Don't compare GNU/Linux with Windows or MacOS - they are not in the same game

Don't compare GNU/Linux with Windows or MacOS - they are not in the same game


Recently a blog post entitled "Why Desktop Linux is its own worst enemy has come across my feed-radar a few times. It's yet another in the long line of "Linux ain't ready yet" jeremiads and it doesn't really say anything new yet it got on my nerves. Why?

Unsupported statements

Like many such pieces, this one starts by making a statement as if it were fact while presenting no actual evidence. The "fact" used here goes along the lines of "Microsoft has shot itself in the foot with Vista and the only ones benefiting are Apple". Even if there were statistical evidence, the premise is mistaken. As I have said before the "success" of GNU/Linux cannot be measured in the same way as a proprietary OS. Apples and air people, you're comparing apples and air. I mean how can you tell how many Ubuntu installs came of a single CD?

The "success" of GNU/Linux cannot be measured in the same way as a proprietary OS

Speaking about unsupported, the post then reels out that hardy perennial "not working out of the box". Apparently "even getting MP3s to play can be while using Linux can be problematic". Which distribution was that then? If playing MP3s out of the box is your thing try one of the distros that comes with that feature out of the box: Linux Mint, PCLinuxOS etc. Okay perhaps I've made that too difficult for people there, I mean perhaps they want "Linux" to be just "Linux": all distros being the same. You know like Windows with its seven different versions of Vista (with their range of hardware requirements) from a single supplier. To be fair the author does try to sound like she's not blaming GNU/Linux:

"Perhaps it's not the fault of Linux as much as it is the pervasiveness of Windows. Microsoft, after all, works with various hardware manufacturers to ensure their hardware works with Windows. And you still have driver problems crop up on occasion so what else can you expect with Linux."

Hmm, Microsoft works with hardware manufacturers? These would be the same manufacturers who said they weren't ready and yet Vista was still released. The Vista release was such a sore-point for Microsoft that a month ago another blog on the same site claimed that Microsoft is asking manufacturers to start testing the next Windows release now. But what follows that bit had me in absolute fits of laughter:

The average consumer just wants to be able to pop a CD into his optical drive, wait 10-15mins and have a working operating system.

Most of the average users I know would rather buy a new PC than upgrade Windows

The average consumer wants what? And in how many minutes? Has this blogger ever tried to install Windows? Sorry but this is just a ridiculous claim. Show me this average consumer who wants to install their OS? Show me any modern OS that installs in 15 minutes (best I've achieved is 18 and I'll assume live CDs are not allowed here). Most of the average users I know would rather buy a new PC than upgrade Windows. No, users wanting to install an OS in 15 minutes is a pure straw-man argument.

Another old chestnut

So what's next? Ah the old "why there are so many GNU/Linux distributions?" question is given a new coat of paint:

"What is it with the collective egos of Linux coders that if one distribution doesn't suit them that they have to go and make a new one"

GNU/Linux comes in different flavours to fit the end-users' needs, wants, desires and just for-the-heck-of-it sense of curiosity

Talk about missing the point. One reason there are "so many" distributions is because there can be and the ones that keep going are the ones that people find useful. Not every GNU/Linux user likes the idea of installing a whole desktop system just to set up a firewall, or a router or to re-cycle some of that hardware that Windows won't even get out-of-the-box for. Some want to play MP3s out-of-the-box, some want a small install footprint, a faster boot time or a longer support lifespan. Windows comes in different sizes to fit Microsoft's sales message. GNU/Linux comes in different flavours to fit the end-users' needs, wants, desires and just for-the-heck-of-it sense of curiosity. The author does offer a solution to the "problem":

Instead of rallying behind a single distro and making it the OS to beat, Linux grokkers tweak and promote their own Linux 'flavours'.

So let's dip into some analogies. Why are there so many amateur sports clubs, spare-time inventors and (best of all) tech weblogs? Why don't any of the sports enthusiasts just get behind another local club and make it the club to beat? And what's with the collective ego's of tech bloggers that if one poorly supported anti-GNU/Linux rant doesn't suit them they have to go and make a new one?

Here's a wake up call to those who keep asking the "why so many distributions?" question: as long as GNU/Linux is available under the GPL, there will be those who will tinker, tweak and create for no reason other than to enhance their own experience and maybe help a few others along the way. We tweak because we're allowed to and because, sometimes, it's fun.

The final straw

Their next point is not the one I laughed at the most, but after the others it was the straw that broke the camel's back (and I say that as a part-time Perl monger). Apparently the real litmus test is the documentation. Windows is -- quite cleverly-- not mentioned here but instead the praises of Apple are sung loud and clear. GNU/Linux documentation is apparently too archaic. Ubuntu is given some credit for doing a "pretty decent job" but apparently that's not enough. And as for the rest.. "Well?". Well what?

I've never purchased an Apple computer so I couldn't tell anything about the documentation. What I would say is that at the prices they charge for their machines I'd expect it to be perfect. But what documentation are we talking about here? The GNU/Linux system is made up of thousands of applications, some of which are well documented and others not so well. So are we talking about GNU/Linux documentation here or free software documentation? No I'm not nit-picking because you see there are thousands of excellent Howtos out there on a whole range of subjects and not all of them will have the word "Linux" attached but they all apply equally to applications that are run on GNU/Linux. Okay so maybe it could be better collated but so could a lot of things, including my CD collection. Another thing to remember is that there is a whole secondary market of books that has grown up to fill this very niche. Several of the publishers involved sponsor articles in this magazine and you'll find reviews of many tomes here as well. Yes they cost money but I'll wager it's less than the cost of that Mac that came with the shiny brochures.

The dummy race

Their closing argument is this..

In essence, until Linux becomes dummy-proof, it's not going to win over consumers. Make it easy, make it accessible - until Linux programmers get that, it's more likely that Apple will perhaps double its user base in the years to come at the expense of both Windows and Linux. It's not about the best OS winning - it's about the OS with the best user experience and Linux still isn't there yet.

What's not about the best OS winning? And winning what? The market? That's changing as we both write, and a few years from now probably won't resemble the current one. When I read that statement though I was reminded of the quote by Richard Cook:

"Programming today is a race between software engineers striving to build bigger and better idiot- proof programs, and the Universe trying to produce bigger and better idiots. So far, the Universe is winning."

If you make something simple to use, there will soon come along another set of users too lazy, stubborn, or apathetic (but rarely too “stupid”) to learn how to use it properly

"Dummy-proof" is a moving target. It's one of those circle-of-life things (as Disney like to say). If you make something simple to use, there will soon come along another set of users too lazy, stubborn, or apathetic (but rarely too "stupid") to learn how to use it properly. The answer is not (always) to make it easier to use --by which most people seem to mean "hide half the functions"-- but to make learning it more interesting. The quality of a user experience should not be judged by the cuteness of the help avatar or the number of steps in a wizard (or even by calling it a wizard). It should also be about how much it enhances your life/work, widens your perspective and awakens the child-like hunger to learn in you. It should make you want to show off what you can do to your friends. Well it is if you ask me but I'm not sure any software, free or proprietary, has achieved that yet. Still it's a good target.

Conclusion

Free software makes users participants again they are not mere consumers

As I said there have been a long line of posts like this, so why did this one get my goat? I think it was because it addresses software users as consumers. You see, for me, free software (including GNU/Linux) reverses the trend that separates software supplier and consumer. I am one of those who baulks when train drivers address passengers as "customers". I dislike it when TV viewers or radio listeners are treated like cash machines. Making someone a consumer reduces their contribution to fiscal terms and people are more than that. Free software makes users participants again they are not mere consumers. Sure they may start out that way, especially if they migrate from Windows, but in my experience they usually learn pretty fast that this game has different rules. Rules which let them play instead of just spectate. So you can't really say that because GNU/Linux is not addressing the needs of consumers it's failing, because largely it's not trying to. There is no free software marketing team, no advertising budget, no "mission statement". Free software and GNU/Linux are simply there. Pick them up, use them if you want. If you don't well that's fine, it's your freedom but as I said before -- the game has changed so comparing apples and air is a bit of a waste really.

Category: 

Comments

John Calcote's picture

Thanks so much for this perspective on the free software world!

I compare this article's sentiments with those of another (entirely different) topic that is near and dear to my heart lately -- that of rising gas prices. I read an article on Slashdot this morning indicating that Mercedes was planning to phase out fossil fuel cars from their line-up by 2014. Now how are they going to do that!? I don't know, but I'm behind them 100 percent! It'll be a pain in the neck for the consumer (for a while), but ultimately, we'll all get something very important to us. What we'll get from it is NOT an easy time at the filling station, but rather, the ability to NOT have our financial lives controlled by oil companies. We'll have a choice.

This is what GNU/Linux is REALLY all about; giving us the freedom to NOT have our choice of software controlled by one company, whose goals and choices are not in the best interest of the consumer, but rather its own.

Don't get me wrong, I'm not an RMS bigot, but I do love the idea of having a choice. And with choice comes complexity. You can't have it both ways, so I'll take choice over ease of installation any day.

craigTFD's picture
Submitted by craigTFD on

... except that having used Linux since the days that Red Hat was distributed on floppies, we have much greater choices today, and much greater complexity, and ease of installation on any of the major distributions is unrivaled. Whether it is on PPC (Apple), x64, i386, the entire process just keeps getting better and better, because of all the wonderful people working on it. That's freedom AT WORK.

Rambo Tribble's picture

Commendations on an excellent analysis. Also, your prose is exceptionally well-constructed and coherent.

I'd offer only one consideration: You may not appreciate the actual purpose of these uninformed opinion pieces. They are constructed for one purpose only, to reassure those who embrace ignorance that they are right to do so.

Since the inception of the industry, much of the popular media has been devoted to this activity. From William Randolph Hearst to Rush Limbaugh, empires have been built on pandering to those unwilling to embrace learning or facts before coming to conclusions. Fortunes await those who cynically offer this debilitating comfort to the unquestioning hordes. That fact is not lost on those who market Microsoft or Apple products.

It is, I think, reflective of this social malaise that one often hears, "Ignorance is bliss," proffered as some estimable literary quote. It is not. What the gentleman actually said was, "When ignorance is bliss, 'tis folly to be wise."

Ryan Cartwright's picture

Commendations on an excellent analysis. Also, your prose is exceptionally well-constructed and coherent.

Thanks. You should have seen the first draft :o)

I'd offer only one consideration: You may not appreciate the actual purpose of these uninformed opinion pieces. They are constructed for one purpose only, to reassure those who embrace ignorance that they are right to do so.

You're probably correct in that it becomes one of the ways they are used but in these days of hits and visitors (and advertising) I think another reason behind such pieces is to attract readers.

Popular media has also long used the idea of publishing a contentious or controversial piece to bring in the outraged as well as the interested. Those in the UK would probably recognise this as the "Freddy Starr ate my hamster" technique. Hey, perhaps that's how they got me there?

cheers
Ryan

craigTFD's picture
Submitted by craigTFD on

Ryan Cartwright makes the point that "Windows" (which I take to mean Vista) and "Mac OS" (which I take to be OS/X 10.5 and later) are not in the "same game". http://www.freesoftwaremagazine.com/columns/dont_compare_gnu_linux_windows_or_macos_they_are_not_same_game

I beg to differ.

There is no doubt that Vista has had it's share of problems, especially with drivers as Ryan points out. Ryan fails to mention the vast library of device drivers that Linux has, production tested, ready and stable. It's true that many companies have not yet seen the wisdom of preparing open source drivers, but even the closed source drivers for Nvidia and ATI cards have evened the scales. NDISwrapper levels the playing field for networking, and basically everywhere you look, open source and Linux are taking back the field from proprietary software.

But Windows is much more than Vista. From Windows 3.1 to 95 to 98 to ME to NT to 2000 to XP and Server 2003, the Windows environment covers a vast array of hardware. Linux, in some form, runs on all of it, from embedded systems, to laptops, to desktop, to servers, to clusters and clouds. Whether we like it or not, we are in competition for the hearts and minds of consumers, in every one of those areas.

Historically, some of the kvetches were correct. Multimedia has been a sore point, especially due to proprietary formats and patent restrictions. But my install of SuSE 11.0 has community repositories that solve those problems, and once correctly (and easily) set up, out perform their Windows counterparts. I have a Vista box primarily used as a DVR, and the files it creates cannot be read by my other Windows XP boxes. Yet my Linux box plays them in Xine without a hitch. Why do I have Windows boxes, you say? I have been waiting for Myth TV to evolve, and for my precious games to work. Now that has happened, and while many in the open source community fault TransGaming for Cedega being closed rather than a set of updates to Wine-X, I just don't care. Today I can put together a Linux multimedia box or a Linux gaming box that is more compatible and better performing than the Windows alternatives. Will people buy such boxes? I have no doubt that they will. Will Microsoft retaliate? I have no doubt they will.

I have to agree with Ryan. Most people do not want to install or configure their computers. They just want to have them plug in and work. The interesting thing is trying to do that with Linux boxes on a commercial scale. At most computer stores (or manufacturers for that matter), they have a configuration on one hard drive that they clone, then "reseal" so that the box does a limited configuration for a new serial number / activation code. AutoYast can do something similar, but does not clone the other updates - community repositories, etc. - and the SuSE system is now drive specific (thanks to the Microsoft agreement?) so that you can't just Ghost copy it to another drive. I'm sure some script kiddie can recreate the manual edits I had to make to the boot loader and related files, and for that matter, I probably could do it in a few weeks. But suffice it to say that the commercial redistribution of working configurations - err, needs work.

We don't have to rally behind a specific distribution. The market already does that selection for us. Try http://distrowatch.com/ where you can find all the major distributions and minor ones too. I prefer KDE based Kubuntu to Ubuntu, KDE 3.5 to KDE 4, and I'm entitled to make those choices, just as you are entitled to make different ones. But if you are trying to replace Vista, and now provides a big opening to do just that, you'd better provide a familiar environment for those people voting with their dollars. I think that a Kubuntu or SuSE KDE 3.5 based box is the way to do that, but I would not presume to tell the FluxBox crew that they can't have it their way, or that their window manager isn't better for some purposes. But for the specific purpose of giving Windows users an environment they will intuitively understand, I think that Novell has chosen the correct course (they also supply GNOME).

I've never gotten adequate documentation with a Windows system, unless I bought the whole programming environment. My favorite distro's have good documentation on the disk, and outstanding, helpful user communities online.

What is without a doubt is that Vista has left many users wanting, and like a wounded wildebeest bleeding on the savannah, the hungry lions are moving in on it. Will Linux take the market share, or will it be left to Apple and others? I think Linux is finally in a place where it can compete as successfully as it has in the server market, by companies that understand its strengths and weaknesses. Why do you think Sun Microsystems and IBM and Novell are all backing it? Do you really think it is some form of corporate community service? Or is the reality that they have been carefully crafting a tool to end Microsoft's dominant market share, segment by segment? If so, qui bono? Who benefits? Thanks to Free Software Magazine, and the entire Open Source movement, all of us.

Ryan Cartwright's picture

Ryan Cartwright makes the point that "Windows" (which I take to mean Vista) and "Mac OS" (which I take to be OS/X 10.5 and later) are not in the "same game".

I beg to differ

Actually I was really referring to any incarnation of those OSs. The "game" I referred to was the one where marketing, number of users and sales figures reign. GNU/Linux is not in that game. I wrote about this in an earlier post.

Your post makes some excellent points and in general I don't disagree. The only thing is that you've disagreed with something I wasn't trying to say.

The point I was making was that comparing any proprietary offering with a free software one on the basis of sales and numbers, and thereby adjudging the free software one to be "failing" is futile. Free software does not operate under the same rules, it is not restricted by the same principles and its users can take part rather than merely consume it.

While I take your point about GNU/Linux being ready (and I agree it is) I think we need to remember that by nature it requires its users to learn it. Thus it doesn't have to become a shoe-in for Windows or MacOS, it doesn't have to do things the way Windows users expect because that might not be the best or most efficient way. Developing free software gives you opportunities to take a step back and think "What if it wasn't this way at all?". Once you stop selling licences (and there's little point if others can give away your software) then the future of your product becomes less dependant on keeping existing customers. If the Python team sold licences do you think they could have started on the route that means Python 3 000 will break code written for earlier versions? They've torn up the rule book and started afresh and my hat is off to them, it's a brave move but one that could make Python better in the long run. Microsoft and Apple no longer have that luxury. They can't make that radical kind of change without losing their all-important customers. Apple did when it was make or break but now? Not a chance.

cheers
Ryan

Terry Hancock's picture

Of course, it has to be appreciated that users who don't want to upgrade can always keep using the 2.x series of Python -- I'm sure that 2.6 will be maintained for a LONG time (by somebody, if not necessarily by the core Python development team).

OTOH, Python 3 looks very promising -- it's always been one of the benefits of Python's development strategy that they aren't afraid to deprecate stuff when a better way comes along. It requires you to move along with the language, but it keeps code clean and understandable to other developers.

I think the main distinction about the Python 3 release is that some of the "deprecated" stuff will actually disappear from the language entirely, resulting in simplifications to the interpreter (e.g. who uses "oldstyle" classes anymore? -- everything is "newstyle" classes these days).

But again, it's not like Python 2.x will disappear from the Earth. And that really is the distinction from proprietary practice, where the company is always trying hard to kill their own old products in order to make people buy the new ones.

Ryan Cartwright's picture

Of course, it has to be appreciated that users who don’t want to upgrade can always keep using the 2.x series of Python — I’m sure that 2.6 will be maintained for a LONG time (by somebody, if not necessarily by the core Python development team).

True but it's still a brave decision. This point highlights yet another plus point for free software. With proprietary licences the older version gets dropped after a while and users are left out in the cold - no matter how many of them there are.

cjcoats's picture
Submitted by cjcoats on

I have just recently had the "opportunity" to watch new installs for both Linux and Windows. The Linux install (with Mandriva 2008.1) very nearly was the "pop a CD in and 15 minutes later it's done" -- it was a DVD, not a CD, and it took closer to an hour, given that I was selecting all the developer packages and then downloading all the package-updates off the net. My wife's new Windows machine has been a nightmare--so far, she has spent at least a hundred hours trying to get everything right. (That is, counting three different re-installs when the networking driver updates screwed up irretrievably...)

MrHasBean's picture
Submitted by MrHasBean on

I write this response from the perspective of someone who owns and runs systems running Windows, MacOS and Linux, and I have supported and developed for these systems for many years. I'm wondering if somewhere along the line you have missed the point of the original article? I have read and agree with some of the things you say, yet in relation to the original article and its context your points actually contradict each other in places.

To begin with I'm not sure I understand the comment about the measurement of the success of GNU/Linux? You suggest that there could have been many installs done from a Linux CD and therefore the number of CD's / Downloads isn't a true gauge. Is this not the same for Windows and MacOS - not SO much MacOS although I'm pretty sure there are many older systems out there running Leopard that were installed from a "gifted" copy. The "how many computers running this OS are accessing the internet" figures tend to be in the same ballpark as other distribution figures so I think its fair to assume they are somewhat accurate.

You then rightly point out that there are so many flavours of Linux for very good reasons, yet at the same time you rightly say that people don't want to be installing their own OS, they will just go and buy a new system pre-installed. Are you suggesting therefore that the manufacturers should be offering pre-installed on their systems whatever version of Linux the consumer determines is suitable for their purposes? How does the consumer first determine this? Should the manufacturer be recommending certain ones for certain applications? This is a logistics nightmare. This point also contradicts your conclusion where you then talk about making consumers into participants. Surely if the average user doesn't want to even install their OS they certainly don't want to "participate" in getting it to do the things they want the computer to do for them?

Further on that point, when you discussed the flavours of Linux you seemed to be suggesting that this was partly due to specialist distros - firewalls etc. I really don't think this is what was being discussed in the original article. Specialist distros are very useful, and rightly you wouldn't want to use a generic "workstation" distro to install such systems if there is a better alternative available. I use a number of specialist distros myself. However the point is not lost that there are probably too many "workstation" distros for consumers to easily make a choice. You mentioned Microsoft's plethora of Windows versions, and I agree that they have gone way over the top. However your comparison is flawed because at least Microsoft (in their collective mind) are targeting their versions at different market segments. What I believe the original article was saying is that the choices in Linux would be akin to Microsoft offering 18 different "Home" versions depending on what you wanted to do with your "Home" version. I'm not suggesting here that there should be one version of Linux that everyone should promote and develop, however an agreed set of standard components that are available and installed by default by every "workstation" distro would not harm Linux's accessibility. For example in the automotive market at each market segment there are a set of standard features you would find on every vehicle, its the extras that are added by the manufacturer and the overall difference in presentation that attract different buyers - a bit simplistic I know but you get the idea. Your sports club analogy is also flawed in that sports clubs service different sports and different locales, so there is really no comparison there.

I appreciate what you were trying to do. I love free software, but the thing that gets on my goat about many people who support free software is they talk over and over about people's rights, yet they completely ignore, to the point of being anti, the rights of those people who want to make a living so they can feed their families from their efforts developing software. It is as much the right of the developer to either release his / her software under one of the open source licences or charge for it, as it is the right of the consumer to use free software or purchase what they use. I use both. I also develop both. They don't need to be mutually exclusive.

I agree with the message of the original article. GNU/Linux is NOT ready for prime time, just the same as MacOS isn't ready for some corporate markets. Is that a weakness? Not at all. It is ready for the market segments that want to adopt it, just like MacOS. I believe over the coming years as more agreements about standards are developed that could very well change for Linux. But then, the desktop market as we know it is a racehorse nearing the end of its track life - just waiting to be put out to pasture, so what we're discussing here will likely be moot in a few years anyway.

Ryan Cartwright's picture

To begin with I'm not sure I understand the comment about the measurement of the success of GNU/Linux?

That point was the whole purpose of the article so if you've not understood that - sorry. Let me try again...

There's really not a lot of point comparing proprietary OSs and a free software OS because they operate and have been developed under different principles. If the original article had ripped into the commercial offerings of say Redhat or Novell it would have been understandable as those companies are trying to compete in the Microsoft world. But GNU/Linux itself is not. There is no single entity making decisions about direction, features or what should or shouldn't work. It's a huge collaborative and almost organic work. Now that may be a disadvantge in the Microsoft world but to say "Linux is failing" in that context is not really valid because "Linux" is not a single thing and in that way cannot be trying to compete with Windows.

You suggest that there could have been many installs done from a Linux CD and therefore the number of CD's / Downloads isn't a true gauge. Is this not the same for Windows and MacOS

No, not at all. Ubuntu can be downloaded and burned to a CD, that CD can be used to legally install Ubuntu on any number of PCs. As others have suggested you can also burne more than one CD from a single download. When comparing the success of proprietary offerings you cannot really count the illegal copies as there's no direct revenue for the company there and in that world, revenue is what counts.

Are you suggesting therefore that the manufacturers should be offering pre-installed on their systems whatever version of Linux the consumer determines is suitable for their purposes?

I'm suggesting manufacturers can offer a range of OS options. Some already do. They offer Ubuntu by default and have others as an option. It's no extra cost for them or the buyer.

How does the consumer first determine this?

How does the consumer first determine which version of Vista to buy with their PC? The answer is most of them don't until they've tried it. What's to stop a retailer offering free Live CDs of various distros for potential buyers to try at home or in the store. If it's online they can offer the download. But even if the buyer feels they've made the wrong choice they can switch to another distro for no cost and little effort (they even know all their hardware already works with GNU/Linux). This has happened to my father in law who was having great trouble with Ubuntu and creating DVD slideshows. As that was one of the reasons he got the PC, he switched to Debian because he knew it worked under that.

This point also contradicts your conclusion where you then talk about making consumers into participants. Surely if the average user doesn't want to even install their OS they certainly don't want to "participate" in getting it to do the things they want the computer to do for them?

That's not what I meant by participate. I meant they get to take part in directing the future of their software. Also turning a consumer into a participant is not a one-off thing. It's a process that begins when they first try free software (if you ask me).

Further on that point, when you discussed the flavours of Linux you seemed to be suggesting that this was partly due to specialist distros - firewalls etc. I really don't think this is what was being discussed in the original article.

True and perhaps I was being cheeky there but many of the distros of which people complain there are "too many" are not desktop ones anyway.

However the point is not lost that there are probably too many "workstation" distros for consumers to easily make a choice. You mentioned Microsoft's plethora of Windows versions, and I agree that they have gone way over the top. However your comparison is flawed because at least Microsoft (in their collective mind) are targeting their versions at different market segments.

Surely the point of the different desktop distros is that they are targeting different subsets. If you want guaranteed free software across the board you probably go for Debian. Want the ability to play MP3 and video regardless of licencing issues would mean you go for something like PCLinuxOS, Mint etc.. Sit somewhere in the middle and you'd probably go for Ubuntu. There are of course millions of reasons why people go for a particular distro but the fact that they can is a good thing.

Your sports club analogy is also flawed in that sports clubs service different sports and different locales, so there is really no comparison there.

In my local area there are approximately 7 (adult) amateur football ("soccer" if you are from the US) clubs. They play in different leagues and at different levels. Some play for fun, others for fitness, others take it very seriously and they have varying sized fan-bases. On that basis the analogy stands. The reason I said "sports" rather than football was to avoid the confusion caused when referring to a sport with that name.

I appreciate what you were trying to do. I love free software, but the thing that gets on my goat about many people who support free software is they talk over and over about people's rights, yet they completely ignore, to the point of being anti, the rights of those people who want to make a living so they can feed their families from their efforts developing software.

I've not mentioned cost anywhere have I? Are you mixing up free as in cost with free as in freedom? People are perfectly at liberty to sell free software. I think they'll find it harder to sell licences if those licences give their users the right to give the software away but they are at liberty to try. In that case the market changes to selling services rather than software itself. In fact that is one of the reasons why you can't really compare GNU/Linux and Windows. GNU/Linux is not so much in a different league to Windows but in a different game: Football and Rugby, Baseball and Cricket - you can't compare the success of a team in one with that in another.

It is as much the right of the developer to either release his / her software under one of the open source licences or charge for it, as it is the right of the consumer to use free software or purchase what they use. I use both. I also develop both. They don't need to be mutually exclusive.

Agreed - so what does this have to do with my article? I write and use free software too BTW. Don't confuse the call for freedom within licencing to be the same as calling for zero cost software.

I agree with the message of the original article. GNU/Linux is NOT ready for prime time,

Okay this proves it - you did not understand the premise of my article and that's probably my fault, sorry. I was not arguing whether GNU/Linux is ready for the "prime time", I was saying to make such a statement is worthless and then tried to say why.

For the record I think GNU/Linux is ready for pretty much any use. I don't know what "prime time" is but I don't think GNU/Linux is a drop in replacement for Windows. Not because it lacks the features but because of the reasons I give above. It's a different approach a different way of doing it, a different game. Just as some skills learned in Baseball or Cricket are transferable the same applies to Windows and GNU/Linux. Moving from one to the other requires a different way of thinking. You will have to grow into playing the new game.

I know some have called this a weak argument -- "Linux is 'special' so it can't be compared" -- but actually I reach this view after several years of having to compare free and proprietary OSs in various contexts. More and more I find that there is very little common ground on which to compare the two so - from whatever side of the fence ones sits - comparing becomes a somewhat futile exercise.

cheers
Ryan

craigTFD's picture
Submitted by craigTFD on

This has been an enlightening post, and a terrific exchange, but your idea that Linux, Windows, and Mac OS/X are not in the same game is simply wrong. I think it is more correct to say that Linux is a different Strategy. Linux is on offense in the server market, Microsoft is on defense. Microsoft has been on offense in desktops, Macintosh and Linux are now forcing them to play defense. A goal is scored everytime a person votes with their dollars. Remember that the original purpose was not freedom for all, but the elimination of Microsoft as a competitor.

Your point about the number of systems being counted as being wildly inaccurate is similar to the "Linux is special" argument (Linux IS special, but not in this way). I don't know how many copies of Knoppix I'm responsible for, but their are techs who I've never even met who are using copies of the CD's that I gave to other techs (hopefully they will learn to download their own). Novell can give you a pretty good count by the number of systems that check into the repositories; I'm sure the Debian distros could do the same. But the point is that market share is not just in the mind, and ALL marketing numbers are somewhat "squishy". There are also experiments, like Dell and Wal-Mart, the new laptops, and IBM/Sun offerings, that have a chance to REALLY change the numbers out there. And of course, there are the embedded systems, the home systems, that never connect to the net and thus don't exist.

This is kind of like following a hockey game on TV, and if they don't put the little red spot on it, you have no clue where the puck is at any given point in time. That doesn't stop either team from scoring, just because you can't (or won't) see the puck. The score still matters too, at least to get you into the playoffs.

So our farm teams and amateur clubs are deep. We have fewer professional clubs, but that may be because our "full monty" strategy of play is not well understood. They may laugh at us. That's OK because we have nothing to hide. Maybe we don't have the fancy logos on our t-shirts (well, we don't wear shirts anyhow, it's part of the strategy, you can see our code), but we do have sponsors, and moreover, we have the passion for the game. Fill the stadiums, and play the game. Is our "full monty" strategy better? Play the game and we will find out.

Ryan Cartwright's picture

Not sure who are addressing here me (Ryan) or one of the other posters. I'll assume it's me.

This has been an enlightening post, and a terrific exchange, but your idea that Linux, Windows, and Mac OS/X are not in the same game is simply wrong. I think it is more correct to say that Linux is a different Strategy.

You and plenty of others (on [n]either side of the fence) keep talking like GNU/Linux is a single thing with strategies, marketing and business thinking behind it -- it's not. Some companies that use it within their business are those things but they probably don't make up half the GNU/Linux user base (can't prove that just feels right). In that sense GNU/Linux is not in any game at all, it's just there. Use it, don't use it -- it'll still be there in the morning. Think of the FUD strategies Microsoft et al have tried - have they actually succeeded in averting their users' eyes from GNU/Linux? I'm not so sure. Those strategies can't work against something that does/can not see itself as your competitor.

Look I want to see GNU?Linux used in more places by more people because I want to see people freed up when they use their computers. I've looked at this for some time now and I cannot figure out a way to identify an achievement that signifies when GNU/Linux can be said to have "succeeded" and it it can't succeed -- it can't fail.

Linux is on offense in the server market, Microsoft is on defense. Microsoft has been on offense in desktops, Macintosh and Linux are now forcing them to play defense.

Use of GNU/Linux is on the rise in server, as I have said above there's no strategy to get Debian, Ubuntu, Gentoo, Fedora or Linux From Scratch into the server market. The rise in use has come about by word of mouth and peer-recommendation not any marketing strategy. The same has applied to the rise in desktop usage. Have you seen any marketing strategies pitching Ubuntu against Vista? I haven't. It's largely been by word of mouth and then the computer press caught wind of it.

A goal is scored everytime a person votes with their dollars.

But it's not about dollars, yen, punds, euros or any other currency. This is the point -- GNU/Linux (or more exactly free software) has reversed the trend that saw software marketed like apples or books. This is why software patents are bad because they try to fit a marketing principle to something which doesn't fit that mould.

Yes people can and do make money off the back of free software, yes businesses can thrive using it but how many of those actually turn a profit from sales of software alone? How many proprietary companies do (apart from Microsoft who use monopolistic tactics to do so)? I don't know the answer to either but I reckon it's fairly low or else we wouldn't see the drive towards service provision, long term support contracts and other such ways to make money from software.

Remember that the original purpose was not freedom for all, but the elimination of Microsoft as a competitor.

Whose original purpose? Stallman's? He wanted to share code freely as he had done before. Torvalds'? He said this thing wouldn't amount to much. Redhat's? They wanted to make some money making it easier to install GNU/Linux. Debian's? They set out to make a completely free OS.

I think you might find some problem proving that any key person/body involved with the development of GNU/Linux said their purpose was to eliminate Microsoft as a competitor. I'll be happy to read it if you can find it.

Novell can give you a pretty good count by the number of systems that check into the repositories; I'm sure the Debian distros could do the same.

Unless Novell embed a unique install code into each installation surely the best they could do is extrapolate something from the IP address and the hit count on the repos. Debian don't even collect stats on popular packages without asking first so I'm pretty sure they won't do the same even if they could.

But even so, they won't allow for figures on every distro or those not on the 'net (okay there are probably fewer of those now). And even if they did the numbers wouldn't matter for much except to the companies involved. I'm pretty sure Debian is not interested in number of installs - why would they be? Numbers like that only work if you can convert them into revenue or marketing figures (to increase revenue).

I will agree that this discussion has been useful. It's certainly attracted a lot more attention than I expected it to. Sometimes you write something because it means a lot to you and you find this happens, other times nobody seems to care - sniff, sniff.

craigTFD's picture
Submitted by craigTFD on

http://www.roughlydrafted.com/RD/Home/1052F769-AC24-4892-8937-E3E90BCFD5CC.html
http://marketshare.hitslink.com/report.aspx?qprid=8

Operating System Market Share
June, 2008

Operating System
Total Market Share
Windows
90.89%
Mac
7.94%
Linux
0.80%
iPhone
0.16%
Playstation
0.03%
SunOS
0.01%
Nintendo Wii
0.01%
OpenVMS
0.00%
SCO
0.00%
OpenBSD
0.00%
NetBSD
0.00%
HP-UX
0.00%
AIX
0.00%
FreeBSD
0.00%
SCP
0.00%

If you don't have Windows in your sights, you are missing the game. 90% of the games are played in their stadium.

craigTFD's picture
Submitted by craigTFD on

http://www.itnews.com.au/News/36258,linux-suffers-server-sales-slowdown.aspx

The game is different in the server market. Microsoft is on offense, Linux on defense. This strategy has resulted in Windows Server being modified so that it can run unmodified on top of Linux.

You know you're good in bed when they want to be on top.

kgoetz's picture
Submitted by kgoetz on

I did like the comment "Why are there so many ... tech weblogs?" I hadn't looked at it like that before, it was ammusing to see.
Incidentally, I did install Ubuntu 6.06 in 8 minutes (from a live cd), but it was on 2gb of ram and a 15,000 RPM SCSI disc, which I agree most "normal users" (heh) dont have.

PS.
Having finally got to this point in my post, I'm finding this resizing comment box *amazingy* iritating. my touchpad mouse keeps resizing it all over the place. Only way i can escape from it is to tab to the 'preview comment' button.

frankone's picture
Submitted by frankone on

Well, you guys can argue over the merits of whatever, but the simple fact is that without support for basic hardware that people have, Ubuntu is not going to do the trick. For the first time ever, I installed something other than Windows, which I have been doing for some 20 years now, and while it did install great and looks great and is faster and seems easier and is less exposed to virus programs, it can't/won't see my sound card and no effort for the past two weeks has been able to change that. I have search over and over for any kind of fix and tried everything I have found on the Internet, to no avail! Without sound, this dog don't hunt!

While it has been an interesting two weeks, and while I really do like the o/s, without basic functionally and support....postings for help even go unanswered.... I an not interested to staying with what doesn't work.

Have a great time debating the ins and outs of babble, but this guy is out of here.

John Calcote's picture

It's an old adage: People who want something for nothing general get what they pay for. It's a real pet-peeve of mine to hear people say "they" really need to improve GNU/Linux hardware support... Just who are "they"? Well, I'll tell you. "They" are people, who fit into one of these three categories:

1. Individuals around the world who add support to GNU/Linux for a particular hardware feature in their spare time so they can use it themselves for their own purposes.

More often than not, such patches or enhancements do not constitute good quality, general purpose code, but rather, hacks that suffice for these hackers' personal needs. Not that they're bad programmers--quite the contrary, often they're some of the best. But they're motivated to do the wrong thing--get it working just enough to be usable for their own needs. We're more or less lucky when they sometimes go the extra mile to make it right.

2. Corporate entities who want to invest as little as possible to get this "free infrastructure" they thought they were buying into in order to get GNU/Linux up to speed for their own business purposes.

While these additions are generally of somewhat better quality than those that come from category 1 people, they still have their own reasons for doing the work, and these reasons are not often centered around providing a really great GNU/Linux platform.

So often I hear in my own work-place statements born of ignorance from management indicating their desire to "leverage free work from the open source world" to accomplish some business goal. I'm sorry, but nothing is free in this world. The reasons for using open source in this context should be centered around ubiquity and learning curve for their customers, NOT around some imaginary free contributions to future products.

I try to warn them when I can: More often than not, a fair amount of work will have to be contributed back to the open source community before the "free" part is usable by the product.

3. Distro companies like RedHat and Novell and Ubuntu, who have a vested interest in seeing Linux do well. But even these companies have resource constraints. They have to be very careful how they spend these resources, or they won't get the return on their investment that their shareholders expect.

This is why RedHat dropped their desktop market a few years ago--they simply weren't making enough money on that end of the Linux field. Foolish people who thought RedHat was in it for religious reasons couldn't understand this move at the time it happened. A wise business doesn't get into religious wars over who owns the desktop. If they do, they fail, and then they help no one.

This is not to say that no one cares about the desktop. Linux distro companies are still spending a lot of money on improving the desktop, but solely for the sake of infrastructure required to ensure their value propositions remain viable. These (somewhat) altruistic endeavors CANNOT take first position within the business, or the business will fail, and then no one will benefit. It's a fine line to walk, and for the most part, they're doing it well.

The bottom line: Support for a particular device, or desirable new features will happen when it happens. If you want it to happen faster, then you'll have to pay for it yourself--either with your time and effort or by paying someone else to do it for you. It's not about strategy or competition. It's just the model. It's just the way things are done in the free software world. We pay for Windows with our money, we pay for GNU/Linux with our patience and effort (and sometimes with our money). But we do pay in either case.

craigTFD's picture
Submitted by craigTFD on

John is correct. There is no such thing as a free lunch, and the success of GNU/Linux is due to the personal efforts made by the huge community of adopters. John is also correct that there are many levels of adopters, some who just want to boot it and use it, some who want to comment on it, beta testers, developers, etc. We come from a wide variety of commercial and non-commercial interests.

But all of this is somewhat far afield from Ryan's original point. I'm going to restate it here, and probably paraphrase it very badly, but he was saying that GNU/Linux can't be counted, and the numbers wars of marketing are just a shell game. My point was that market share of servers has had little trouble counting Linux boxes, because they count iron installed. In the desktop market, the same thing is true. Counting iron installed, there is very little market share for GNU/Linux (<1%).

The other point is that Windows Vista is failing badly. People just don't like it (Although like many, I live in a Windows/Mac OSX/Linux world, and it works just fine - on brand new, state of the art hardware). There is an opening, and the community has brought the evolution (my favorite mail client) of Linux to the point where it not only competes with Vista - its better. I'm writing this on my Vista/SuSE 11.0 box at work, on the Vista side. I would use the SuSE side but I screwed up and installed KDE 4.0, which is NOT ready for primetime - yet. After the comments and improvements of the community, it will either evolve into something useful, or the community will choose something else.

As a former zombie coder, I resent the implication that hacked together monsters are not just as good as the "perfect" ivory tower logical poems. I come from the engineering side of things, and working now is better than a design still being implemented six months from now. Work now. Polish later. And with a huge community of people to polish, GNU/Linux has a bright shiny body indeed. Something that even the ivory tower crowd has to admire.

Ryan Cartwright's picture

But all of this is somewhat far afield from Ryan's original point. I'm going to restate it here, and probably paraphrase it very badly, but he was saying that GNU/Linux can't be counted, and the numbers wars of marketing are just a shell game. My point was that market share of servers has had little trouble counting Linux boxes, because they count iron installed. In the desktop market, the same thing is true. Counting iron installed, there is very little market share for GNU/Linux (<1%).

I did state my point several times above your comment - you could have just copied it ;o). Check my response to MrHasBean above.

As for your point about how the market counts servers (and can count desktops) I say they can't. I'm sorry I don't get the reference to "iron installed" - perhaps it is cultural?

craigTFD's picture
Submitted by craigTFD on

But it is a worthwhile discussion, no? The point I'm making, despite protestations to the contrary, is that it WILL be counted. Every piece of iron, every box, runs something. Those somethings are counted. With web servers it is pretty easy, if they are running IIS you know it isn't a Linux box. If it runs Apache, the question is harder. Our software is better because it runs in all the environments. But the final point is that environment has to exist somewhere. There has to be a metal box with some impure sand to make the magic incantations go round and round. It doesn't make any difference how many times it gets downloaded, just to sit in an ISO file on a rusty metal plate (hard drive). It doesn't make any difference how many coffee cup coasters (CD's, DVD's) contain that ISO data. It only makes a difference when the magic incantation is performed in the metal sandbox. And sandboxes CAN and WILL be counted, whether you like it or not.

Let us do what we can to make those numbers as high as they should be.

Terry Hancock's picture

"Every piece of iron, every box, runs something. Those somethings are counted."

Other than the fact that I post here and tell you, you have absolutely no way to "count the iron" in my household. Nor do I really want you to.

I have, as I've stated elsewhere on this site, four desktop PC/workstations running Debian. Every one of them is some kind of "whitebox" that I assembled from components.

Two more systems that I inherited are going to have either Debian or some other free software OS running on them soon. These machines were sold with Windows pre-installed on them. So Microsoft counted them as "Windows" sales, but they aren't running Windows now.

Now how common is that? I have no idea, but I don't think you do, either!

There's a lot of this world that is running on old, second-hand, or hand-built computers. Windows is simply awful on those systems, and installing Debian Linux on them is a good deal easier, as I can attest from experience with both.

I mean, do you realize just how much computing power has been sold at dirt-cheap prices due to companies retiring and replacing their old systems every year or two? What do you think happens to all those systems? People like me buy them and put them back to work, that's what -- waste not, want not.

If you try to make projections based solely on the corporate market, you're missing a lot of the action. That would be like trying to figure out what cars were driven most by looking only at company fleets or using only the statistics of new cars sold.

For me, as a "do-it-yourselfer", free software is a complete "no-brainer". This is the "do-it-yourself" way to set up computers. Windows is not friendly to my demographic. I don't know just how many people are in my situation, but I'm reasonably sure I'm not alone.

Trying to use the same kinds of figures that people use for proprietary software is a serious mistake. No one should care how much money is made selling Linux. We need to care how strong Linux is as a platform, and that isn't even a direct function of how many instances are installed, let alone how many are on commercial contracts. It's a function of how many active, contributing members there are to the projects that support the software.

There may be useful statistics for evaluating the success of Linux, but they aren't the same ones that proprietary producers quote. So, making direct comparisons will always be suspect.

craigTFD's picture
Submitted by craigTFD on

"Most of the average users I know would rather buy a new PC than upgrade Windows"

IMHO, SuSE 11.0 with KDE 3.5 is much better that Vista. More compatible. Faster. More applications. More games (yes, I use Cedega), and a better gaming experience.

If these boxes were widely known and widely available, we wouldn't have to talk about Vista anymore.

Dell, Wal-Mart, ASUS, slowly we are seeing Linux boxes offered as the first choice, rather than an aftermarket upgrade.

The age of dual boot is coming to an end. The age of Linux as first choice has arrived.

craigTFD's picture
Submitted by craigTFD on

If you want to be relegated to the hobbyist shelf at the bookstore, that's fine. Many people build cars from scratch, many of them never street legal. Those cars don't ever get counted in marketing surveys, but they bring new technology into the marketplace that helps all cars become better. Next up are the "kit cars", which are still constructed in a hobbyist garage (see http://www.dmv.org/wi-wisconsin/custom-built-cars.php). Perhaps you live in Florida, and are used to your vote not counting. Neither excuse works for me.

I use Linux. I am proud of that fact, and I want it to be counted. I too have assembled boxes that have never had a Windows Sticker on them, and I have converted boxes that do. Too bad that there may be others like us, whose many Linux boxes will never be counted by anyone, never matter. Of course, with Linux "market share" down below 1%, it wouldn't really matter if we got an accurate count and doubled our numbers. And don't kid yourself Terry, the vast (99%+) number of those refurb and off lease systems don't run any version of Linux, but some version of Windows.

But even more heartening, is the idea that those MicroSerfs will never be liberated, that they will have to spend their entire lives in service to Bill and Steve. You don't seem to care if no one ever makes a dime off Linux. I take the opposite view. I want there to be billions and billions made off of Linux machines and related services. I want people to have a better operating system, a better computing experience, to use the best software on the planet. I want people to be free, especially in a civil rights sense. Perhaps I have misunderstood you Terry, but it seems that if Linux "makes progress" on its own, it really has no other reason to exist.

Don't kid yourself Terry. Boxes matter. Numbers matter. Linux matters. And Linux should stand up and be counted. Linux should lead the world. Linux should succeed. And everybody will be better off because of it.

Terry Hancock's picture

Okay, the car marketplace analogies aren't working too well for you. Let's consider a different example: wine and water.

Any comparison which is based on "money made" will inevitably favor the much more expensive product, even if it is far less important. After all which makes more money: selling wine or selling water? Which one is better for you? Which one is actually drunk more? Which one is critical to your survival? How many bottles of water did your municipal water supplier sell last year?

If you looked at the "market share" of wine versus water, would you have any chance of predicting which is actually drunk more?

Do you see how those comparisons could be meaningless or pointless? Likewise, comparing GNU/Linux and Microsoft Windows successes in terms of sales or even total money made is nonsensical.

Your suggested approach -- counting operating systems delivered on turn-key systems sold -- would miss enormous numbers of operating system deployments. It's like trying to estimate water drunk by examining sales of bottled water.

Now, I'll admit that you're probably right that most installations on second-hand machines are Windows -- but you have no evidence for it. Just hear-say. There's no justification for faux quantitative statements like "less than 1% 'market share'".

How would you get that number? There's no way you can know it, which is what I was trying to say. There are ways to estimate for servers, but for workstations and PCs, it's a very difficult task (one possible estimator is to look at browser strings from clients visiting websites -- but there are all kinds of sampling problems with that approach, too. For example, you'll tend to get very different answers for different sites).

The tests of that kind that have been done suggest much larger figures for free software browsers and GNU/Linux operating systems: more like 10%, not 1%, IIRC.

A more important objection is that Linux does not need a large "market share" to be stable. It is quite possible for anyone who wants to run GNU/Linux to successfully do so, even if it does have a very small share of the total number of deployed operating systems.

Because GNU/Linux costs (vastly) less to develop (due to the advantages of openness and user-participation in development), it needn't make as much money to succeed as more expensive proprietary products. Thus, comparing the money made from GNU/Linux to the money made from Windows is highly deceptive.

Trying to directly compare sales figures props up a false impression that sales figures ("money made") is as important a metric for free software as it is for proprietary software. Proprietary software companies die without sales, but for free software projects they are optional and have little bearing on the continued stability or availability of the software.

You shouldn't imagine that the fact that the free software marketplace turns over less money is in any way a bad thing -- in fact it's part of what makes GNU/Linux great: it is vastly more efficient.

Author information

Ryan Cartwright's picture

Biography

Ryan Cartwright heads up Equitas IT Solutions who offer fair, quality and free software based solutions to the voluntary and community (non-profit) and SME sectors in the UK. He is a long-term free software user, developer and advocate. You can find him on Twitter and Identi.ca.