Slashdot Mirror
What Lies Ahead For Linux
An anonymous reader writes "Here's an interview with Stacey Quandt, a Linux and open source industry analyst. She explains why she feels Linux will overtake Windows as the number one operating system within the next three years." There's some interesting tidbits on what it takes to be an industry analyst as well, and some looking back to when most analysts were unaware of Linux.
Nice out-of-context hyperbole. She was referring to shipments of new boxes in the server market. In terms of desktop market share, she says that mere parity would take "a long time", and she's looking forward to a modest 10% share (essentially changing from a "fringe" player to a commonly-supported niche player) as a significant milestone.
http://alternatives.rzero.com/
Did they include some pictures... it, uh... helps me with my advocacy. Oh, they did. h0t!
He wants his headline back.
I thought the Slashdot editors would have caught this mistake when they read the article...
;-)
Oh. Nevermind.
I would doubt anyone would agree with the statement that Linux could overtake windows in 3 years, it will take a lot longer and more team work from the linux people to make this happen, not to mention Linux better start getting the support of gamers who can drive the sales of OS purchases.
**It runs through my veins like radioactive rubber pants! Do not deny my veins!**
Analysts exist solely to pimp products for vendors. When an XYZ Analyst tells you that XYZ is going to take over the world in 3 years, you can safely ignore it. That holds true whether XYZ==Push Technology or XYZ==Linux.
I like that name. That's a very clear title for Microsoft. It definitely would get the attention of someone undecided about MS vs Linux.
"Well, you could buy OS and related products from a convicted monopolist, or you could get these open source products (and buy professional support) from these (_list_) vendors."
.sigs are for post^Hers.
How many years before the server/desktop OS becomes irrelevant? The apps make the platform valuable, not the OS.
Yeah, right.
Maybe I'm a minority, but I use MS Windows XP at home to listen to music, watch DVDs, play games, surf the web, and (very infrequently) be productive. I download every security patch and always scan downloads. And my system has always been rock solid stable, and I've never had any kind of security breach (I use Norton Internet Security, too). Yes I payed a couple hundred for all the software (more for each game) but it's been worth it to me. And P.S. I love Linux - I develop Linux apps for a living.
Mine is Good
Anyone who has read more than 2-3 reports from the "big boys" like Gartner can easily answer that one. Not much, save zero morals/integrity.
I worked for a company which dealt exclusively with whitepapers written by the big analyst houses. The reports were widely known to be staggeringly poor, often blatantly wrong. It was hardly surprising that they were a royal pain in the ass to deal with on a technical level; getting them to use FTP to upload their content was nearly impossible. IT industry experts who can't figure out FTP. Special.
I've seen numerous comments here on /., on stories about both pro and anti linux analyst reports, talking about how much of a joke these companies are. Most of the analyst groups do huge amounts of "commissioned analysis", which is then passed off as being legitimate, unbiased analysis- when it is nothing of the sort.
Analyst groups have turned into little more than for-hire technical marketing (the computer industry's version of "military intelligence") who spew out documents just technical enough to impress/confuse the top brass.
Please help metamoderate.
I agree with you that stability is a minimum, but it depends on if you are discussing servers or desktops. In being a stable, a server has accomplished a big part of it's requirements. The desktop is a completely different ballgame. Linux on the desktop is a very popular /. topic, and everyone always wonders: "will this be our year." It's been a long time coming, but I still think we have some distance to cover. Linux is very powerful and very stable, and pretty user friendly these days. It is not yet ready in terms of program installation, especially when we are talking about Joe Sixpack. From my experiences with friends and family, the "average" user has little or no knowledge of the command prompt and no desire to learn to use it. It has to be so easy that the user can click on a program and have it installed. There are solutions that are getting close, such as RPM and APT, but there is still some ground to cover.
This isn't necessarily the biggest or most important problem with Linux on the desktop... I have seen many articulate and thoughtful discussions on this subject before. It is my honest belief that Open Source will have it's day, as many users are already switching to Firefox and Thunderbird among other things. Linux's day will come, but it will still take some time and honest, constructive criticism.
Damn... right when I got my karma up....
Oh no.
Linux has been "overtaking windows in the next 3 years" for at least the last 6 years,
probably more. Don't get me wrong. I'm all for running Linux on some non-critical servers.
But on desktop it's a freaking joke. I mean, you still can't play multiple sounds in Linux
at the same time (unless you use a laggy userland daemon that takes a second to unpause your mp3,
or buy a "supported" audio card with hardware mixing). For some reason Linus has a problem with
putting kernel audio mixer (something Windows had since Win98 (and probably '95 but I'm not too sure))
into kernel-mode, so now all "desktop linux" users are stuck with playing one audio stream at a time.
Even FreeBSD, a much less "desktop" oriented OS (at least it isn't claiming to be "the windows killer"
on the desktop every few months), has kernel audio mixing support since like 5.x-CURRENT. So this
was one tiny nitpick about audio, something people on "desktop" will probably need sooner or later.
How about video? Windows supports almost every known video card out of the box, while to get any kind
of decent graphics in Linux you need to buy a "supported" video card. How many "corporate desktops"
you know of that run on exotic "custom ordered" hardware? They all use precanned HP/Dell/Whatever
desktops with generic onboard video and audio. Unless Linux will automatically without *any* problems
installs on this class of hardware, forget using it for corporate desktops.
And to sum this up, I guess the real reason Linux isn't going to be overtaking anything "in the next
3 years", is the group mentality of Linux users in general. There are literally hundreds of half-assed
"distributions" of Linux. And new ones seem to be popping up at an amazing speed. Compare that to
the *BSD family, where there is only one "distribution" for each flavor (FreeBSD, NetBSD, OpenBSD) and
once you know one, you should have no problems migrating to any other *BSD family. In Linux, every
distribution seems to want to invent their own packaging system, configuration system, etc etc.
People, this is not how you win users. You win users by creating a standard, easy to use system.
Forget the 100's of distributions. Create a single standard and make everyone use it. Then, only then
you might have some chance at a "desktop OS".
Servers, on the other hand, should be linux's play ground
I used to think that, but after doing some work with Win2003, I'm not so sure.
Imagine how much harder physics would be if electrons had feelings! -Feynman, maybe
Microsoft is an "unstoppable" mega-corporation. Any legitimate competition is crushed by the might of Microsoft. Try to develop a for-profit operating system to compete with Windows and you'll get crushed. Try and develop a for-profit word-processor to compete with Word and you'll get crushed. Microsoft has reached the top of the food chain.
Legitimate for-profit companies cannot compete against Microsoft. Due to this fact, "free" software, such as Linux and Open-Office, has bubbled to the surface as the only possible contender in the evolutionary struggle against Microsoft. Providing "free" software is the only way to possibly compete against Microsoft. There would not have been a need for "free" software if Microsoft had not crushed all possible means of fair competition.
This lack of competition also hurts Microsoft because: a competitor, in general, only needs to be better than his next closest rival. If there are no close competitors then Microsoft does not need to improve. If it does not improve it will stagnate, whither, and die. It will be overrun by the weeds of small "free" software projects just waiting to get out from underneath the shadow of the mighty giant Microsoft.
My majors were chemistry and Asian Studies in college. Am I working in a chemical factory in Asia now? No. Am I a geek reading Slashdot at work and replying to you? Hell yeah!
If you cannot go beyond judging a book by its covers, you should not be judging.
Probably blow my good karma with this, but oh well
:)
I agree with everyone that Linux has become more usable and more security oriented(depending on the admin), but the bottom line is that as far as corporations and windows in the workplace goes, I doubt linux will grab a significant user base because of some basic reasons:
1. Alot of corporations will cling to windows because 99.9% of their userbase is on windows right now. They realize that there is cheaper alternatives out there (linux) but they rather stay with what they are using because it will cause less headaches for the IT dept. and operations as a whole will run smoother without messing with the OS that they are using.
2. Users in the workplace are comfortable with windows because it is what they know. Applications are not quite as cryptic and windows is truly a morons operating system which is what the vast majority of users in the workplace are.
3. The cost of hiring systems administrators is pretty close of linux vs. windows, but the cost of deploying software and the simplification that microsoft has deployed in this area is still untouched.
again, my argument is staged more to linux in the workplace and not in the end users hands which is probably where linux has more potential to grow.
prepare to see this posting get modded all over the place
Those who trade in their freedom for security, deserve neither.
What people will care about is, "Can this run my digital camera? Can I run the Sims on this? No? Oh. Convicted monopolist? I don't care, I don't use my computer that much anyway. I just want to play games and use my camera..."
The USSR, the plantation system, the railroad barrons, the oil barrons, the shipping tycoons.
... Well the fact is, MS's isn't competing against an opperating system, they are competing against a superior paradigm - and their half trillion market cap is nothing compaired to the yearly output of global industry. If they don't go with the flow, they will get squissed like a bug. like it or not.
Alot of times people have this misconception that something can be too big, too huge, too much talent and resources behind it to fall from greatness. This isn't true. How many times have we herd that "MS won't let it happen"
I think one hurdle Linux deveopers need to get past is their naming. Acronyms, abreviations, and random letter groupings culled from what the app does is not "user friendly".
"huh? Where is the install wizard?"
"oh, well you use RPM to get it."
"wha? What does how fast my motor is going have to do with installing my Video watcher thing?"
Maybe because they are names. And they come with a nice clickable installer. If they were called fbrsr and mrd, and required a full build from source on Windows, how much do you think their user base would grow?
I am relatively computer literate, but if I have a choice between something that needs building from source, and a nice installer, well, computer savvy or not, I am also lazy.
If you are one in a million, then there are six thousand people who are just like you.
This is pretty simple to sum up in my mind. Although my desktop is still running windows at work, at home I rarely see it.
But when asked the question why I have moved to Open Office from Microsoft Office, and why I have moved to Linux from Windows, what is the answer?
It's mostly about rights and freedom. I'm not yet willing to admit that I am a full out FSF supporter, though I have been a supporter of the Open Source movement. Microsoft's licensing tactics (and not just theirs but the general tactics of many other folks have led me as far away from proprietary "treat-the-custer-as-a-theif" software as I can possibly get.
Linux is great, and it has been an incredible learning experience (I've honestly never felt so dumb sitting in front of a command prompt as I did during my first Gentoo installation).
I was never a *NIX user. I never had any desire to run anything other than Windows because I was happy with the product.
But they forced me to look elsewhere, and when I did I learned what I was missing.
So IMO, what lies ahead for linux is more users...and I don't believe that is limited to the server. From the desktop side, the strides that have been made in KDE and GNOME in the last couple of revisions have made them dramatically nicer to work with. From the server side...not having to have a GUI running on a server is quite a bit more efficient.
Back in the day I remeber Microsoft recommending you change the screen saver to the black screen instead of one of those OpenGL screen savers on your Windows NT SQL server because the screen saver would bury your processor. I couldn't help but think why do I have this huge GUI running on what is supposed to resemble a somewhat powerful database server?!!
"God is dead!" - Nietzsche
"Nietzsche is dead!" - God
Industry analysts are often wrong. If they were on target all of the time they wouldn't give out advice. They'd instead make a killing on the stock market.
link
Replace "NT" with "Longhorn" and change the dates and it still works!!!
- "When you want something with all your heart, the entire universe conspires to give it to you" -Paulo Coelho
As an ex-analyst who moved back to software development, I would add a few other things for my fellow Slashdotters:
... "In that time we'll see tremendous growth" ...'"Tremendous" means that we're going to see it move from being a fringe market..." I suppose I agree with Stacy about her actual conclusions, but the phrasing struck me as being about as optimistically phrased as one could expect given the underlying statements about Linux on the desktop.
;-)
1) If you want to be an good analyst, you need to be able to write English; preferrably fairly easily and fairly well. Speaking skills can be learned on the job. Overcoming writers block probably can't.
2) Tech skills can give an analyst an important filter and BS detector which can be a competitive advantage versus other analysts. However, ability to communicate with techies does not pay off. Techies aren't spending thousands of dollars for insight. Managers are. Ability to communicate with management and market the value of the service you provide is the paramount skill for an analyst.
3) In my view, the important milestones that lie ahead for Linux all have to do with success as a database server. That's where the most critical business data is, that's where the money is, and if a company trusts their data to Linux, what will they not trust Linux for? It's also a technology space that's complementary to Linux's existing strengths in webservers and web services, and it plays well to Linux's developer (not end-user)-orientation while avoiding the desktop usability and UI-training issues where Linux continues to play catch-up. In terms of specific milestones, I would track the percentage of applications being deployed in Fortune 500 with Linux hosting the database. And I would track the growth of applications employing open source databases. A Linux firmly entrenched as a database platform is a Linux not easily dislodged by Microsoft-induced desktop trendiness. Witness the billions upon billions continually invested in mainframes and AS/400 if you doubt me.
4) I'm personally agnostic about whether Linux will ever make headway on the desktop. If pressed for a conclusion, I confess that I doubt it, although if I was afraid of the Linux advocate hordes, I might couch it like Stacy did: "potential for a lot of innovation"... "a lot of potential for Linux to become a much stronger play there"... "next milestone to look for is when Linux takes 10% of the market"
More constructively, in terms of adding to that 'desktop milestone' analysis, another milestone to watch for is when Linux desktop developers spend more time trying to understand how the Mac OS X guys tackle the usability problem than they spend trying to copy the Windows approach blindly in the techy details while missing the bigger picture.
I used to get paid 20k... now I'll settle for 2 karma. Ah the price of doing what you love...
--LP
I think a lot of the decision on which system (politics aside) comes down to whether or not you want to futz around with the OS or not.
I use Windows, also, and have found it to be very stable - but then again, I also don't experiment a lot with software, having figured out which stuff works for me AND I don't try to push the bleeding edge on hardware.
Windows is very stable, now. In common desktop usage, I daresay it can be as stable as Linux.
I used Linux for a while, but it lacked a lot of the software I needed to get my design work done.
To celebrate the occasion of my 1000th post, I will post no more forever on Slashdot. Goodbye.
I'm in agreement, but a few points:
:)
1) Stability is a big part, and I don't want to put words into your mouth, but other facets cannot be ignored. Performance. Ease of maintenance (service interruptions? reboot?). Remote administration. Batch-administration. Security. Lack of bloat (see Performance and Security as well). Available server applications. And lack of preparation or unique application training to accomplish these things. It's my personal observation that Linux beats out Windows in every area.
2) As far as I can see, most realistic people think Linux will take another 3-5 years to hit 10% on the desktop, including big Linux figures.
3) Administration is still the killer for Joe-user, but for companies with an IT department this isn't an issue. Considering Linux's put-your-home-and-usr-directories-on-NFS ability and how easy it is to mirror a box (no unaccessable 'files' on the filesystem), a company can roll out Linux without admin hassles. I honestly think this will be where it starts. People will use it at work and take it home (for work reasons or personal reasons). Companies will demand hardware support, user base will grow, and the snowball feeds itself.
Cheers
~Dalcius
Rome wasn't burnt in a day.
You're not in the minority, you're in the vast majority. It's just that most people in your position don't have any reason to discuss it. They just happily go about using their computer and having it work.
I'd rather be lucky than good.
I used to think that, but after doing some work with Win2003, I'm not so sure.
Win2003 is ok. It's just a version of Windows that sucks 20% less.
My main problem with Win2003 is that ther'es hardly any upgrade path.
With *nix you can grow as you need to from Linux to FreeBSD to Solaris to IRIX to AIX.
Hell, Linux has it's own upgrade path - Linux on ARM -> Linux on Intel -> Linux on PowerPC -> Linux on Sparc -> Linux on POWER5/6 etc.
With Windows, once you outgrow your 4 way Intel box - you're screwed. (We'll there Windows Advanced Server - but from what I've seen, its a bitch to keep running and the hardware it runs on sucks)
Moneyed corporations, non-working 'poor' and criminal prisoners are turning productive citizens into tax-slaves.
Here's why against your arguments:
... reformat. When entire chunks of companies are looking at nothing but the BIOS info they'll SERIOUSLY re-think the whole matter.
... will completely trash the system leaving you staring at nothing but ... BIOS.
... Sure, the need for better GUI based configuration routines are being worked on and coming. I will say there is nothing like coding for Linux sitting in front of OS X. :)
... while Windows will sometimes work, sometimes won't. Some Windows applications won't work right, or at all. Heck, some Windows patches require you to run around manually rebooting problem systems -- I've seen 1/10th the headaches dealing with NBM systems.
... and watch as business WILL roll with Linux ... and care to bet what the home users follow with? I can't count how many Linux distro CD's I've sent home with people who's 95 or Me box did this or that and won't work right anymore... One of the reasons Microsoft made it to the top was BECAUSE of the pirating going on. Ssshhhh, here, take it. It'll be OK. Well ... we, the geeks, FUCKED IT UP. We, the geeks, WILL fix it. The best part? It's not illegal this time...as Microsoft is pinching their users with activation keys and phoning home.
1. One of these Blaster type worms will come along. AV software won't catch it while it migrates through web servers (and then clients using IE), also via Outlook, and of course the direct connections. Login
1b. Another real option (based on Microsoft's history of code writing) is that one of these updates that comes along -- which EVERYBODY is trying to install quick and fast
2. OO or WordPerfect (for Linux) sure don't seem cryptic to any of my users. Click File, Open,
3. Have you deployed large scale software roll outs for Linux? Or patched hundreds of systems that needed it due to, oh my gosh, a flaw that was found (and typically fixed if it is serious within 24 hours)? I've done it for Windows, Linux, and OS X. OS X wins hands down (GUI or command line is trivial to deal with), Linux can easily be made to work "magic"
May you be modded up
An average computer user today does not type. Well, you can expect him to type a letter in a wordprocessor. But that letter comes from his head, and is easy to read. Computer commands do not come from his memory (as opposed to you, me and other computer professionals), and so he would need to ask, in this order:
In other words, the consumer is now totally GUI oriented, and only UNIX heads still remember what a command prompt is.
Note that Redcarpet, Up2date, RHN (and probably others) do auto-dependency like APT-GET.
I used the Redcarpet GUI when I first started with Ximian. It was very helpful and almost magic for a newbie. After a few months, I got 'rug' - the command line version (called 'rc' at the time). It was so much more flexible and efficient, I have never used the GUI since. For example, instead of paging through package listings with blearly eyes, I can just list them to a file with rug, and search the list with vim.
I started out my Dad on the GUI, and although he is no unix head, once familiar with the concepts, he found the 'rug' command line version much more efficient as well.
So my point is that you need both. I always recommend providing a solid scriptable command line interface first because it gives you the most bang for the buck (or hour). But then, a good GUI helps sell the program to new users by looking pretty and inviting and by helping to visualize concepts used by the program.
Because analysts tend to play the role of pundit they can come across as insightful or just plain idiotic. [note: fixed typo on "plain"]
:-)
Best quote ever. Darn, it's refreshing to find an honest, non-pompous analyst.
Yes, one thing. I go to a lot of events where I can be the only woman in the room with a bunch of guys, and that's fine. I have no issues with that, really, except that I just think that more diversity in the Linux ecosystem is always good. I think it is great that Pamela Jones created Groklaw. It would be great to see more women developers involved too- there are a few, but seeing more of them would actually be better. The growth of Linux in India, Brazil, China and other countries may foster an increase of women in the community. I think that's probably one of the things that, if I could effect any change, it would be to encourage more women to enter the Linux ecosystem.
That is actually a facinating point.
I've tended to find that as a very rough, general rule, women tend to do a better job of getting along with people than men, and take longer to get angry. If I had a choice between a male or female manager, and was choosing only based on ability to get people to work together and only with knowledge of the gender, I'd probably pick the female manager.
This is especially true for the open-source world, where nobody is *made* to work together. Communities form around how well people deal with each other and work together.
My guess as to why there are few female developers comes down to drive. This isn't that there aren't driven females, but there is a difference in the psychology here. I was reading an article (listed on fark and Metafilter) on why many fields of science generally have breakthroughs done by relatively young people -- developments and interest in work for the sake of work and glory fall off after a certain point. The article drew a link between drive to impress females and the attempt to rack up accomplishments under ones name. (I got a kick out of this, and it stuck in my head -- apparently, my subconscious has been trying hard to improve my sex life by convincing me to code up new algorithms). Anyway, point is that there's at least some research evidence for the male personality being an easier fit for OSS.
Linus' claim for support of "a law to get geeks laid" could have been OSS's undoing.
May we never see th
Knowledgeable users will choose the best software for a specific job they need to do and the best knowledgeable users are those that take the time to investigate both Windows, Linux & other OSes as providing the possible solutions to computing problems that they need to overcome.
It is more important to focus our attention on Open Source software to ensure that the scourge of proprietary formats is wiped from the face of our planet. There is nothing wrong with using commercial software as long as there is an interoperability with Open Source software such that everyone can exchange the data they want to with any people that they need to.
I fully accept that there are security issues in Windows just as much as there are learning curve concerns with Linux.
But the Windows community should embrace Open Source software much more readily than it currently does - for example, Mozilla/Firefox should now be the number one browser because it is free, available on most platforms, and conforms to the HTML standard much more than IE has ever done and will do.
Furthermore, we should all stop being hypocrites. If we are not prepared to pay the going rates for commercial software then we should all actively seek to use (and better) the Open Source alternatives. It is wrong to sit back and wait for OpenOffice.org (for example) to reach 100% compatibility with MS Office while using an illegal copy of MS Office - instead, we should use OOO with equal passion and give our opinions (and time) to the OOO developers to ensure that the product (and others) go in the direction that we need them to go.
It's now the time where we should all grow up a little and take some responsibility for ourselves and how run our computers. MS exists because there is a demand for their products and, if you don't like their products or the way they do things, the best way to get them to change is not give them your money... it's that simple.
I still use Windows 2000 and MS Office because I quite like both as products and because my place of work provides both as tools to me - however, I use Linux more because I've worked hard to learn it and am fortunate to work in a company that embraces Linux also. Going forwards, I will strive to migrate fully to Linux because I personally loathe MS's business strategy and will never pay them (or anyone else) money to access the data and information I already own due to proprietary format licenses.
But in the mean time, I have both work and leisure activities to do on my computers and I am not going to make either harder purely because media pundits believe every computer user is on one or the other side of a non-existent Windows v Linux war.
Be aware of some of the dangers of commercial software, sure, but otherwise use what's best for the job you need to do - as a result, you will be more efficient and find your computing experiences much more fun.
Gentoo Linux - another day, another USE flag.
When people say software installation is not easy, don't spit back the trite, unhelpful answer, "Just type [apt/emerge/urpmi/...] package_name" That's generally not what they're talking about.
.tar.bz2 file and tell it to install.
I've been looking at and reading about Linux for a while and recently begun using it, and I've already found several programs I want to use that are not in any packages. They just come in source, and I don't know how to compile/install source code yet. I have asked on Slashdot before if there is a graphical front-end for compiling programs from source, but haven't seen any answers yet. It would be nice to have a program where I can select the
If I have to run gcc manually or something to compile it, how do you know which compile options to use?
I've found no problems with packages--they do install and uninstall easier than Windows programs, but there needs to be an easier way to install non-packaged stuff. I don't think package dependencies are much of an issue on any of the several distros I have tried now.
We may experience some slight turbulence and then...explode. -Capt. Mal Reynolds