linux was mainstream five years ago.
for the drm wars
The best test environment is production. - Me
chrome://browser/content/browser.xul
How come every thing is "5 years away" but never seems to get here. I'll bet the writers for the Jetsons anticipated space cars in 5 years too.
The source of bullshit for years and still counting.
Five years to mainstream Linux -- I'd say they were being optimistic about desktops. But servers? When is this report from, 1997?
What I'm listening to now on Pandora...
Apparently, 'Running the majority of web servers worldwide' doesn't count as mainstream.
This means that most of the software the current /.'r is running, won't show up in enterprise level distributions for several years. So yeah, five years off doesn't sound that far off the mark.
Since when did operating systems become a religion?
Gartner indicates that 'mainstream' use of open source in IT environments may be 5 years away.
I wonder where he has been. I started using Linux IN the datacenter some 4-5 years ago now. One system was up for almost 4 years running DNS and Squid. For DNS, we occassionally patched it, for squid we had a job that restarted it once per week at 11pm on Sundays. It didn't make it to 4 years because the UPS had to be upgraded. We had bets if she would reboot, the na-sayer lost.
And it was a no-name left over PC to begin with. The moral of the story is that it isn't the hardware as would the other OS have you believe.
Not adopting Linux where it is suitable has more to do with an inability to change fo the better and what shares does the CEO/CFO/CIO own.
Why does Linux keep getting faulted for installation issues while Windows gets a pass?
Linux installation is not a reason to avoid switching at a corporate or oem level.
I downloaded and installed Suse 9.3 64 bit on my new dual Opteron the night before last. The installation went really smooth but of course there was a hiccup. I had to install sensors. That involved a trip to a web site, yasting around a bit, etc.
It would be easy to blast Linux for not automatically doing everything and retreat to M$ land, except that Windows 64 bit doesn't even have drivers out of the box for my SATA hard drive and thus wouldn't work at all. If I really wanted fans to work badly enough, and could not get a device, I could write a kernel module myself and all Linux hardware stuff has excellent documentation to at least get me started.
The bulk of all OS distributions are either OEMs or corporate rollouts. OEMS have a team that prepare images for a fixed hardware, and so do corporate rollout centers. Whether you wade through driver compatibility issues on Windows or Linux doesn't matter. Both systems have similar problems and Windows wizards at that level don't really help someone who should already be an expert on the topic.
I would think that OEMs might consider locking down Linux PCs so that end users do not have the root password. So they can't break it...
This is my sig.
I think it's really important to distinguish from Linux the server platform and Linux the desktop platform, as you say. I run GNOME from an Ubuntu distro on the desktop, and it's.... pretty good. But it's not XP. No Quicktime or WMV plugin means a lot of websites like CNN and Yahoo don't really work well. Xine is ok for DVD content, but overall it's a bit slow and uses more resident memory than what I consider an equivalent XP system does.
Linux as a server has arrived, and has been here for awhile.
Just like Fusion is always 10-20 years in the future for commercial usage.
It depends on what your definition of "mainstream" is, of course. Right now, more people are using Linux than ever used Microsoft's DOS. Or Windows 3.1 for that matter.
Define your own reality - don't let others define it for you, with metrics based on the sales price of the OS, or the net revenue from OS sales. Linux strength is it's low cost, so it will never win at that game.
But right now, many people worldwide use Linux, or even BSD, even if it's what runs on their cell phone or inside their networked self-repairing robot-dog-feeding fridge.
And, to paraphrase Martha Stewart, that's a good thing.
-- Tigger warning: This post may contain tiggers! --
Kind of like this article from 5 years ago, or this one from 3 years ago, or this one from Dec. 2000.
In 5 years, there will be an in every garage. Yawn...
Until you can get the easy things doable by the masses, then you have a chance at taking the desktop.
Storm
Gartner is a respected company. Many companies pay TONS of money to get their two cents. Considering this, for those of you who do not like what Gartner is saying about Linux...how about you counter their findings with your findings. Here are the rules:
You should be neutral (tough for this crowd)
You should be logical (shouldn't be tough for us, but will probably be)
You should perform qualified research with backup sources.
Publish
Profit
Saying "Gartner you suck, you don't know what your talking about. You are five years behind the times" is really lame and inflammetory (if not trollish). Proving them wrong goes a LONG way.
I mod down so you can mod up. Your welcome.
Neither you nor the Gparent post understood one of th e main points of the great-GP post.
To answer GP:
Not in the software thats available, but in sheer choice of software.
MS Has Windows XP home and Windows XP Professional, designed for the general required use, its easy to tell epopel to get the correct version.
Sure, and it is really easy to tell the same people to buy Xandros or Lindows or even Mandriva. Just tell them to use Mandriva! Do not tell them to use just "Linux" because then you will give them problems.
Its like if someone bought a computer and you tell them "you can install Windows 98, Windows 2000, Windows NT 4, Windows XP, Windows 2004 or MS-DOS if you like the command line"
Of course you will know that person and if he is not computer saavy, you will tell him just to install Windows XP Home or pro.
Make the same with Linux, as GP said, Open Source software is all about Choice, Linux IS OSS, so it is all about choice.
The problem here is that we (Linux advocates) continue to try to push this motto "Linux is Easy" or "Linux is for geeks only" or whichever but the key element is that Linux (nowadays) is all that.
"Linux is Easy" for your Grandma if she uses Lindows.
"Linux is easy" for John Sysadmin if he uses Slackware (just an example okey? do not bash me)
"Linux is for geeks only" if we are talking about Gentoo
"Linux is difficult and not functional" if talking about Lindows for John Sysadmin.
Do you understand? I think it is time to stop thinking about "linux" as the operating system -per se- and start to think about distributions there ARE distributions for every kind of person and whichever the person, if you recommend the wrong distribution it will be -difficult-
Ubuntu is an African word meaning 'I can't configure Debian'
The article keeps flicking confusingly from Linux to Open Source.
Open Source is already mainstream. I don't have colleagues at any major enterprises that don't use it, and the smaller enterprises tend to use it for a larger percentage of their operations.
Linux adoption is however far slower, and I don't know anybody at all using it (commercially) on the desktop. I'd be surprised if Gartner's 5 years is correct, especially given the way Sun's Java Desktop hasn't exactly been the most successful business venture ever seen.
So does anybody have access to the Gartner report that can clarify whether it's Linux that's 5 years away, or Open Source software?
(Not that I rate Gartner especially highly anyway)
I'm curious. Which distros did you try? All of the major distros, upon which others are based, have some form of package system. Gentoo has emerge, which, if I remember correctly, is very similar to your beloved Ports sytem. Redhat/Fedora Core has had RPM, and now has yum. Debian has dpkg and apt. Slackware also has packages of some sort, although I have no idea in what form.
It sounds to me like you didn't put a whole lot of effort into trying to use Linux and expected it to behave the same as FreeBSD.
The average home user (I'll say a 60 year old whippersnapper is getting there ) is becomming more accustomed to being able to do more advanced things on the computer with just the click of a button. Backing up a dvd is a perfect example. Or even better, transferring a VHS to DVD. On windows, they buy a piece of crap software that barely works, but does what it says it will do, and thats copy VHS to DVD. Stop and think how many steps it takes to do that in Linux. Every one is saying its mainstream now. IT IS NOT! It is industrial right now. Mainstream is when it is a house hold name. I actually keep a windows box around for several reasons (mostly multimedia in nature)just because I can get what I want to do done in a quarter of the time with that particular os. I use linux on all of my other computers because it does all that I need for those particular machines. When they make a distro that is so dummed down that my 60 year old father can pick it up and go "Why didn't you show me this earlier??" it will be ready for mainstream. There is too much configuration to do for now (/etc/hdparm.conf, ipchains, samba, etc.) that isn't done automagically or through a very easy to use UI (Suse is an exception, but it is so crippled in other respects that it can't do what else I need it to do with out doing the dependancy dance). Personally, I think Ubuntu is a good start, but it isnt perfect either. All of the distros have something to them, but none are ready for my old man...yet. Till then, It will always be 5 years away. After then, it will be the now.
Stop signs are only Suggestions
The study is more about the market state than the technical readiness of either Windows or Linux, don't confuse the two. Technically Linux may well be ready and in many ways is better than Windows as a server, but this doesn't automatically translate to higher adoption in the market, as Windows has massive unfair advantages e.g. huge marketing budget & sales team, 'network effects', critical mass, desktop monopoly etc. When they say Linux is "ready" to start becoming "mainstream" they don't mean it is now technically 'good enough', they mean, the market is at a point where it is willing to adopt Linux enough that Linux is poised to reach the required critical mass. In other words, in five years or so a significant percentage of companies will be 'ready' to adopt Linux as a server platform. It's not so much a case of "is Linux ready for the market", it's more about "is the market ready for Linux".
I tried Linux *FIRST*
I tried SuSe, Mandrake and one other I think. Problem was, every time I read a HowTo or other document, right after discussing how to do something, it always added "your distribution may do this differently." And surprise, surprise, they did. So I tried FreeBSD. I got the book "Absolute BSD" and went through that. And every example worked as published. Yes, I understand that this is due largely to the "FreeBSD is an operating system, Linux is a kernel" thing, but the problem I ran into was a lack of distribution-specific information for any given linux distribution.
FreeBSD was simply easier to understand for me. I do recognize that is a highly personal choice, though.
Merde, il pleut encore!
here, we're not moving to XP from 2000 as it's not worth it: we're moving to longhorn as and when it emerges
So, you don't see the point in moving to XP, but you do see the point in going to Longhorn... What does Longhorn offer that's significantly different from XP anyway? Transparent windows?
I suggest you take 15 minutes out of your life (vs > 45 min for Win X) & do an Ubuntu installation. The only potential user hurdles are because it has to be installed- it it came on the machine there wouldn't be much difference between it & Windows. I say this because I've installed it for Joe non-technical user and they're fine with it.
This is the case now- not some nebulous prediction.
This is before you even get to Linux distributions. And still before you get to the distribution, you've the C library to contend with. There are several for Linux, and this will determine what you can run and how it will perform. So, choosing a different C library essentially changes the whole picture.
Oh, and on top of that, you need to decide how generic/specific the kernel and C library are to be, with regards to hardware. Do you want something that'll run on any white box with the processor of your choice? Or do you want something that runs specifically on your machine, as it stands right there and then?
Then, there's your choice of init system. There are quite a number listed on Freshmeat and there are probably as many more again that aren't listed. You also don't need a classic init system, if you've something specific in mind and it won't change - you can write a program that starts up whatever you like.
Assuming a fairly standard init, to have a working Linux system, you need to be able to connect to it somehow. There's about a dozen getty-type packages out there, with different strengths and weaknesses. There are several login programs. For graphical logins, the number od xdm-like packages is unbelievable. I've lost track of the number of username/password systems, which may or may not use PAM or something similar.
So far, the number of combinations is astronomical. We've not got as far as a working system, all we've got is a skeleton that'll allow connections and trigger things.
My personal preference would be to have a meta-distribution that is compiled on a central system, where you pick the options from a pick-list and it builds the distribution from your choices.
It's a small world and it smells funny; I'd buy another if it wasn't for the money; Take back what I paid (SoM)
yum update
Downloads and updates everything. No user intervention required. I love it!