South Africa native and current London resident Mark Shuttleworth, founder of Canonical Ltd. and the Ubuntu Linux distribution, told DesktopLinux.com Friday in an interview that widespread adoption of Linux on the desktop – so long-awaited by many people. “Yes – I think Linux will be the dominant platform. It already defines the landscape in the server space (from supercomputers to YouTube). The desktop is just a matter of time.”
Interview: Mark Shuttleworth
144 Comments
2006-11-19 4:14 am
archiesteel
BTW, I also checked my Firefox user agent string (the default one for Firefox 2.0, which I didn’t alter) and it doesn’t mention Linux. It does say Ubuntu/Edgy, but nowhere does the word “Linux” appear, so unless that site’s statistics tool has a filter that tags “Ubuntu” as “Linux” (which I doubt), then I won’t be counted as Linux when I visit it.
Again, this is the default option of the most popular browser on the most popular distro, a fact that supports my argument and weakens yours. Sorry.
2006-11-19 5:16 am
NotParker
BTW, I also checked my Firefox user agent string (the default one for Firefox 2.0, which I didn’t alter) and it doesn’t mention Linux. It does say Ubuntu/Edgy, but nowhere does the word “Linux” appear, so unless that site’s statistics tool has a filter that tags “Ubuntu” as “Linux” (which I doubt), then I won’t be counted as Linux when I visit it.
http://en.wikipedia.org/wiki/User_agent#Browsers
Wikipedia suggests otherwise.
“Firefox 2.0 on Ubuntu Linux — Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.8.1) Gecko/20060601 Firefox/2.0 (Ubuntu-edgy)”
Maybe you are confusing the string sent to web servers versus the user agent you can set.
You should read up on it.
Try this site to detect your browser settings: http://projectip.com/
As for now, I’ll just assume you are ignorant or lying.
Edited 2006-11-19 05:21
2006-11-19 5:32 am
archiesteel
Okay, you’re right about the default Firefox string it seems, but you still didn’t know what you were talking about for Konqueror. I visited the site with Konq (my usual browser) and it registered as IE on Windows 2000. When I casually glanced at the Firefox user agent value in about:config, I thought it was an accurate representation of the string sent to the web server (like it is in the Konqueror config dialog). I was wrong on that specific account, it seems.
So I’ll admit I spoke too soon as far as the default for Firefox on Ubuntu is concerned, but only if you admit you had no idea what you were talking about when you posted the Konqueror link, or the useless nature of counting individual hits for web statistics without somehow trying to figure out which ones come from the same PC. As you haven’t provided counter-arguments for these points, I’ll take it you have abandoned trying to defend these points.
In other words, my argument still stands.
FYI, I have better things to do than lie on web sites. Unlike you, I don’t have a hidden agenda to discredit anything or anyone. That’s why I have a good trust rating, and you don’t.
Edited 2006-11-19 05:34
2006-11-19 6:13 am
NotParker
Okay, you’re right about the default Firefox string it seems,
Of course.
or the useless nature of counting individual hits for web statistics without somehow trying to figure out which ones come from the same PC.
You could read up on web analytics on Wikipedia:
http://en.wikipedia.org/wiki/Web_analytics#Logfile_analysis_vs_page…
But the fact is, many major web analytics companies put Linux usage at .4% or less.
Maybe you have a scientific reason backed up by research that proves Linux is over or under represented. I’d love to hear it. Anecdotes don’t count.
2006-11-20 8:44 pm
archiesteel
I’ll repost my previous post, since it was modded down for no good reason:
Okay, you’re right about the default Firefox string it seems, but you still didn’t know what you were talking about for Konqueror. I visited the site with Konq (my usual browser) and it registered as IE on Windows 2000. When I casually glanced at the Firefox user agent value in about:config, I thought it was an accurate representation of the string sent to the web server (like it is in the Konqueror config dialog). I was wrong on that specific account, it seems.
So I’ll admit I spoke too soon as far as the default for Firefox on Ubuntu is concerned, but only if you admit you had no idea what you were talking about when you posted the Konqueror link, or the useless nature of counting individual hits for web statistics without somehow trying to figure out which ones come from the same PC. As you haven’t provided counter-arguments for these points, I’ll take it you have abandoned trying to defend these points.
In other words, my argument still stands.
FYI, I have better things to do than lie on web sites. Unlike you, I don’t have a hidden agenda to discredit anything or anyone. That’s why I have a good trust rating, and you don’t.
——-
I know about Web Analytics, and this is precisely why I’m saying that they suck to establish things such as market share.
I gave you precise reasons why they are inaccurate, reasons which you choose not to try to reply to, therefore I must conclude that you concede defeat.
2006-11-19 10:33 am
hal2k1
//most Linux browsers masquerade the OS as something else,
Evidence? //
http://en.wikipedia.org/wiki/User_agent#User_agent_spoofing
https://addons.mozilla.org/firefox/59/
2006-11-19 10:38 am
Babi Asu
At the same time, this is also evidence that Windows browsers masquerade the OS as Linux.
Please give a smart and convincing evidence.
Edited 2006-11-19 10:39
2006-11-19 9:58 am
deb2006
Well, there is:
– South America
– Africa
– Middle East
– China
– Russia
– Japan
Linux is gaining market share especially in South America, Africa, and in China.
So, no, I don’t think Mark Shuttleworth is overoptimistic in any way. I like the way Linus Torvalds described it: Linux may not not conquer the desktop today, maybe not even tomorrow, but it is going to hapen.
Edited 2006-11-19 10:02
2006-11-18 7:11 pm
diegocg
Msft, with their enormous influence, will make sure Linux catches up.
I hope they’re more succesful than they’ve been for now. They’ve been trying to stop linux for a long time and haven’t succeed
2006-11-18 7:39 pm
valnar
That’s part of the reason. The other part is Linux itself. It’ll never replace Windows because, well, it simply wasn’t meant to. Underneath the hood, Linux is vastly more complicated than Windows as far as the user needs to be concerned. Windows is complicated too, but most of that is hidden from the user.
Windows was made from the ground up to be reliable and simple, once again, as far as the user is concerned. It runs applications easily, it works, and everything meant for Windows installs okay.
Linux is a Unix clone and from its initial design, was never meant for end users as a desktop. So unless you want to change Linux to something it is not, it will *always* be too complicated for the average user. Having multiple distributions and package managers doesn’t help either. The thing that makes Linux so great will be it’s Achilles heel. Most don’t want that many choices. Linux gurus do.
Here is another analogy:
Windows is a Toyota. It just works. There aren’t too many options, and the few are available for anyone to use.
Linux is a hotrod. It can be made to do anything you want, provided you know what you want, and know how to do it. If you gave a hotrod to a grandma, she wouldn’t know how to use it, customize it, or even keep it running for long. Sure, as her hotrod mechanic you could always build one for her, but the power behind it she will never understand. They are also far more complicated than a Toyota.
-Robert
-
2006-11-18 8:10 pm
AlmindorThis is flawed logic, and ubuntu proves that. You need to realize that linux can hide stuff exactly like windows. The main problems are actualy hiding it good enough with so many choices.
Right now the only department where linux lacks is ABI.
-
2006-11-18 8:21 pm
czubinand why can’t linux hide these nerdy components?
look at mac os X , it’s unix but still more user friendly then windows
Initial design has nothing to do with ‘human computer interaction’; it just must stay hidden, easy accesible, …
-
archiesteel
The nerdy components *are* hidden in some Linux distros, such as Ubuntu. They are still available, though, just like they are on OSX.
I’ve put total newbies in front of a Linux computer many, many times, and they managed fine. “Linux is hard” is just more FUD from MS fanboys and astroturfers.
-
2006-11-18 9:34 pm
g2deviThe Mac OSX is a good counter-example that Unix has to be hard. Ubuntu is a close second, although there are a few issues to be resolved:
* We need a good open source NVidia driver (Apparently, X.org is working on one now and it should be ready in about 8 months. Yah!), since relying on closed source is problematic.
* X.org needs to fallback to “safe mode” if there are configuration issues and not just crash. The next version of Ubuntu is working on a stopgap for this issue but the X.org people are working to implement this directly within 8 months or so (Yah! again X.org)
* We need open source alternatives to Flash and Java since these are used everywhere. It’s easy to install both, but it leaves users confused and you have to wait for Macromedia to put in the effort (this may take a long time, witness the lack of Flash 8 on Linux). Fortunately Java will be free software soon (Yah Sub!) so this is one hurdle out of the way. Gnash will take time, but hopefully it’ll get there.
* More open source audio and video codecs need to be supported universally. It’s easy to use “common customizations” like Automatix and EasyUbuntu make upgrade or do it yourself, but it weakens the user experience again and in some cases, it can cause problems with upgrades since additional repositories may override the officially supported ones. The next version of Ubuntu will try to make it easier for the user by detecting not yet installed codecs and either provide the user instructions on how to install them or install the codecs (after the user agrees that it’s okay and reads the licensing issues).
* Upgrades need to take into account that “common customizations” happen and either try to work around them (e.g. undo them if unsafe for the upgrade) or let the user know that he/she is rolling the dice if the upgrade proceeds. Apparently the next version of Ubuntu will handle this too.
* Compiz or Beryl eye candy needs to be preinstalled (I don’t care for it, but joe/jane users do). The next version of Ubuntu will address this and some other distros already address this.
So overall, Linux (or Ubuntu at least) isn’t too far from overcoming its last remaining issues for Joe/Jane Users. Give it a year. These things take time. Remember, Windows didn’t *really* start to get popular until 15 years after DOS was introduced, and Microsoft was *trying* to make a desktop environment from those early days. Linux is a bit more than a decade younger than DOS and has only turned its attention to usability (for Joe/Jane user) in the last few years. Before that, it was more focused on the Unix user or Windows Power User that likes to play with new stuff. You can’t expect miracles, but it is amazing how far it’s come in such a short time.
-
valnar
Indeed, it has come a long way. But my original statement holds true. An end user has to marry himself to a particular distribution, and those applications written for either Gnome or KDE to get seamless functionality. This doesn’t even address all the other Window managers. Like I said, what makes Linux unique will never go away. Everything that runs on Windows works on Windows. The same cannot be true of Linux. And because of so many different formats, package managers and binaries, it’s impossible for any given user to download ANY program he feels like and know that it will work. I don’t see that going away anytime soon.
(Yet another example)
The day the new DOSBox 0.65 came out, I had no problems installing it on Windows. It worked great, including sound support. Ubuntu was months behind. And even now if I want the latest (non-CVS) version, I’ll need to run Edgy, not Dapper. Sound still doesn’t work right. ‘Too many cooks in the kitchen writing their own audio standards… So basically I have to install a whole new version of the OS just to run the latest application? Tell me how a regular user will deal with that?
Two choices. Install from source, or upgrade their OS to the latest Ubuntu. Pretend I’m computer illiterate before you respond.
Linux needs to come a long way – that’s true. But if the course it’s headed in doesn’t change, it won’t matter if its 5 or 50 years from now.
Robert
Edited 2006-11-18 22:19
-
2006-11-18 10:39 pm
Redeemanonly one tiny problem in your doomsday “same cant be said for linux” theory… you act as though applications using gtk+ cant be run from within kde, and reversed, which clearly isnt true.
-
-
2006-11-19 12:04 am
MoochmanI modded you up, but I also want to point out that Linux is heading in new directions on exactly this front (software installation). Most visible and likely to succeed projects: Autopackage and klik. Autopackage promises to make all new software releases easily packagable for all Linux systems simultaneously while providing end-users with the easy, wizard-style install they’re used to from Windows. If it becomes widely adopted enough, it will mean that even newbies will be able to install all manner of new versions of software just as easily as on Windows without waiting for the official packages. Meanwhile, klik makes it even easier: just click on a package name online and it installs and runs itself, not even admin rights are necessary. Klik, however, will likely never get rid of the package availability delay problem, since its packages AFAIK are based off of Debian’s. Nevertheless, with these two options (not to mention other solutions like the Smart package manager and GoboLinux’s App-Directory approach) still fresh on the scene, the future of Linux software installation looks _very_ promising.
EDIT: I forgot to mention that many software vendors tend to favor binary, package-manager-independent installers anyway (witness Java & Firefox, for instance), so that’s another solution to the problem. And finally…. Who are we talking about here? If it’s really the know-nothing users, how likely are they really to download the latest version of their software as soon as it comes out? If they care that much about the latest and greatest, I’d guess they’re more likely to be power users.
Edited 2006-11-19 00:11
-
hal2k1
//Everything that runs on Windows works on Windows. The same cannot be true of Linux. And because of so many different formats, package managers and binaries, it’s impossible for any given user to download ANY program he feels like and know that it will work. I don’t see that going away anytime soon. //
Not correct.
All of the same packages work on all distributions.
Take for example the CD/DVD burning application, K3b. This is a KDE application. There is not one distribution that cannot run it.
There are 20,000 packages in the Ubuntu repositories. All of them work on all variants of Ubuntu … Ubuntu, Edubuntu, Kubuntu or Xubuntu. You can get all 20,000 to work on RedHat, or SuSe, or Mandriva or Slackware.
The only differences between distros are the packages installed by default. This does not mean you cannot change the default installed packages.
Please desist with 100% incorrect FUD.
-
2006-11-19 10:26 am
jbauerPlease desist with 100% incorrect FUD.
I think you are the one who should stop shouting FUD when you disagree with ideas from others. What he is saying is totally correct, no matter how you twist it to make it look false. By the way, I guess Mr. Shuttleworth was spreading FUD too when he called for package standardization?
http://www.markshuttleworth.com/archives/66
-
hal2k1
//What he is saying is totally correct//
What he was saying was totally incorrect.
Read my lips. Almost every package for Linux will work on every distribution.
The package formats are not compatible, it sometimes takes some lesser distributions a good while to catch up with some applications, and it wouldn’t hurt to have a common standard, but it isn’t strictly necessary and in NO WAY KNOWN does it mean that any Linux applications will only work on one distribution or another.
This claim from the original post “Everything that runs on Windows works on Windows. The same cannot be true of Linux.” was and still is 100% pure unadulterated FUD. It is ABSOLUTELY incorrect.
In fact, it is quite backwards. In reality, far more often for a given Linux package it is far easier to get it working on all varieties and all versions of Linux than is the case for Windows packages to get working all varieties and all versions of Windows.
Tell me, where can I buy “Office 2003 Windows 95 backport”?
In fact, the very discussion of backports http://en.wikipedia.org/wiki/Backport mentions the fact that backports are far less common for closed-source applications.
Edited 2006-11-19 10:56
-
jbauer
You’re twisting his words again, he’s not saying that certain apps only work in a given distro. What he is saying is that if you want to download a new software from the internet, you’ve got to make sure you download the right package for your specific distribution, and the specific version you are using.
Since developers usually have no resources or knowledge to create every package possible for every combination that exists, some users are always left in the cold at least until the distributions packages the software. If it’s not a popular software, it might never be packaged at all.
Now, to make my point, take a look at the download page for basket (impressive application, by the way, I encourage everyone to try it), and tell me that what’s there is sane:
http://basket.kde.org/download.php
-
hal2k1
//You’re twisting his words again, he’s not saying that certain apps only work in a given distro. What he is saying is that if you want to download a new software from the internet, you’ve got to make sure you download the right package for your specific distribution, and the specific version you are using. //
Nope.
Here is the direct (totally untwisted wording) quote:
“Everything that runs on Windows works on Windows. The same cannot be true of Linux.”
That is wrong, wrong, wrong. 100% incorrect. Utterly backwards, in fact.
It turns out that you are the one twisting words, not I.
And further, even this imagined problem: “you’ve got to make sure you download the right package for your specific distribution, and the specific version you are using.” is not a problem. Just use the package manager provided with your distribution, and you simply do not have to worry about any of that at all.
Or alternatively, use the Smart package manager, point it to any and all sorts of repositories, and it will work out the best solution for you.
Edited 2006-11-19 11:08
-
jbauer
Come on … if you’re going to quote him, at least do it right, and don’t cut the quote in the middle of his argument.
Everything that runs on Windows works on Windows. The same cannot be true of Linux. And because of so many different formats, package managers and binaries, it’s impossible for any given user to download ANY program he feels like and know that it will work.
This is getting too surrealistic. I give up.
-
hal2k1
//it’s impossible for any given user to download ANY program he feels like and know that it will work.//
100% incorrect. Utterly wrong. Just use a package manager, select whatever you want, and it will work. Clearly it is not IMPOSSIBLE at all.
It is more of a correct statement for Windows (say a Windows 98 user) than it is for Linux.
//This is getting too surrealistic. I give up.//
So you should. You are wrong.
-
2006-11-19 3:49 pm
dagwJust use a package manager, select whatever you want, and it will work.
Assuming they have a package of the particular version of the particular program you want. And that all the packages for the relevant dependencies are equally up to date. And assuming that the package isn’t broken, which tends to be the case with many of the more obscure packages which have been abandoned by their package maintainers.
-
hal2k1
//Assuming they have a package of the particular version of the particular program you want. And that all the packages for the relevant dependencies are equally up to date. And assuming that the package isn’t broken, which tends to be the case with many of the more obscure packages which have been abandoned by their package maintainers.//
Correct. Which is what they check on before the put packages in distribution repositories. And which is why there are 20,000+ packages in Ubuntu’s repositories, and a similar number in Debians.
Which is why installing from your distributions package manager just works. Even for some extremely obscure packages.
Thank-you for making the point.
BTW, there is nothing like this available for Windows. Windows is light-years behind here.
Edited 2006-11-20 08:41
-
jbauer
BTW, there is nothing like this available for Windows. Windows is light-years behind here
Sure, that’s why I’ve been using Firefox 2 in Windows for weeks, while I’m stuck with Firefox 1.5 in my Suse 10.1 workstation. So much for the package management brilliance.
And no, I don’t want to upgrade the whole distribution just to be able to use a new version of a new program, that’s insane. Neither I want to hunt down unsupported RPMs or obscure repositories through the Internet. No one expects you to upgrade Windows just to install a new program, and in fact, hell is raised by angry users when that rare fact happens.
But hey, it’s always nice to put your hands over your ears while shouting “dahdahdah everything is perfect in LinuxLand dahdahdah”. If I had to point out a single reason why Linux desktop isn’t here yet is because of that kind of thinking.
Edited 2006-11-20 12:10
-
hal2k1
//But hey, it’s always nice to put your hands over your ears while shouting “dahdahdah everything is perfect in LinuxLand dahdahdah”. If I had to point out a single reason why Linux desktop isn’t here yet is because of that kind of thinking. //
I don’t know why you like to pretend there is a problem here when there clearly isn’t.
If you want cutting edge, then point to cutting edge repositories. If you want stability, then point to stable repositories.
You can get both, for just about any distro. The only ones for which this isn’t true are small “fringe” distros. But if you wanted good availability of packages, why did you opt for a fringe distro (such as Arch) with a small repository base in the first place?
Don’t pretend it is a problem to switch from one distro to another, either, when it just isn’t. I myself have done this multiple times without losing any data, home directories or even personalised application settings.
If you are interested in fetching all sorts of apps from alien repositories rather than waiting a few days for them to become available for your chosen distro, then you can do that too. You can get Smart package manager even for Suse.
Seriously, you don’t have a single leg to stand on here.
Get a clue, and go whinge about something that is actually real. Like for example the lack of Office 2003 for Windows 98 users.
Edited 2006-11-20 12:30
-
jbauer
Yep. It’s outrageous that I cannot use Office 2003 in an almost 9 nine years old Operating System. That I need to go through black sorcery to install this, you know, completely unpopular program that no one has heard about called Firefox 2 in my 6 months old distribution is perfectly normal, and acceptable, right?
But you are right, it’s not a big deal. I can always upgrade my distribution or switch to another one, if my time is worth nothing. Or I can stick to Windows, where installing Firefox 2 took me exactly 3 minutes the day it was out. And I bet that is what many users would do, and that among other things explains why Windows is so successful and Linux is, unfortunately, still nowhere on the desktop. But you can call that FUD too, if you so desire.
I see also that my last comment has been moderated down almost instantly. Well, truth hurts sometimes.
-
2006-11-20 3:58 pm
elsewhereThat I need to go through black sorcery to install this, you know, completely unpopular program that no one has heard about called Firefox 2 in my 6 months old distribution is perfectly normal, and acceptable, right?
http://repos.opensuse.org/mozilla/
There you go. Up to date builds for 9.3, 10.0, 10.1, SLES/D and factory. Simply need to add it as a repo to Yast, Smart or whatever you’re using. No black magic. What more do you want?
I’m not aware of any mainstream distros that offer “supported” application version upgrades post-release. But most of them do make them available through standard repository channels, just not by default and there’s a valid reason for that.
openSuse, frankly, has one of the best methods for supporting newer and bleeding edge applications for their distros, and in fact even does some packaging for non-Suse distros via their build service which simplifies package creation and maintenance for the devs, leading to wider availability of packages. You’ll find both openSuse and community generated packages, ranging from backports of popular applications to things like svn release snapshots of kde and gnome devel libraries.
Version freezes in linux distros are a strength, not a weakness and besides they are easy enough to work around.
-
hal2k1
//
http://repos.opensuse.org/mozilla/
There you go. Up to date builds for 9.3, 10.0, 10.1, SLES/D and factory. Simply need to add it as a repo to Yast, Smart or whatever you’re using. No black magic. What more do you want?//
Exactly.
As I said before, “If you want cutting edge, then point to cutting edge repositories. If you want stability, then point to stable repositories. You can get both, for just about any distro. “
So that then is the solution for “cutting edge” repositories for OpenSuSe. Just add that to your list of repositories, and you are good to go to install Firefox 2 on SuSe.
-
archiesteel
Sure, that’s why I’ve been using Firefox 2 in Windows for weeks, while I’m stuck with Firefox 1.5 in my Suse 10.1 workstation. So much for the package management brilliance.
Complain to SuSE then. My Ubuntu Edgy laptop uses Firefox 2, installed from the repositories.
-
2006-11-19 11:17 am
netpythonYou’re twisting his words again, he’s not saying that certain apps only work in a given distro. What he is saying is that if you want to download a new software from the internet, you’ve got to make sure you download the right package for your specific distribution, and the specific version you are using.
Not necessary allways true.
The dependencies aka “the right app” are handled by advanced package managers such as apt,emerge..
You’re right when you say all is distro specific.
Edited 2006-11-19 11:18
-
hal2k1
//Not necessary allways true.//
In fact, nearly always untrue.
//The dependencies aka “the right app” are handled by advanced package managers such as apt,emerge.. //
Exactly. Just use the package manager that comes with your distribution, and you will have none of these “imagined” problems at all.
//You’re right when you say all is distro specific.//
Yes, but it totally ignores the undeniable fact that the vast majority of useful Linux applications are found in all distribution repositories, and that there exist package managers that can work with repositories of different distributions.
It also ignores that the so-called “problem” is rapidly fading into non-existance with efforts such as the Linux Standard Base and the Portland Project.
Edited 2006-11-19 11:33
-
hal2k1
//Now, to make my point, take a look at the download page for basket (impressive application, by the way, I encourage everyone to try it), and tell me that what’s there is sane:
http://basket.kde.org/download.php
//
Sigh!!
Windows think!
Don’t download the app from basket.kde.org, just get it via your package manager.
For example, the version in Ubuntu repositories is detailed here:
http://packages.ubuntulinux.org/dapper/kde/basket
for dapper, and here:
http://packages.ubuntulinux.org/edgy/kde/basket
for edgy. The package managers sort all this out for you.
The version in the repositories is going to be a little behind the latest version, but using the version from the repositories avoids every single issue you are trying to claim as a problem.
So users have two choices: (1) get the very latest package from the web, and be prepared to go on your own to get it installed (not impossible, but admittedly a bit ambitious for inexperience users), or (2) wait a few weeks and get it from the repositories (where the distribution maintainers have worked out all the kinks for your distro).
Either way, you can still get all applications for all distributions.
In the Ubuntu repositories, basket is available for all releases of Ubuntu, and it has even been made available for feisty which isn’t released yet.
http://packages.ubuntulinux.org/cgi-bin/search_packages.pl?searchon…
Edited 2006-11-19 12:01
-
2006-11-19 12:30 pm
ahwayakchih//Now, to make my point, take a look at the download page for basket (impressive application, by the way, I encourage everyone to try it), and tell me that what’s there is sane:
http://basket.kde.org/download.php
//
Sigh!!
Windows think!
Don’t download the app from basket.kde.org, just get it via your package manager.
You’re right that using package manager is very easy, but only after user knows where it is in menu. Of course it’s because of “windows-like thinking”, but we can’t avoid that unless we can wait ~80 years for all Windows users to die (and only if all new computer users start from something other than Windows).
Wouldn’t it be MUCH easier for everyone to have “special” type of link for packages? Just like we have mailto://[email address] we could also have package://[package name] (maybe with additional path like “/[version number]” for manager to use as a hint). And each distro would have own handler for such link (so it wouldn’t work as a link to download anything from that webpage, but as a link to start installation process of package found in repositories of distro).
-
netpython
You’re right that using package manager is very easy, but only after user knows where it is in menu.
The same is valid for installing an new program on windows.You have to get acquainted first.
Wouldn’t it be MUCH easier for everyone to have “special” type of link for packages? Just like we have mailto://[email address] we could also have package://[package name] (maybe with additional path like “/[version number]”
where do you draw the line?
There will allways be users who don’t even know what todo with the links you suggested.
I think a GUI package manager looks the most familiar with what the majority of PC users know from working with windows.AKA windows add/remove,although linux package managers such as apt-get,emerge are much more advanced in my opinion.
Any slightly advanced windows user with average intelligence will without doubt feel in no time at home behind a userfriendly Ubuntu,PCBSD,etc desktop.
A lot of PC users don’t want to learn new windows programs much less a new OS.
What would be more interesting in my opinion is the experience of two equivalent newbies where one begins to learn windows and the other PCBSD or ubuntu.
If IBM hadn’t given MS a change a long time ago we would propably all use OS2 Warp now.MS didn’t reach the current level of marketshare with fierce competition.
Edited 2006-11-19 13:12
-
ahwayakchih
The same is valid for installing an new program on windows.You have to get acquainted first.
Yes, but majority of users are used to Windows-way of doing it.
where do you draw the line?
There will allways be users who don’t even know what todo with the links you suggested.
I think a GUI package manager looks the most familiar with what the majority of PC users know from working with windows.AKA windows add/remove,although linux package managers such as apt-get,emerge are much more advanced in my opinion.
I didn’t mean that package manager should go away, or anything like that. I just ment that adding special handler for links to packages would make things even easier for users. It would be just an additional way (or “interface”) to use the same system.
I never used “add/remove” thing to “install” something. Only for “uninstall” and only when application didn’t have own uninstall link/icon/app/whatever.
I started using package manager after Ubuntu (well… i used it on Knoppix too, but it was much less “friendly” there) and liked it. Even though i like it and use it i would find it much easier to click on link in webbrowser than to open package manager, find application by name, double click it and click “apply” button. And i don’t mean it in a way that “it’s too complicated for me to click more times”, but in a way that “it requires more time, and i’m lazy”
. -
netpython
I didn’t mean that package manager should go away, or anything like that. I just ment that adding special handler for links to packages would make things even easier for users. It would be just an additional way (or “interface”) to use the same system.
just curious,have you ever tried PCBSD?
Their PBI system is as easy as it gets,namely double-click on the file and the rest is pretty dead on.
-
ahwayakchih
just curious,have you ever tried PCBSD?
No, i have not tried it.
Their PBI system is as easy as it gets,namely double-click on the file and the rest is pretty dead on.
But You have to download package first, right? And it has to be in that specific PCBSD’s format, right?
If so, then it’s not different from .deb on Ubuntu or .exe/.msi on Windows (download, double click
. Yes it’s easy but it has to be specific format handled by distro/os.With the special html link i tried to describe, format/distro wouldn’t matter, as long as application is available in repository.
For example: there is “MyApp” application which is truly portable. So on webpage there are links to Windows, OSX and “Linux” versions (and sourcfe.tar.gz too of course
. There is one huge, default “Download!” link which changes target depending on OS used by visitor.“Linux” link uses package://MyApp/1.0 URL.
When user clicks on it, distro checks app’s name, finds it in it’s own repository and launches installation process. So it doesn’t download app from that webpage, but from repository.
If there’s no 1.0 version it can warn user that 1.0 is not available yet, but there is 0.99
.Now imagine that people create repositories and managers for Windows apps (Maybe using Tucows?), for OSX apps, BeOS/Haiku, ReactOS, SkyOS, etc… There would be only 1 link needed! one link for all different formats/binaries
. Of course there would be additional links for those who want manual installation, don’t have/use manager or if app is not available at all in repository they use.Additionally, when app is missing in repos, it could be “reported” (anonymously – just app name, and maybe site on which link was found) to repos admins so they would know about it and maybe add it to repo. That would make user life even easier because he/she wouldn’t have to write mail/forum post/etc… with question about app.
Edited 2006-11-19 15:45
-
netpython
The PBI package system is half way the scenario you described.It’s an nice ideal to protect the average user from overly complex installs.Though the scenario will undoubtedly have a tough overhead in terms of bandwith,dependency,development,etc..I’m afraid it’s economically committing suicide at the moment.
-
2006-11-20 1:51 am
WrawratSo users have two choices: (1) get the very latest package from the web, and be prepared to go on your own to get it installed (not impossible, but admittedly a bit ambitious for inexperience users), or (2) wait a few weeks and get it from the repositories (where the distribution maintainers have worked out all the kinks for your distro).
More like a few months if you are using a stable distribution with strict backport policies. For example, a new version of GAIM (like NetworkManager integration) came out after Ubuntu Edgy. I am pretty sure we won’t see it until Feisty, even with the existence of the edgy-backport repository.
Similarily, I have encountered some issues with quodlibet with Breezy. Unfortunately, packages in the Universe repository don’t get the same love as the ones in Main, so I had to wait.
Another anecdote: I have recently scoured the Net for a LaTeX GUI in GTK+ that didn’t suck. I found one on GnomeFiles that looked great, only to realise that it wasn’t in my repositories. Now, I have the proper knowledge to compile apps and deal with the possible issues (missing libraries, compilation errors), but I had other things to do (like writing that report in LaTeX) and wasn’t really interested in doing maintenance. In the end, I had to look somewhere else.
While I prefer repositories over standalone packages, they have some shortcomings. Not only you have to use a specific package format, but you are at the mercy of the maintainers.
I believe the original poster meant: a specific package won’t work in all Linux distributions, i.e. vim binaries inside a DEB package might not work in, say, Arch Linux. Of course, the same thing can be said for MS Windows. However, the compatibility work made by the developers to ensure backward compatibility and the relative homogeneity of the environment (compared to the ever-increasing number of Linux distro) leads us to a situation where most self-extracted installers can be used in all the recent versions of MS Windows.
I wonder how difficult it would be to set up a web service where developers could upload their software with a install script indicating its dependencies, ask it to compile packages in a virtual environment for the popular distros, then make them available in a web interface. Might be a good compromise since users could download software packages that could integrate with their package manager.
-
hal2k1
//For example, a new version of GAIM (like NetworkManager integration) came out after Ubuntu Edgy. I am pretty sure we won’t see it until Feisty, even with the existence of the edgy-backport repository. //
What new version?
According to the gaim website, the current version is 1.5.0, and they are still in the release cycle for version 2, having just posted the 5th beta. So version 2.0 isn’t released yet.
The version in the Ubuntu edgy repositories is 2.0.beta3, and the version available from here http://www.debuntu.org/gaim-2.0.0beta4-deb-package-ubuntu-edgy specifically for edgy is gaim-2.0.0beta4 .deb
That is more than reasonably current.
//I believe the original poster meant: a specific package won’t work in all Linux distributions,//
That isn’t what was posted.
// i.e. vim binaries inside a DEB package might not work in, say, Arch Linux. //
Since you can almost always get the same version of vim specifically for your Arch Linux, why would you try to install the deb bianry?
If you are one who itches for the cutting edge, and your distribution is “fringe” and never has cutting-edge packages, then why do you have that distro?
If you want cutting edge, get a cutting edge distro. If you want speed and don’t mind installing from source, get Gentoo. If you want stability and a good desktop and heaps of packages that work, get Ubuntu. If you are an utter newbie to Linux and want a dead-easy desktop-oriented distro with a reasonable number of reasonably current packages, get PCLinuxOS. And so on.
None of these ramblings about packages on Linux are actually valid complaints. There is a distro for all desires, and they all have decent GUI package managers. The install problem on Linux is well & truly solved, and it has been for a few years now, and Linux is way ahead of Windows in this respect now.
Find another topic on which to try to FUD.
-
valnar
It is not FUD. You are acting like a typical Linux bigot. My examples were exactly indicating software NOT in the package manager.
-
archiesteel
This is what you said: “it’s impossible for any given user to download ANY program he feels like and know that it will work.”
That is clearly FUD, and modding yourself up with sockpuppet accounts won’t change that fact.
-
ahwayakchih
We need open source alternatives to Flash and Java since these are used everywhere. It’s easy to install both, but it leaves users confused and you have to wait for Macromedia to put in the effort (this may take a long time, witness the lack of Flash 8 on Linux). Fortunately Java will be free software soon (Yah Sub!) so this is one hurdle out of the way. Gnash will take time, but hopefully it’ll get there.
Java is GPL now. And Macromedia (or more exactly: Adobe, because Macromedia was bought by Adobe) released Flash 9 beta for Linux (i’m using it currently and i don’t have any problems with it). Yes, it will take some time before they are easly installable (for Joe/Jane Users), but i don’t think it will be a long time.
-
2006-11-18 9:35 pm
anda_skoaUnderneath the hood, Linux is vastly more complicated than Windows as far as the user needs to be concerned. Windows is complicated too, but most of that is hidden from the user
I certainly agree that internals are well hidden on Windows, sometimes they are hidden too well
However I strongly disagree with the idea that internals cannot be hidden as good on Linux.
Numerous use cases like network appliances, set-top boxes or digital video recorders show that it is possible and doable.
Since operating a desktop system requires more features than operating a set-top box, it is just more work to hide the internals properly.
-
archiesteel
“It’ll never replace Windows because, well, it simply wasn’t meant to. Underneath the hood, Linux is vastly more complicated than Windows as far as the user needs to be concerned. Windows is complicated too, but most of that is hidden from the user.”
If this argument was valid, then Mac OSX would be the dominant OS on the market. It isn’t, therefore your argument is baloney. Sorry.
The reason Linux is not more prevalent is that it doesn’t come pre-installed, and that MS Office is not natively available for it – in other words, two reasons due to Microsoft dual monopolies of operating systems and office document formats.
Oh, and games, too, but with the game consoles we have today, this is becoming less and less relevant.
-
2006-11-20 9:59 am
collywolly“Linux is a Unix clone and from its initial design, was never meant for end users as a desktop”
Sure, but so is OSX, and no one seems to deny that that makes a good Desktop. There is a problme that Linux can do too much, and as such confuses users that are used to the Windows way of doing things. Macs do things differently, but look how many of their users rant about how they wouldn’t go back once they switched.
2006-11-18 11:49 pm
rayiner
There are a couple of problems with that argument. First, noting that Linux has been around for 10 years (actually, its about 15 years) is fairly irrelevant. A product doesn’t build market share without actually competing in a market in a focused way, and its only relatively recently that Linux companies have focused on the end-user desktop. What are the major players in the desktop Linux world? RedHat, SuSE/Novell, and Canonical. Of those, SuSE never saw the desktop market as a primary focus, RedHat still doesn’t, and Novell and Canonical have had Linux desktop products for barely two years now. Meanwhile, both MacOS and Windows were conceived from the very beginning as end-user desktop products, and have been marketed that way for more than 20 years. Moreover, they’ve been competing in that market while being backed by multi-billion dollar companies for what whole time.
Second, the actual Linux desktop share is 2-4 higher than your estimate, or well over 2%. You vastly underestimate the impact Linux is having in Latin America and Asia, which together account for a huge and under-developed market. There is a huge potential in those parts of the world for increasing Linux’s market share.
That said, I don’t think Linux will become the number one platform in desktop space. Note, however, that it needn’t be in order to be the dominant platform. There are a lot more cell phones than there are personal computers, and regular people interact with server or embedded computers more often than they interact with desktop machines. With Linux running on everything from cell phones to servers to supercomputers, it is already a dominant (if not THE dominant) platform even without a large desktop market share.
2006-11-19 2:39 am
archiesteel
“I’m not a basher, I’m just realistic.”
Then you should know that Linux has about 5x the market share you claim it has.
2006-11-19 6:32 am
backdoc
OK. These are valid valid points. But, I think a bigger problem for Linux is that each application is an island. Where as in Microsoft, each application is aware of each other. In fact, the integration is too tight for my tastes. But, it’s that tight integration that keeps corporations dependant upon Microsoft.
This is a tough nut to crack for Linux.
-
2006-11-19 11:12 am
TerracottaIn windows 95 (the first real popular windows), every program was a seperate island, and it was that that created the succes of windows: third party apps, by big companies, that’s what linux lags at the moment, game companies are just a fraction of the other companies that create decent windows software. If you go completely KDE you’ll find better integration than on windows vista. There still lies truth in what you said but it’s not the main problem, and it’s been adressed to by the opensource community of i.e. KDE a lot better than by microsofot
2006-11-19 4:03 pm
cerbie
Drivers and applications? Which ones? A real modem is cheap these days, and that’s the only major area where drivers truly lack. Applications working together is a little pain, but constantly getting better, and is now at least as good as Windows (mainly due to the clipboard generally working as I think it should between apps of different GUI toolkits).
Applications in general, though, are great. What is lacking?
I’ve not had any major driver problems recently that I can blame on anyone but Logitech, and honestly, that doesn’t, and didn’t, surprise me. However, even that is being worked on (using evdev), and appears to be more of a pain due to X and Linux interaction than either one on their own.
Now, I use Windows for my PC, and don’t see anything wrong with it. However, I don’t lack anything when using a Linux box now, I just everything tweaked like I want in Windows, and haven’t gotten disgusted enough at MS to truly switch.
Hardware support is generally superb in Linux, but getting the software configured to use is not. X is a royal bitch sometimes (MS and Apple both have setting the display up done far superior to anything else I’ve found), wireless equally so (but at least it is getting better), and sound can be, too. All of this even with perfect or near perfect hardware support.
2006-11-18 7:38 pm
samad
“That’s what’s happening in practice, and it’s the reason we see Microsoft waging a heavy war in education departments around the world, to try to keep those departments on Windows.”
I agree somewhat. At my university all of the computer science classes are all Linux/Unix based. Most computer science and bioengineering labs have computers that don’t even have Windows installed. I think Linux has a pretty solid following in the university, especially among computer science students who are willing to experiment and try different things. Macs also have a pretty solid following too, thanks to OS X.
2006-11-18 7:58 pm
acobar
I disagree. (Of course I may be wrong)
For most of these 14 years of existence Linux did not have a decent office suite. How long is OpenOffice in a good shape floating around?
The same can be said about a good browser, photo edit software, music/photo collection suites and the like.
Fact is, now these all around needed goodies are really getting in shape. Microsoft has an advantage, on Office and GUI, but it is getting less evident faster each day. Even the speed difference is getting less noticeable, thanks to new hardware (I’m talking about Office).
Hardware is also getting broader and better support. Speaking for myself, I hadn’t had a single miss on driver support for a year, thanks this time to broadband connection (bye bye soft modem drivers hell), with my customers.
I think also that much of changing on interface in latter MS offerings has to do with this catch up game, i.e., to make harder to its users to switch to alternatives.
Also, I would not compare Windows XP with Linux (on just OS aspect). I’m really getting tired of fixing Windows machines.
And never forget about development. Now java is more attractive than ever. Thanks a lot SUN.
So, yes, GNU/Linnux/OpenOffice/Firefox/Thunderbird are now a real threat to Microsoft kingdom.
-
hal2k1
//Even the speed difference is getting less noticeable, thanks to new hardware (I’m talking about Office).//
OpenOffice 2.04 has caught up with, and arguably beaten, Microsoft Office for speed.
This had nothing to do with Linux, nor hardware. OpenOffice was inefficient, and this has largely been fixed now.
-
2006-11-19 8:58 am
stestaggInterestingly, a lot of work on the recent speed improvements was done by the Novell OO team.
2006-11-18 7:58 pm
porcel
The ease of use of Linux is there. The apps for most people or a great deal of the computer-using population are there.
Now it is down to a war of words between people who have too much at stake in keeping the windows monopoly around because that’s all they know and that’s where their bread and butter is.
Freedom always triumphs. Sometimes, it takes 5 years, sometimes it takes 10, but it always does.
For those that talk about the Linux desktop and how it’s been around for ten years, I’d say stop rewriting history. The first serious desktop was KDE, a project which was started in late 1996/early 1997, one that now has thousands of apps. KDE didn’t release something which the average user could use reliably until KDE 3. That is the date for the beginning of the Linux desktop.
Even then considering it took Microsoft 25 years to deliver WinXP, which is still insecure and not all that user-friendly, I’d say the Linux desktop is doing just fine.
The refinements in the current KDE release and the upcoming ones are truly breathtaking. The amount of innovation and code written since the release of KDE 3 are also what has Microsoft and its apologists busy trying to rewrite history.
-
samad
“The apps for most people or a great deal of the computer-using population are there.”
Personally I feel if open source had its killer app to woo over a significant segment of the computer industry, then Linux would begin to enter more significantly into the mainstream commercial world (and I don’t mean as servers). For example there aren’t any decent open source replacements for Adobe products. If there was an open source project that was superior to Photoshop, Illustrator, or InDesign, there would be a lot more interest in Linux amongst graphic designers and desktop publishers. A single license of Adobe’s Creative suite goes for something like $1000USD. If there was an open source alternative, think of all the organizations switching over to save those kinds of software costs.
2006-11-18 8:22 pm
dikrek
OSX (which is really BSD with a great GUI) and Windows (especially now with vista) offer a ton of cool usability features, adequate performance, support and future-proofing.
Moreover, a lot of the more important free software runs on pretty much any platform.
I was doing a stint as a speaker once at Veritas Vision. Novell’s CEO gave a speech and then asked what Linux would need to be successful.
I told him “games and Visio”. People laughed but think what most people use their PCs for (not necessarily what OSAlert readers do):
1. Common stuff like surfing the web, email, word processing, spreadsheets etc
2. Games
3. Specialty apps such as photo editing and music creation
4. Interoperability with what they use at work.
Linux can provide #1 but severely lags for the rest.
What would help Linux tremendously is if Microsoft wrote Office for it (ALL of Office), and Adobe compiled all their stuff for it. The rest would then follow.
However, see it from the point of Adobe, for instance:
They have to write for an OS that has a standard kernel but does NOT have anything else as a standard! Not a GUI, nor sound, even underlying graphics architecture. They’d have to pick Gnome or KDE and a couple of compatible distros, then half the Linux community would be up in arms while the rest would tell everyone how they were right all along.
That’s just not a viable business model.
I could go on and on but the fact of the matter is that Linux ain’t even close at becoming a desktop standard.
Why not start the same agrument about Solaris 10? At least it’s a single unified distro, performs fantastically and has full support from one of the biggest and most respected computer companies in the world.
All Linux really is is the kernel. Everything else is distro-specific.
D
Browser: BlackBerry7290/4.1.0 Profile/MIDP-2.0 Configuration/CLDC-1.1 VendorID/102
-
2006-11-18 10:18 pm
LB06I think you’d be surprised how many people only need #1 and to some extent #4 (which means reading .doc and .pdf).
-
Moochman
Hmm…. I think it would actually be a LOT easier to make universal Linux apps than you think. Firefox can do it. I recently heard that Adobe is planning on using Eclipse as a development environment, maybe that says something about their plans….
-
2006-11-19 2:56 am
kaiwaiOh bullchunder; Adobe simply can say, “GNOME, either on Ubuntu or Fedora Core” and with that you would hit the majority of users without any problems.
Adobe doesn’t want to do it because they’re Microsoft’s bitch, just like HP is Microsoft bitch; they sit around claiming to be a victim but they do everything possible to prop up Microsoft monopoly through their inaction.
Adobe could make a tonne of cash selling their wares on Linux or some other operating system, but the problem is, its not that they can’t, its because they don’t want to.
What is the alternative? pray that maybe a group of coders will pull their finger out of their ass and create a clone of Adobe products; setup two computers and re-create the whole software stack, icon for icon, menu layout for menu layout, feature for feature.
-
hal2k1
//I could go on and on but the fact of the matter is that Linux ain’t even close at becoming a desktop standard. //
http://en.wikipedia.org/wiki/Linux_Standard_Base
With these two tools, your entire argument is moot.
Recently, OpenOffice released version 2.04 as a set of RPM packages only. Poeple said “what about Debian”?
Debian is compliant with the LSB and using the Smart package manager (works on every distro), one can install from the following types of repository:
* APT-DEB Repository
* APT-RPM Repository
* DPKG Installed Packages
* Mirror Information
* Red Carpet Channel
* RPM Directory
* RPM Header List
* RPM MetaData (YUM)
* RPM Installed Packages
* Slackware Repository
* Slackware Installed Packages
* URPMI Repository
Since Smart works on all distros, and Smart supports an “RPM Directory”, that means that OpenOffice 2.04 can be installed on any distribution from the set of RPM files that were provided on the OpenOffice.org website.
//All Linux really is is the kernel. Everything else is distro-specific. //
Not correct. Your comment is out-of-date. 2004 called and it wants its FUD back.
-
jbauer
Since Smart works on all distros, and Smart supports an “RPM Directory”, that means that OpenOffice 2.04 can be installed on any distribution from the set of RPM files that were provided on the OpenOffice.org website.
Really? Then the FAQ is wrong:
Does Smart’s wide support for packaging systems mean I can install .rpm packages on my Debian system? Or .deb packages on my RPM-based system?
Short answer: no.
Long answer: no, unless you have the non-native package system installed in the same machine. Even then, that doesn’t mean it would be a good idea.
Smart is not meant as an universal wrapper around different package formats. It does support RPM, DEB and Slackware packages on a single system, but won’t permit relationships among different package managers. While cross-packaging system dependencies could be enabled easily, the packaging policies simply do not exist today.
This is not at all different from what you can already do. In fact, Debian has been shipping the RPM package manager for a few years now. “Possible” does not equal “good idea”, and everybody should stick to their native package format.
http://labix.org/smart/faq#head-15e431ff69330db0ad3f078e2139c58ba7b…
2006-11-18 8:34 pm
acobar
The only thing I forgot to mention in my last post is were Linux is really lagging far behind: games. This is the real wall to be targeted if someone would like to see Linux more pervasive on homes.
Well, I don’t give a flying duck about games, but most of the people do. That is another reason Microsoft is pushing so hard on their OS graphics capabilities (besides the huge market for games).
Anyway, if the game console war keep its pace, people will have no alternatives but to have a computer and a console (what most that I know already do), and it will be less a problem.
2006-11-18 9:00 pm
TechGeek
It doesnt really matter what the US and Western Europe do. We are what, 1 Billion people strong? The 5 billion other people on the planet are pretty much in the poverty zone. You cant expect them to pay $400 for an OS when they make less that $4000 per year.
Not only that, but Windows is NOT going to be the OS of the future. Even MS has said that the future of computing will be online services. That includes the OS. I think Vista will be the last monolithic home OS that MS builds. And lets face, Linux is already an efficient client OS. The future of the home computer will be things like the PS3 or the Wii. How often do those devices get viruses or malware? And when they do, you can reboot them from a CD/DVD or from a file share at your ISP. and You are back up and running, no muss no fuss. Only IT people and hobbyists will have big open ended systems sitting at home.
-
2006-11-18 9:56 pm
camo rBlogger please, those people in the poverty range do what we do, PIRATE. And that speaks volumes, use a very capable, highly reliable free os or put up with hacks and cracks and use a somewhat reliable, kinda insecure os.
View this whichever way you want, windows has the pc by the proverbial shirt and curlies. If people are willing to use windows in all it’s flavors despite it’s flaws, then i guess the customers have spoken.
Yeah yeah, people don’t know about linux, blah blah. Tired of hearing that. If you want a product to be seen, you develop an identity and sell that baby for all it’s worth. Ubuntu is trying to do just that, and i guess it’s kinda working.
-
Moochman
>Even MS has said that the future of computing will be online services. That includes the OS.
Eh, actually I think this is where MS misses the point and the argument for Linux becomes even stronger. If MS actually does transition all of its software to a services-only, pay-until-you-die model, Linux will win by virtue of its being free, not because it somehow provides a better service-computing model.
>The future of the home computer will be things like the PS3 or the Wii. How often do those devices get viruses or malware?
Um, “these devices” just came to market, so it’s hard to say, isn’t it. We certainly can’t base expectations on previous models, though, since previous models didn’t support online connections and installable software to nearly the extent of the current generation. Aside from that, what makes you think that current consoles are the future of the home PC? If that were true, then no one would bother buying PCs anymore. The truth is, people like a certain amount of control and customization over the software on their computers. The form factor and price point of the home PC hardware may change, but the vulnerability factor that comes from having an open software platform (in contrast to the limited platforms the consoles provide) will not change.
2006-11-18 9:46 pm
Obscurus
Another thing holding Linux back is the general lack of support for anything proprietary. For better or worse, .mp3 is the standard format for compressed audio, with .wma a close second. .ogg is supported by only a handful of portable devices. For this reason, any Linux distro that wants to play with the big boys needs to support .mp3 in a seamless, automated way. The user should not have to do anything more than click “OK” on a dialogue to download and install the codec when they double click on an .mp3 file. At most.
There are tons of other proprietary file formats that need this kind of seamless support.
Also, while most Linux users at the moment are the sort of people who like to spend time choosing distros & packages from hundreds of available options, the majority of computer user do not like to have to make a lot of decisions. If you want to lure Windows users to a differnt product, it needs to be percieved as dramatically superior in the ways that count to the end user. It doesn’t matter if Linux Distro X has a better file system or a some other “technically” better xyz, it only matters to the end user that their experience is intuitive, simple, easy and uncomplicated. Most people do not care about software licences or philosophical principles, they just want to get their work done.
For most people, an OS is simply an interface from which to launch and manage applications, it is not a toy in its own right. To become successful, Linux needs to be so simple to use that a mildly retarded person can use it, it needs to standardise the interface so that people can recognise Linux when they see it in operation, and Linux must put all of the politics and philosophy aside if it wants to achieve double digits in the desktop OS market. Otherwise, another BSD derivative may well rapidly overtake Linux as the Desktop *nix of choice for many people.
All of the features that make Linux great at what it is make it a bad choice for a desktop OS – a desktop OS needs to be idiot proof, have limited customisability, come with few preinstalled applications, and have no hang ups about closed or open source (among other things).
-
2006-11-18 10:24 pm
lord_robAll of the features that make Linux great at what it is make it a bad choice for a desktop OS – a desktop OS needs to be idiot proof, have limited customisability, come with few preinstalled applications, and have no hang ups about closed or open source (among other things).
When I see all the mess that many different users do to their Windows (spywares, malwares, …), I wouldn’t say that Windows is idiot proof either. They can say that Windows is easier than Linux, but actually this is false. Many users *think* they understand Windows, but many times they need their tech-savy friend to solve all the problems they caused.
-
Moochman
It’s worth noting that Windows has a bare minimum of codecs installed by default, as well. This won’t change in Vista either. Truth is, installing all the codecs needed is actually a bit easier on Ubuntu, since you can just run a script to do it for you, whereas on Windows you have to deal with finding and manually installing codecs and complete software packages (like Quicktime) from a variety of different websites. I should know–I *just* did this on a new Windows PC, and it wasn’t too fun. Sure, there are sites with codec packages pre-made, and stuff like Quicktime Alternative and Real Alternative if you don’t want to clog your system up with giant junk-heaps of background tasks, but good luck figuring out which packages you need to download, whether they overlap with each other, and whether you can get them all to run in your media player of choice!
Edited 2006-11-19 00:42
2006-11-18 10:05 pm
Southern.Pride
For me Linux was a very difficult start (beginning with RedHat 6.0 – Caldera Linux 2.3 if I remember correctly back in 1999. The boxes looked really nice with excellent looking desktop displays. Once Windows was wiped out and the install process was started the frustration began, finally after the final product was up and running loss of functionality was clearly present. After several years of different installs and new versions I settled on RedHat permanently. I am glad now because at work the Open-Source movement began about 5-6 years ago small but took off quickly. Now I run Fedora Core 5 on my laptop & workstation, plus I got my RHCT 2 years ago.
In summary I would say Linux on the desktop alienated a lot of users on the front end, now with the multi-media support and nice desktop managers to choose from it is ready now more than ever.
Ubuntu may be my next distro to switch over to, since everyone at work is running it now. Lastly, if the hardware support improves like it has been doing the only place to go is up!
2006-11-18 11:56 pm
JeffS
The person siting 1/2 of 1% market share is way off base. Those surveys are very unreliable.
In fact, Gartner’s stats in 2004 said Linux had 5% market share, and had exceeded MacOSX.
That’s quite a difference. Is either one right? Probably neither one. Those surveys are very unscientific, and have very large margins error.
Another important factor is that those surveys only count units shipped (sold), either as a boxed, over the counter version, or a pre-isntall (like 99.999999% of all Windows installations).
Well, Linux is sold in boxes over the counter, and Linux is pre-installed in some PC’s. But that is an extreme miniscule amount of the over all usage of Linux. The fact is, the very vast majority of Linux usage comes from free downloads, or as free discs in magazines or books, or CDs passed out at LUG meetings. That is stuff that the Gartners of the world are completely incapable of tracking.
Thus, those market share surveys don’t even come close to estimating overall Linux usage.
And to others saying that Linux doesn’t work, or it’s too hard – I say POPPYCOCK!! All the Linux desktop oriented distros are so easy to install and use it’s pathetic. You can literally get say, PCLinuxOS (only one example of many super easy distros), installed in 10 minutes, with all hardware working, all media codecs working, flash and Java working, games, Office suite, multi-media programs, programming tools, and graphics programs, all working 100% out of the box. No fuss, no muss. That is an irrefutable fact, except for in some very rare cases.
This is much much much much easier than Windows, by a long long long margin. Professionally, I’ve done so many Windows installs (of all versions and designations and languages), that it’s impossible to count. I’ve also done many many many Linux installs. So I know exactly what I’m talking about.
I’m absolutely not saying all this as a “Linux zealot”, or geek, or hyper idealist, or whatever. I’m saying all this from a completely pragmatic point of view, from someone who simply likes great technology, from whatever source.
Mr. Shuttleworth is exactly right. Linux will be a major desktop player. It’s only a matter of time. The only things holding it back is end user inertia and constant Microsoft FUD.
Edited 2006-11-18 23:59
2006-11-19 2:40 am
kaiwai
He has some good things to say, but he forgets one important thing; a win for *BSD/OpenSolaris is still a win for Linux – atleast they’ve moved alway from Microsoft, which is the more important goal.
By having a more fractured marketplace what it does force companies into doing is adopting openstandards rather than trying to do a Microsoft and create things on the fly – or atleast create good ideas, but fail to get the industry involved by creating an impartial pannel to promote those technologies in an openstandards way (XPS, XAML and Winforms and others come to mind).
OpenSolaris is making HUGE progress; if you look at the change logs for each build, build 53 being the latest, and the changes have been emense; FreeBSD are making huge scalability improvements, and the desktop/user layer such as GNOME and KDE as making progress.
*NIX won’t take over the desktop today, but what I will find humorous is when opensource applications are so good, the excuse of ‘xyz is not available’ will be a thing of the past when opensource will be providing superior applications.
One that comes to mind which is superior is Amarok; superior to iTunes/Mediaplayer in everyway; supports MTP and iPod out of the box, the KDE environment is more heavily integrated than Windows – global spell checker for instance, which is accessible through any KDE based application.
2006-11-19 10:30 am
AkiFoblesia
let’s see… linux proponents say linux will dominate the OS market
Windows was out…
Windows camp got really shut quiet
But wait…
the other camp seems to create so much noise…
are they celebrating?
let’s take a look…
oh no! more wars…
they say: “my own distro is the best!” “i’m creating a new distro based on ubuntu which will outperform all other distro”…
-
-
AkiFoblesia
oh yeah.
how free. he can choose one distro to use at home, then would not know how to use everything else elsewhere…
nice freedom
2006-11-19 4:50 pm
buff
I think Shuttleworth is right that people mostly use Windows still since it is what people are used to and change is difficult.
There is the reality though that desktop linux is not quite there for users that need some Windows applications. For example Flash, which I need for my school’s online learning, is always one version behind on linux causing me to switch to Windows. Also I attempted to run Turbotax in linux using Wine and it wouldn’t work right. I thought about buying crossover office but it was just easier to boot up Windows XP. When I don’t have to reboot in Windows XP mode to get all my work done then I will know desktop linux is ready. Very close but no cigar.
Edited 2006-11-19 16:51
2006-11-19 5:36 pm
Lambda
I’ve been using Linux professionally since ’97 (back in the old FVWM days) and people have been saying desktop Linux is going to make a breakout on the desktop since around 2000. That hasn’t happened, and it’s not even close to happening.
The reality is that handwaving about “Linux” is stupid. What Linux? Gnome, KDE, what distro? How is this supposed to happen? Shuttleworth doesn’t say anything about that.
“Linux” is just too fragmented. Either there should have been a desktop reference – which is being tried, there should have been the canonical distribution that a desktop distribution should be compatible with in order to be a “desktop linux”.
But the other problem is that Gnome and KDE are trying too hard to follow in the footsteps of Windows and Mac. They follow the same static model. So forget about KDE and Gnome. They’re going nowhere.
If I had Shuttleworth money, and really cared about the desktop, I would want to differentiate myself and invest the money into something from the groundup. Think live environments like you see in Smalltalk or Lisp – coupled with a new toolkit that depends on hardware-accelerated OpenGl. I would also develop something like Applescript that is very english-like, so that the average user can write little scriptlets.
People are fooling themselves if they think business as usual, with everybody just repackaging the same old stuff, is going to change anything. It’s almost 2007, and desktop linux is still “just around the corner”.
-
2006-11-19 10:36 pm
diegoviolaThere’s no need to for invent the wheel, X.Org has AIGLX, Composite, Damage, Fixes, NVIDIA has GLX_EXT_texture_from_pixmap, there is also Qt 4 which is a excellent widget library and can be accelerated with OpenGL for compositing, GTK+ uses Cairo and can be also accelerated with Glitz, the framework is there, and has been for some time already, haven’t you seen Beryl? that window manager uses Composite and GLX for all the Compositing operations, KDE 4 will also implement some compositing effects in Kwin, I don’t understand your point there. Stop trolling.
Edited 2006-11-19 22:37
-
Wrawrat
I believe his point wasn’t about the looks, but rather the functionalities. Apart for the development model, the leading DEs are not fundementally different from the leading desktop OSes. Even though they share different features, they are mostly offering a similar experience because they are using the same paradigms. Adding layers of bling-bling won’t change this, no matter the amount.
I don’t see this as a troll, but rather as a call for being more daring by stepping out from the main path. Stuff like Mezzo (coming with Symphony OS) is a good start, but it could go further. The average person might not be interested in learning a new system that does his tasks the same way, but they might be interested in something that tries to solve his problems differently. That’s how the free software desktop could stand out, IMO.
-
Lambda
I believe his point wasn’t about the looks, but rather the functionalities. Apart for the development model, the leading DEs are not fundementally different from the leading desktop OSes. Even though they share different features, they are mostly offering a similar experience because they are using the same paradigms. Adding layers of bling-bling won’t change this, no matter the amount.
Exactly. Over the years, I’ve been constantly surprised at the conservative, conformist attitude of open source desktop developers. There was the ill-fated BlueEyedOS project, which was to put BeOS on top of the linux kernel. But besides that, I haven’t seen much in doing something different.
You mentioned Mezzo (Symphony OS). That’s a fresh approach, but I’m talking about starting from the botton-up and not the top-down. I would look at the kernel as a POSIX subsystem, and Xorg as my graphic subsystem and go from there. Basically, everything on the desktop would be live objects. There really wouldn’t be the need for preferences because ever desktop object could be inspected and changed. Desktop objects could be combined, embedded, ripped apart, etc.. All desktop objects would be scriptable at runtime. I’ll probably write up a real submission for OSAlert so I can flesh out some of these ideas in more detail.
On the other side of the spectrum are the KDE and Gnome desktops. I’ve talked to some prominent desktop developers over the years in private candor discussions, and most have agreed that from a marketing and penetration standpoint, if things could be done over, it would have been wise to have collaborated on “the” desktop, “the” toolkit, “the” interop system from the beginning. Everything else would be pretty much on the margins like Enlightenment. But that didn’t happen, and is problematic from an ISV perspective.
Now as long as OSX and Windows remain closed-source, there will always be a market for a free, open source desktop, so it’s not like linux desktop will go into irrelvance. It’s just that it might be another decade or so, and given that, if there were resources to be had, it couldn’t hurt to do something different. But unlike 10 years ago, there’s a herculean amount of work to be done for something totally different, and unless you have philanthropist deep pockets backing you, or have a sound plan for a ROI, it’s a tough uphill battle – but not undoable.
-
2006-11-20 12:18 pm
h3rmanI think you might be right that there’s a great deal of conservatism in software development, and you’re talking specifically of the DE’s GUI. The thing is just, humans are generally conservative, and GUI software is very complicated, and increasingly so. On top of that, you have users that want things to work in a way they can understand with a minimum of effort. If *n*x DEs/GUIs were too much unlike anything else, *n*x would get comments like, “sure you ain’t got no market share, you gotta make it like those other DEs.”
Maybe this will always be much more about evolution than about revolution. The first thing brought us from the apes to where we are now (well… I’ll not be glorifying our present state of affairs), whereas most we know of revolutions is they “eat their own children.”
I also disagree with you that the existence of both Gnome and KDE side by side is somehow a “problem”. It’s more a blessing. Imagine one monster project that decides how the major *n*x DE works, looks and feels, how many criticism that would arouse, how many forks and competing projects would come out. You’d have a worse situation than now. Say, I don’t like Gnome, there’s no one that’s gonna listen to my whining over it. Just use KDE and shut up, they’d say.
As for Linux desktop world domination, well, what goes up must come down. But the impatient will always be disappointed. I personally don’t understand all the fixation on market share. Rolls Royce never cared for that. Just focus on quality and the rest will follow.
-
Lambda
I think you might be right that there’s a great deal of conservatism in software development, and you’re talking specifically of the DE’s GUI. The thing is just, humans are generally conservative, and GUI software is very complicated, and increasingly so. On top of that, you have users that want things to work in a way they can understand with a minimum of effort. If *n*x DEs/GUIs were too much unlike anything else, *n*x would get comments like, “sure you ain’t got no market share, you gotta make it like those other DEs.”
The novice users *would* work in the same way. There would really be no paradigm shift for them. This is really about power users and those between power users and the novice. Think about html and how it brought the ability the average guy to create content. But I really need to write up a real story for OSAlert.
Maybe this will always be much more about evolution than about revolution. The first thing brought us from the apes to where we are now (well… I’ll not be glorifying our present state of affairs), whereas most we know of revolutions is they “eat their own children.”
The problem with going the evolutionary route in the open source desktop space is that Microsoft and Apple don’t sit on their hands and wait for desktop Linux to catch-up. So I believe that will always put desktop Linux in a bad position in comparison to them.
I also disagree with you that the existence of both Gnome and KDE side by side is somehow a “problem”. It’s more a blessing. Imagine one monster project that decides how the major *n*x DE works, looks and feels, how many criticism that would arouse, how many forks and competing projects would come out. You’d have a worse situation than now. Say, I don’t like Gnome, there’s no one that’s gonna listen to my whining over it. Just use KDE and shut up, they’d say.
This is a misunderstanding that is repeated over and over, under the guise of “choice”. There would always be alternatives. It’s just that if you go the conservative route, then it makes sense to collaborate in order to present a unified front to ISVs. ISVs have a hard time targetting desktop linux because essentially they have to stick with a subset of what is going to be on the average desktop. For example, Kparts is a powerful KDE component, but an ISV can’t use that unless they want to restrict their market.
As for Linux desktop world domination, well, what goes up must come down. But the impatient will always be disappointed. I personally don’t understand all the fixation on market share. Rolls Royce never cared for that. Just focus on quality and the rest will follow.
Microsoft potentially “coming down” decades from now doesn’t mean that KDE or Gnome is going to “go up”. I don’t care about market share. I use Linux as an engineering workstation. I don’t hate Microsoft. I tend to run both windows and linux at the same time. It’s the Microsoft haters that have the fixation with Microsoft.
My point wasn’t even really about market share, but doing something creative and different, instead of just following the lead of Microsoft and Apple.
-
hal2k1
//For example, Kparts is a powerful KDE component, but an ISV can’t use that unless they want to restrict their market. //
I don’t see why (although I don’t use GNOME so I could be wrong here). AFAIK, if a GNOME of Xfce user wants to install an application with a dependency on Kparts, then the package manager will download the relevant kdelibs and whatever other packages are required in order to get that application installed.
-
netpython
This is a misunderstanding that is repeated over and over, under the guise of “choice”. There would always be alternatives. It’s just that if you go the conservative route, then it makes sense to collaborate in order to present a unified front to ISVs. ISVs have a hard time targetting desktop linux because essentially they have to stick with a subset of what is going to be on the average desktop. For example, Kparts is a powerful KDE component, but an ISV can’t use that unless they want to restrict their market.
It’s not that there’s to much to choose from,choice is good.The ISV’s just don’t smell the money yet.Which DE is being used is irrelevant if there was a more unified front of portable code.
Edited 2006-11-21 09:12
-
Lambda
It’s not that there’s to much to choose from,choice is good.The ISV’s just don’t smell the money yet.Which DE is being used is irrelevant if there was a more unified front of portable code.
I think you just re-iterated what I said, except you added the part about “show me the money”. But you have to lower the barriers to entry when the market is small.
-
archiesteel
You make a valid point, however you have to understand that radical new ideas, even if they are real improvements, would not necessarily mean a larger share fo the desktop.
There’s a reason why it’s so hard to break away from the WIMP (Windows, Icons, Mouse, Pointer) paradigm. The reason is that it works. Why do you think books, hammers, televisions have changes so little over the years? That’s because they have reached their functional limits, and there’s nothing wrong with that.
So before wanting to reinvent the wheel and claim it is necessary to ensure success in the marketplace, it’s a good thing to question whether that is the real issue at hand.
-
Lambda
You make a valid point, however you have to understand that radical new ideas, even if they are real improvements, would not necessarily mean a larger share fo the desktop.
The thing that you have to understand, is that it’s almost 2007 and really nothing has changed. But most of the problems with desktop linux are social issues of collaboration, and not necessarily technical problems. There was never a unified front to market. The cost and problems of Windows hasn’t turned out to be enough to garner a significant shift.
Personally, I don’t really care about market share. I’d just like other ideas to be explored and new operating systems to be developed for Linux, or at the very least, new desktop paradigms.
There’s a reason why it’s so hard to break away from the WIMP (Windows, Icons, Mouse, Pointer) paradigm. The reason is that it works. Why do you think books, hammers, televisions have changes so little over the years? That’s because they have reached their functional limits, and there’s nothing wrong with that.
Nothing that I would propose would change the 2d WIMP paradigm. You really can’t with our 2d screens. This is more of an under-the-hood change; a way for medium-skilled and advanced desktop users to make the most effective use of the components of their system.
So before wanting to reinvent the wheel and claim it is necessary to ensure success in the marketplace, it’s a good thing to question whether that is the real issue at hand.
The question I would be asking is why nothing really changing? Why are people still sticking with Windows. Firefox was/is a better browser. actually, the firefox extension paradigm is a great metaphor for what I’m talking about. IE has gotten better with regards to spyware and such, but its the plugins in Firefox that keep them coming back. The desktop I envision would basically be a big introspective, plugin system. But it’s hard to explain it all in a thread post. I hope to submit a story to OSAlert soon.
-
netpython
The thing that you have to understand, is that it’s almost 2007 and really nothing has changed. But most of the problems with desktop linux are social issues of collaboration, and not necessarily technical problems.
There aren’t necessary more problems than with any major OS system.Personally i’m quite happy with running linux wherever i can.As are a lot of other people.
Why are people still sticking with Windows.
Because that’s what they know.
-
Lambda
Let’s start with this gem
Stop trolling
Oh please, saying “trolling” has been reduced to “I don’t agree with what you say”. You people need to grow up.
There’s no need to for invent the wheel, X.Org has AIGLX, Composite, Damage, Fixes, NVIDIA has GLX_EXT_texture_from_pixmap, there is also Qt 4 which is a excellent widget library and can be accelerated with OpenGL for compositing, GTK+ uses Cairo and can be also accelerated with Glitz, the framework is there, and has been for some time already, haven’t you seen Beryl? that window manager uses Composite and GLX for all the Compositing operations, KDE 4 will also implement some compositing effects in Kwin, I don’t understand your point there
No, when your happy with the status quo then you wouldn’t understand.
I wasn’t suggesting to reinvent Xorg or any low-level driver code, the point was to start off with a toolkit that is not stuck in the static world of C/C++, which is what you have with gtk+, Qt.
For a more concrete example, think of Squeak http://www.squeak.org, but something more aesthetically pleasing, and probably not written in Smalltalk. But I’ll explain more to the guy that responded to you, so look below.
2006-11-19 5:52 pm
2006-11-19 9:54 pm
h3rman
Some people talk about the weather,
some people talk about Hollywood,
some people talk about the piles,
some people even talk about work,
but it seems, there’s nothing like good old “Linux vs. Windows.”
2006-11-20 2:12 pm
fireball
My opinion regarding this topic, is that Linux is not The Only alternative for a FOSS operating system.
In future, and right now, there should be a choice of OS for the actual goal you need to achieve. If your goal is having a server, then linux is the choice. If your goal is having a desktop PC, then one must think.
And think about other operating systems, which are also free and opensource.
E.g. ReactOS targets exactly this area – desktop usage, and it already has a solution for the most important 2 problems:
1) Drivers – all hardware manufacturers provide drivers “for Windows” which can be used in ReactOS
2) Applications – number of available Win32 applications is a lot wider than the amount of non-win32 apps (and yes, there exist foss win32 apps too – what a surprise!)
If people would not be that blind to not see anything except linux, ReactOS would evolve much faster than it is evolving now.
-
2006-11-20 7:06 pm
DeadFishManThe problem with that is that ReactOS is way too Windows-like and it may inherit some of its flaws. Please note that I said “may” as it seems that its developers want to put as much compatibility as possible in place whilst doing something different with the desktop (virtual desktops, for starters).
This is nice if Windows is what you want, but then I'd have to give up some stuff that I've grown used to in Linux such as Konqueror, Amarok, bash, REAL scriptability with DCOP (and later with D-Bus). It's awesome to be capable of access my home machine through SSH in order to avoid a corporate firewall, start GNU Screen and access my favorite IRC channels through BitchX, use giFTcurs to download some stuff and stuff like that like I'm doing as I write this.
Windows cannot offer me such flexibility currently and it remains to be seen whether ReactOS will or not. My point is, if a Windows experience is what you're looking for, then more power to the ReactOS project. However, it will be always playing catchup to the real thing so one would be better using Windows that he/she probably already paid for.
I don't use Linux because I hate Windows (though Explorer surely could learn a thing or two with KDE); I use Linux because I love UNIX and I managed to adapt comfortably to the new environment. I'd hate to let it go.
-
2006-11-20 7:49 pm
frik85The problem with that is that ReactOS is way too Windows-like and it may inherit some of its flaws […]
That’s a assumption, ReactOS won’t suffer from GUI flaws as the latest incarnations of Win32 do. And the WinNT kernel is in comparision to all other modern mainstream operating systems (for workstation & server) a very good one. Some deprecated WinAPI functions may sound out-dated but they are still there (mainly in Win16) for compatibility and that’s especially where free open source Unix derivates suffer.
[…] accessing through ssh, bash, scriptibility […]
… everything is possible with open source applications and tools for ReactOS & Win32.
However, it will be always playing catchup to the real thing […]
It will take another 2-3 year until ReactOS hit version 1.0. The native NT API hasn’t changed a lot for years nor does the WinAPI. The only changes are GUI related, and additional services, tools and of course drivers. And that’s where ReactOS advantage come in, no other operationg system beside WinNT serie has access to such a mass of driver collection. Every piece of hardware comes with NT driver, and other operating system have to reverse engineere NT compatible driver to add such functionally to their OS. That task is not only questionable but also leave a wide range of unsupported hardware.
I use Linux because I love UNIX […]
Why would you use Linux when you love UNIX? Linux and Unix are not compatible in general. Both are follow the old posix standard, as do WinNT serie.
Trying to convert a broader mass of people to use Linux is IMHO a no brainer based on a hype promoted by big Linux distro vendors and some media companies.
Linux is great for web-server and embedded devices, but designed for desktop nor for super-computer.
Application and driver(-modul) compatibility is broken, x11 is by design badly outdated, etc. … the list is long.
You cannot assume that average joe compile his own binaries from source code nor use a complicated package manager software.
Staying with age-old x11 is silly, just silly. They add only another hack to add new features or fix something.
To be fair, when I use the term “linux”, i just use it as it is the most hyped unix alike OS and “DeadFishMan” use it too. Although some of the above mentioned flaws are guilty for *BSD and other Unix forks too.
Unix has been designed for single user single account (root). Everything else has been added as “hacks” (as it would be named everywhere else, except unix world).
Multics (yes, some used it until 2000) has been a lot more advanced back then, but even back then the “hype” and money phenomenon was already popular.
-
hal2k1
//You cannot assume that average joe compile his own binaries from source code nor use a complicated package manager software. //
You have GOT to be kidding.
From nothing at all, a package manager can download and install a new application in just a few clicks.
On Kubuntu, as an example to install a complete office suite:
Menu->System->Manage Packages
(enter your login password)
Click on the “search” button on the toolbar
Type “office” and then enter
Select your choice of OpenOffice or Koffice
Click the “Apply” button on the toolbar, and …
wait for the application to download & automatically install.
You are done.
That is several times easier for even a newbie to do than on Windows. On Windows, first you have to get an Office suite from somewhere, which involves either a search on the web or a trip to the local computer store to buy disks. Remember … those above are the steps for installing via a typical Linux Package Manager starting with nothing.
You don’t have to compile binaries for Linux. You are either FUDding like mad or you are hopelessly behind the times.
BTW, on Windows, even if you do manage to get through all the hurdles including CD keys & whatnot to install an office suite, apparently after a little while it will start nagging you to make phone calls & jump through yet more hoops and then stop working if there is a hitch with that.
http://www.osnews.com/comment.php?news_id=16533
Edited 2006-11-21 08:58
-
frik85
From nothing at all, a package manager can download and install a new application in just a few clicks.
ReactOS offer a package manager too. It is simply to use, it will get more features but will keep it’s simple interface. Usability is a key factor. Most GUI application designed and developer don’t realize that :S
I have written about linux package manager because they tend to be extra overloaded (e.g. Yast in SuSE til v. 9/10). Newer package manager tend to be a bit more simple to use but are less feature-rich.
[…] example to install a complete office suite […]
You are not limited on ReactOS & Win32 to MS published software. OpenOffice, AbiWord (afaik soon KOffice) from the open source fraction and a dozen of closed source ones are available.
Hundreds of thousands Win32/WinNT applications and drivers are available and everyone will find suitable solutions without compromisses.
That was my point with the comment before.
You are speaking of DRM (and similar technology), which is a bad movement and should not be extra supported. But it’s not part of a Win32 application in general. You can have the same kind of shit in unix world too. Look at most dvd players’ driver modules, for example.
-
netpython
Unix has been designed for single user single account (root).
Unix has been designed from day one to be a multi-user environment.Much unlike windows 95/98.windows 2000,xp home..
Everything else has been added as “hacks” (as it would be named everywhere else, except unix world).
I would call the professional version of the above a hack.
Linux is great for web-server and embedded devices, but designed for desktop nor for super-computer.
Do some googling and you will be suprized how many super computers in fact run on linux/unix.There isn’t a single one that runs on any windows says enough.
So another sign of you spreading FUD because the super computers cycles are quite expensive to be run by one single user only don’t you think?
-
frik85
Unix has been designed from day one to be a multi-user environment. Much unlike windows 95/98. windows 2000,xp home..
My comment is based on real facts, it was about Unix which originally developed in the 1960s.
Win1-3x & 4x (aka Win9x) are based on DOS, to sum-up it in a simply to understand way. It’s a single user OS for personal computer.
Although, I have never written about Win 1-4, I have written about the WinNT series and ReactOS. That are very different OS in comparision to Win 1-4.
Unix is the small brother of Multics, and has been by design more lightweight but also less secure, less feature, and single user, especially for cheap computers.
WinNT architecture is based on VMS (no OpenVMS) and some ideas came from OS/2. To sum-up WinNT could be called a VMS clone.
Most banks (and other companies which transfer or handle a lot of money) use and trust OpenVMS and not Unix alike operating systems for several reasons.
I would call the professional version of the above a hack.
you can call it how you want, that doesn’t matter.
Do some googling and you will be suprized how many super computers in fact run on linux/unix.There isn’t a single one that runs on any windows says enough.
The fact is no super-computer have been developed and built for years! Most famous “super computer” are the Cray-series (1970s to mid 90s). If you use the term “super computers” nowadays, you may mean “clusters” which are cheap standard computer (or in blades, what ever) connected to each other with a high-speed network.
Google & co. use such sub-optimal solutions, because they are cheap, but very electricity power intensive. That’s why e.g. Google even built there new server halls next to rivers power-plants to save money.
So another sign of you spreading FUD because the super computers cycles are quite expensive to be run by one single user only don’t you think?
You seem to be a kind of troll or a weird fanboy which cannot add 1+1 together or wear blinders., or someone that got his first computer in post millennium era.
-
netpython
And the WinNT kernel is in comparision to all other modern mainstream operating systems (for workstation & server) a very good one.
Yes it’s so good that when a badly written driver is loaded it takes the whole system with it.
-
netpython
http://en.wikipedia.org/wiki/Supercomputer#Operating_systems
windows isn’t mentioned.
If it is so good and multi-user than why not?
It doesn’t make sence to call a multi-user OS a hack.Especially when 7/10 next generation super computers use linux.
or someone that got his first computer in post millennium era.
I’m glad you asked.It was a c64 with a cassette player as HD.And yours?
-
2006-11-21 2:58 pm
OMRebel“I’m glad you asked.It was a c64 with a cassette player as HD.And yours?”
I had the exact same thing. In fact, I still have it, and hooked it up a couple months ago, and it still works fine.
Linux has been around over 10 years and has about 1/2 of 1% of the destop market.
Two things are holding Linux back: drivers and applications. Msft, with their enormous influence, will make sure Linux catches up.
BTW: I like Linux, I am linux certified, work with linux to earn my living, and I’m using linux right now. I’m not a basher, I’m just realistic.
Two things are holding Linux back: drivers and applications.
Actually I think also quality assurance needs more focus. Ubuntu particularly has been suffering in this department lately, which is a shame.
I guess testing all the work that has been done is rather boring, but a very necessary part of delivering a quality desktop. Feature glitz is the sexy thing which makes a distro popular, but I think you’d see people more likely sticking to Ubuntu if the apps you use aren’t full of embarrasing spelling errors, don’t crash half the time, ignores configuration changes and strange things like that, that makes it feel very rushed.
As for the desktop experience (ease of use particularly), Linux has a long way to go, but it’s steadily improving. That part is an old discussion and is more a design issue. Right now I want quality!
I haven’t noticed any of those, I have noticed *stupid* oversites; for example, in the network configuration, how come I can’t setup WPA without needing to drop down into shell, editing wpa_supplicant.conf? I shouldn’t need to do that as an end user, and yet I am.
Oh, and documentation; good lord, that definately needs fixing; the documentation is worse than MSDN, and that is *really* crap.
Not disagreeing with your principles, but you might be pleased to know that wpa_supplicant now comes with a GTK app — wpa-gui — that allows you to create network-blocks for wpa_supplicant.conf. (Maybe Ubuntu don’t use this, but have a check on the commandline in case it’s in there.)
As for documentation, I think there’s no shortage as such (except for certain apps which shall remain nameless) but it is so disparately organised. We have man, doc, info, plus numerous apps implementing their own builtin help systems. For each app you have about 3-4 places you might look to get the right help. A unified, Desktop-agnostic and textmode/GUI-mode transparent help specification would be a big help. Let the UNIX throwback stuff quietly die off, especially ‘info’ — I don’t bloody want to learn a highly counterintuitive set of key-functions a la Emacs just to scroll through some help-docs!
A unified, Desktop-agnostic and textmode/GUI-mode transparent help specification would be a big help. Let the UNIX throwback stuff quietly die off, especially ‘info’ — I don’t bloody want to learn a highly counterintuitive set of key-functions a la Emacs just to scroll through some help-docs!
Oh, please! Man pages are a great resource and have been that way for years. People can complain that most man pages are cryptic and hard to read but there are thousands of man pages out there for GUI apps that are easy enough to understand.
KDE already has a KIO-slave that translate man pages into nicely formatted pages on Konqueror. All that you have to to do is type man:/app or simply #app. You can't tell the difference between a man page and a web page seeing it this way. There you have it: your unified, Desktop-agnostic and textmode/GUI-mode transparent help specification that's still UNIXy in nature.
I'm getting tired of this Linux/UNIX bashing and really wish that people wanting Windows-like usage pattern go back to using Windows and leave Linux to people that enjoy using it and/or is willing to learn how to use it properly.
I let networkmanager handle my WPA. Haven’t touched the terminal yet.
And yeah documentation is really bad, I’m mostly using IBM’s aix doc, qt’s doc etc
“1/2 of 1%”
In North America. Wake up America, realise there’s an entire world out there bigger than you and places where Microsoft don’t exist, and don’t matter. This constant obssession with the American desktop is what’s holding you back.
Well, I’m willing to bet a lot of money that market share in the rest of the world is very close to that too.
I live in Europe and even if twice as many Europeans are Linux users compared to Americans, that’s still a too small user base.
In North America. Wake up America, realise there’s an entire world out there bigger than you and places where Microsoft don’t exist, and don’t matter. This constant obssession with the American desktop is what’s holding you back.
Linux = .4%
European Stats: http://www.webhits.de/webhits/browser.htm
Win95 has double the market share of Linux.
WinME has quadruple the share of Linux.
Edited 2006-11-18 20:11
and
http://www.w3schools.com/browsers/browsers_stats.asp
linux 3.5% …
What’s your point besides showing off statistics?
I like the quote at the end of the page:
“There are three kinds of lies: lies, damn lies, and statistics.”
Benjamin Disraeli
Still to low.
From a business perspective that number is high enough to get a couple programmers rewrite the non-portable code.
I mean as a business you’re missing out 3% off your revenue stream. That’s A LOT!
But I don’t see companies port their applications any time soon simply because they probably never build their application in platform independant manor.(which is simply retarded)
btw there’s no extra cost to write an application to still be able to easily port it later, just don’t hire retards
From a business perspective that number is high enough to get a couple programmers rewrite the non-portable code.
I mean as a business you’re missing out 3% off your revenue stream. That’s A LOT!
But I don’t see companies port their applications any time soon simply because they probably never build their application in platform independant manor.(which is simply retarded)
btw there’s no extra cost to write an application to still be able to easily port it later, just don’t hire retards
How high revenue do you expect from a community that don’t want to use any software that’s not free, end even it’s free, if the source is not given, they still don’t want to use it, and also if you give the source, but it’s not GPL, you will be bashed?
Edited 2006-11-19 10:33
If that was true then why’s photoshop the top most wanted app on linux and not gimp?
People will pay if it’s worth the money.
Even on the windows market you got freeloaders (see cracks, warez etc) and they aren’t the minority :/
I mean as a business you’re missing out 3% off your revenue stream. That’s A LOT!
Depends on what the cost of the rewrite is and the potential profits. 3% can be anything from ‘A LOT’ to a lot less than zero when it comes to profits, which is what really matters.
btw there’s no extra cost to write an application to still be able to easily port it later, just don’t hire retards
Hiring people who can write portable code is an extra cost compared to hiring ‘retards’. Also writing portable code greatly limits you in your choice of libraries and increases the amount of necessary testing, both of which can drive up costs.
Depends on what the cost of the rewrite is and the potential profits. 3% can be anything from ‘A LOT’ to a lot less than zero when it comes to profits, which is what really matters.
True, depends on the balance between your costs and how many people are willing to buy the product. But if your product is desired enough that 3% will be huge.
Hiring people who can write portable code is an extra cost compared to hiring ‘retards’. Also writing portable code greatly limits you in your choice of libraries and increases the amount of necessary testing, both of which can drive up costs.
That’s only true for small projects, when you’re dealing with huge projects then you need to hire people who can do more then just write lines of code.
And it shouldn’t limit the use of your choice of libraries , there will always be non-portable code.
You simply don’t need to mix non-portable with portable code. To better explain myself a short example:
When writing a statistics application then you can easily seperate the GUI(non-portable) and “the datastructures and algorithm’s”. This approach also makes it easier to debug, extend the code etc
If your app is a succes in the market place then you can always hire a couple programmers to fill the gaps with equivalent libraries on the other platform.
“Hiring people who can write portable code is an extra cost compared to hiring ‘retards’. Also writing portable code greatly limits you in your choice of libraries and increases the amount of necessary testing, both of which can drive up costs.”
Writing good, clear portable code actually decreases costs, because maintenance, bug fixes, and upgrades become much easier. Quality doesn’t cost, it saves. “Limiting” libraries simply means using standard libraries that won’t be changing at the drop of a hat, and that don’t require special tricks, again decreasing costs.
Rewriting hacked up, poor code into something that is portable, efficient, and clean is a cost up front, but writing it correctly the first time is a cost saver. I suspect that Adobe has some legacy trash to worry about.
I think exist the truth, the uncomfortable truth and stadistics.
W3C schools is more interesting for its trends because, as they say, it attracts a specialized audience.
For example, Linux usage actually dropped after June of 2005 and has never been higher than that usage 1.5 years ago.
That is evidence that Linux usage has peaked and isn’t growing.
Other sites (with a much, much bigger audience) confirm the .4% for Linux.
Edited 2006-11-18 20:51
From the same w3schools.com web page you *misrepresented* in your last post….
“First get your facts; then you can distort them at your leisure.”
Mark Twain
Other sites (with a much, much bigger audience) confirm the .4% for Linux.
Web site access statistics are exactly that: statistics about configurations visiting web sites.
They are in no way reliable instruments to gauge the installed base of any component of those configurations, not even the webbrowser.
To make such statistics even remotely accurate would require to know that all visitors are using their private machine and not some computer at their workplace, internet cafes or public surf stations.
and
http://www.w3schools.com/browsers/browsers_stats.asp
linux 3.5% …
w3schools is of interest to geeks, not general public. Just like, for example, OSAlert.com. It is expected to see more Linux users there.
If you want to be realistic, then look at some site that gets visited by all types of users. MySpace, Google, CNN, BBC, NBA, Yahoo.. would all be better examples. You will not have more than 1% of Linux users on those sites.
For example, I don’t see my father visiting w3schools any time soon, yet I know he uses Google News on a regular basis.
true! it’s just a mention how wrong statistics can be, I never mentioned those ones were right!
What’s the methodology? If it’s gathered from sites that have a more Windows focus or depend heavily on Windows Technologies, of course they’ll show Windows is more popular. I’d trust Google’s Browser statistics a bit more. I don’t have the current stats, but in 2004, it was pegged at 1%, which is pretty respectable considering that most of the *real* desktop work has only happened in the last few years:
http://www.google.com/press/zeitgeist/zeitgeist-jun04.html
W3Schools and Mihaiu (don’t know anything about them) has Linux pegged at 3.5% and 6.5%. Both are bit high for my expectations, but it does go to show that there are places in the world where Linux is more popular than others.
http://www.w3schools.com/browsers/browsers_stats.asp
http://www.mihaiu.name/operating-system-search-engine-browser-stati…
In my city, I’d peg it at 2%. Linux is pretty popular here, especially at University. We even have a Linux Cafe and all computer and electronics stores seem to at least heard about Linux and have someone on staff that can handle Linux related questions. Most people, however, just use whatever’s in front of them, and in most cases, it’s Windows.
What’s the methodology?
Lets see. Google had Linux at 1% (probably less but they rounded up to take pity on the Linux base) in June 2004.
W3Cschools had Linux at 2.9% in the same month.
Because Google is so ubiquitious, and more likely to favor Linux over people using MSN search, my methodology says W3C schools overrepresents Linux usage by 3x.
On W3C Schools Linux grew from 2.9 to 3.5.
My methodology would conclude Linux usage would be 1.2% max on google today.
Since the deals OSS companies like Firefox make with Google would overrepresent Linux, I conclude .4% is a realistic number.
And places like OneStat and others confirm that .4%.
Ok. .36% on onestat.
http://www.onestat.com/html/aboutus_pressbox46-operating-systems-ma…
any link to google’s 1%?
any link to google’s 1%?
http://www.google.com/press/zeitgeist/zeitgeist-jun04.html
“Win95 has double the market share of Linux.
”
“Win ME has quadruple the share of Linux.
”
And you think those are better operating systems than current state of Linux ?
You’ve just proved to yourself that Microsoft’s monopoly has nothing to do with the quality of their products.
If you hate Linux and open-source so much, why do you even bother leaving comments in these Linux threads?
Oh I know… to irritate people. Correct?
It’s obvious, isn’t it? You don’t read up on a topic unless you have some interest in it, so he’s likely interested in Linux and/or Ubuntu (or is so bored with his life that he reads anything and responds to anything to pass the time). Assuming the former, he’d like to switch or wants Linux to be an option for him if Microsoft continues down it’s path of treating people like second class citizens, but he doesn’t yet feel comfortable with it or has grown so cynical with the Windows culture or seen how Windows beat out superior products time and time again that he doesn’t feel things could ever change. So he’s resigned himself to justifying the cynicism.
As someone that left Linux in 1996 for Windows and then returned in 2001, I can definitely relate to that point of view. The only thing I can say is that from my experience, the best way to fight cynicism is not to fight it — just be open to the good and the bad. Linux’s openness won me back (and Microsoft’s arrogance helped push me the rest of the way). Linux has it’s warts. So does Windows — if you use it at work or work near a help desk, you know at least a half dozen horror stories. Even MacOSX does too (e.g. if your desktop is read only and your app tries to put things on the desktop you’ll get a really cryptic error message and have no idea how to fix it unless you’ve dealt with this issue before). Most of us turn a blind eye to the warts in our OS or gotten used to them or developed hack workarounds since if we don’t, we’d be complaining all day. The important thing to point out to anyone investigating Linux is that if you use Linux, do things *the Linux way* and not *the Windows way* or *the Mac way*, since if you try to make Linux something its not you’ll get frustrated. The same can be said about Linux or Mac users that try to use Windows or Linux and Windows users who use a Mac.
If you’re frustrated with Linux, then either you’re using a distro that’s too techie for your abilities or tastes or you’re unknowingly carrying your old habits (or expectations) into Linux and are fighting the OS when you really don’t need to (in nearly all cases). Because these days, unless you have very Windows specific needs, Linux is a fairly comfortable OS to work in.
If you’re frustrated with Linux, then either you’re using a distro that’s too techie for your abilities
Or you’re running a PPC distro so you get
1) no Adobe products
2) no Real Player
3) Firefox updates weeks or months later than the x86 crowd.
—
And yes, the sort of websurfing I do at both home and work makes use of a lot of multimedia.
Or you’re running a PPC distro
When you are on the PCC platform and you are dependent on top support for Adobe products, Real products and FireFox, you are better off not using GNU/Linux.
For all the niceties PPC may carry, it is more of a niche platform now, than it ever was, now that even Apple has abandoned it.
GNU/Linux started out on commodity x86 hardware and that is where its strenghts still lie.
“Win95 has double the market share of Linux.
”
“Win ME has quadruple the share of Linux.
”
And you think those are better operating systems than current state of Linux ?
You’ve just proved to yourself that Microsoft’s monopoly has nothing to do with the quality of their products.
No, what those figures just prove is what really matter to people when it comes to operating systems are two thing: first, they want something they don’t have to mess much with in order for it two work, and second, they want to have every application under the sun (including games) available for that system.
Why some people just can’t get those two simple facts through their head is beyond me. Windows 9x was a crappy line of OS but it managed to give people that, and they would still be massively used today if XP did not exist because of those two reasons. No Linux distribution can match that yet.
Ever thought those systems exist because of the hardware they run on? I have worked with some of those machines and they don’t need an upgrade, they need a drop out of the window.
This statistic cannot be used to accurately measure Linux installations, because:
a) most Linux browsers masquerade the OS as something else, and
b) the percentage of Linux users with broadband is much higher than for Windows, which means a smaller overall number of connecting IP addresses (broadband connections keep the same IP much longer than dial-up).
Not only that, but the actual contents of the website will skew the results. This is the reason Google took out the OS survey from their Zeitgest, because it was simply inaccurate and under-represented Linux/*nix desktops.
This statistic cannot be used to accurately measure Linux installations, because:
a) most Linux browsers masquerade the OS as something else,
Evidence?
and
b) the percentage of Linux users with broadband is much higher than for Windows, which means a smaller overall number of connecting IP addresses (broadband connections keep the same IP much longer than dial-up).
Evidence? (I mean evidence that you understand how page requests are logged?)
Not only that, but the actual contents of the website will skew the results. This is the reason Google took out the OS survey from their Zeitgest, because it was simply inaccurate and under-represented Linux/*nix desktops.
Evidence?
Come on. The sooner you accept Linux “installed base” is trivial, the sooner you can leave the cult and get a real life.
I know this is hard to understand for someone like you who has an irrational anti-Linux bias, but it’s really a matter of common sense.
If you’ve ever used a Linux desktop, you know that browsers can be set to identify your PC as using Windows. This setting is used more often than not, because many websites are “optimized” for IE and Windows, and may refuse access (of offer less functionality) for non-supported browser/OS combinations. A good example is using Konqueror to access Google Mail: if you don’t masquerade the browser, you only get basic functionality. In order to get the full experience, you need to pass off as a IE/Windows combination.
As for the second point, again it’s a matter of simple logic: the proportion of broadband usage among Linux users is greater than for Windows, for the simple reason that Winmodem support has always sucked on Linux. Also, Linux users tend to be more computer-savvy and spend more time online, which means they are more likely to have broadband than Windows users, and therefore less likely to have a new IP address when visiting the same site twice. Since most counters on websites use the visitor’s IP address to compute their statistics, Linux users tend to be under-represented.
All of this is of course irrelevant to the innate qualities of Linux, and its many advantages over Windows. But I don’t want to expose you to too much logic, as your head may asplode…
I wanted some evidence that a significant number of Linux users alter the OS identification.
You bring up Konqueror. Here is a list of user agent strings in use by Konqueror:
http://www.pgts.com.au/pgtsj/pgtsj0208j.html
Almost all of them clearly identify the OS sucha s Linux or FreeBSD even though they also say they are Mozilla/5.0 compatible.
So … maybe you could read up on user agent strings and then come back to me with some evidence most Linux users alter those strings.
the proportion of broadband usage among Linux users is greater than for Windows …
Actually, by your logic Linux is overrepresented on weblogs.
Since most counters on websites use the visitor’s IP address to compute their statistics
No, they don’t.
They count the number of requests by the user agent string. Then they have to parse out the user agent string for OS names and browser types.
Maybe you could Google “user agent string” and improve your education.
You bring up Konqueror. Here is a list of user agent strings in use by Konqueror: