With all of the discussion of whether or not “GNU/Linux on the desktop” is a dead issue or not, it’s unclear whether anyone’s taken a close look at what the inhibitors to adoption have been.
Here’s where the stridently “free” contingent is at their least useful: the usual argument is that, since they find Linux systems perfectly usable—not noting that they have a self-serving interest in finding them so, in support of their own philosophical stance—they feel that it should be somehow self-evident to the world at large that their choice of platform is an inherently better alternative. When it fails to be self-evident, or at least that obvious fact is somehow not acted upon, free software advocates have regularly blamed Microsoft—although Apple is becoming a much more popular choice lately—for leading some sort of conspiracy to “keep Linux down”.
Sadly, the assessment these folks of “self-evidently better” is on the axis of “promotes software freedom”, which is not a critical value to most actual computer users. Those people are thinking more in terms of “Am I going to be able to do the things I want to do?”—like listen to mp3′s, watch streaming video, surf the Web, read their email, etc.—”Am I going to have to spend a lot of time learning something new?”, “Is there something I can do with it that I can’t do with what I’m currently using?” and “Is it difficult to use?”
“Regular users” are not interested in their ability to read the source code, use it “for any purpose”, fix it themselves, learn from it, or improve it. They’re not developers, nor should they be, any more than people who use toasters should be electrical engineers.
Unfortunately, that simple realization—that most users are not “hackers” or “coders”, don’t want to be, and shouldn’t have to be—runs completely afoul of the viewpoint of a significant (and noisy) contingent of the “free software community”.
There’s a strong “hobbyist” mentality in the community that—while it hasn’t impacted servers, where “user experience” doesn’t mean a lot—has had a significant negative impact on uptake of Linux-based desktop OS’s. Since hobbyists are creating mainly for themselves and people like themselves, there’s more of a premium placed on “clever hacks” than on things like stability, usability, etc., much too often. Since there’s no “project management”, per se, what gets done is what people are motivated to do, which is frequently not what would need to be done to better promote the system as a whole.
At its worst, this “hobbyist” mentality manifests in outright contempt for “regular users”, who are felt to be “n00bs”, “stupid” or unthinking sheep being fleeced by Steve Ballmer and Steve Jobs. There’s no recognition that these computer users may be talented and productive creators in their own areas of expertise, and are simply using a computer as a carpenter uses a chisel. Most carpenters know a good chisel from a bad one, without a deep understanding of metallurgy, and without particularly caring about the politics of the chisel’s maker.
Organizations like GNOME continue to attempt to “market” free software on the basis of its inherent freedom-enhancing qualities, which is something like trying to sell electric drills based on the color of the extension cord. Just as people don’t really care about electric drills but rather about holes, people don’t care about computer operating systems, particularly: they care about creating documents and spreadsheets, using the web, and reading their email.
The community-developed solutions available on Linux, while quite usable in many cases, are generally no better than proprietary alternatives, and frequently worse. Thunderbird is, at best, as good as Mail on OS X, and Evolution compares very unfavorably to it. You can make similar observations about GIMP and Photoshop, or Inkscape and Illustrator, or Open Office and iWork.
When advocates of free software say, “GNU/Linux is just as good”, they’re saying, “On its own merits, it’s not really good enough to justify the effort in switching to it.” To suggest that someone “ought to” use GIMP rather than Photoshop while simultaneously admitting that it isn’t as good, because it would “support free software”, is not an argument that’s going to get you far with a graphic artist who’s quite happy with Photoshop. It’s not a conspiracy: it’s a simple cost/benefit analysis, and for free software to win, it’s got to provide the benefits that prospective users care about.
Unfortunately, the understanding that free software has to compete on its own merits comes hard to the free software community. It runs contrary to their outlook—the software they write is largely for themselves, and their priorities are the reverse of those of typical users. It lets “the unwashed”—non-technical types—into their “club”, which for many, is a hub of their existence, apparently: by defining themselves primarlily as “free software advocates” and by associating almost exclusively with those with a similar outlook, they reinforce their own sense of being “special”, and only increase their resistance to the kinds of changes in viewpoint that would have been necessary to popularize Linux-based systems on the desktops of average users.
Worse still, that viewpoint is carried into the free software community’s attempts to engage with the mobile space. They point to “totally free” (but totally unmarketable, being utterly unready for consumers) efforts like GPE and Opie as “evidence” that there’s some sort of conspiracy, without realizing that a micro-Ubuntu-desktop, or something of the like, is the last thing people want on their phones. Here’s a hint: if you think a command line interface on a phone is a cool, or even an interesting, idea, you’ve got no business attempting to design a mobile device platform.
It’s worth recalling that “hacker culture” got its beginnings as a model railroading club at MIT. In many ways, its outlook has never progressed beyond that of a model railroading club. The inability and unwillingness to understand and even embrace the outlook and requirements of actual end-users of the devices they wish they could “target” more effectively is the single biggewst factor in keeping them from being successful.
Companies like Apple don’t suffer from this confusion: they understand quite well who their actual customers are and what they’re really looking for.
HP, understanding that a platform like webOS has a much greater chance of adoption than anything developed in the community, is reaching out to the homebrewers, a very smart move: webOS doesn’t seem to be very endangered by community-developed alternatives, for the reasons I’ve cited, yet some useful improvements could well come out of positive interaction with that community.
This is the future, it seems. Just as free software invisibly runs much of the web, it will—equally invisibly—run things like Android and webOS, at least from a “kernel” or “networking stack” perspective. But will it win as an end-user solution in its own right? It seems less and less likely. The “Year of the Linux Desktop” isn’t ever coming, and it seems likely that the “Year of the Linux Phone” won’t be either.
“Those people are thinking more in terms of “Am I going to be able to do the things I want to do?”—like listen to mp3′s, watch streaming video, surf the Web, read their email, etc.—”Am I going to have to spend a lot of time learning something new?”, “Is there something I can do with it that I can’t do with what I’m currently using?” and “Is it difficult to use?””
It looks like I might be installing Linux on a coworkers computer. Her copy of windows is hosed, perhaps by file-sharing, an old boyfriend or, act of child. I asked her what she uses her computer for and she stuck pretty close to your examples. Same with a few other coworkers who I’ve switched, at least till they replaced their old computers. Heck on the computer that I’m using right now, Vista broke when I dared to install a new video card without a internet connection available using the driver on the CD. No big deal I had started using Linux on my computers starting from not long after my old Mac was stolen only to find that Windows was still too dumb to read hfs+, and completely stupid in it’s “smart” way of handling drivers. Having used or experimented with Linux and open source on my mac, installing Apple’s file system on Linux was trivial for me. In fact Linux was a breath of fresh air.
So far I don’t know anybody, other than myself who uses Linux because he or she is dissatisfied with mainstream operating systems, but when Windows breaks, friends and coworkers often talk to me and, 99% of the time they are a good candidate for assimilation. So while I agree with your point on user apathy, for me Linux works better and, the people who I’ve talked into trying it didn’t seem to mind it either.
Incisive analysis – though you should probably be more clear with “the year of the linux phone” since webOS, Android, and Meego are all “Linux phones” (not to mention the millions of embedded devices) they just don’t use most of the traditional desktop components, for exactly the reasons you said, no one wants a mini ubuntu in their pocket. Or do you mean the “year of the linux community phone”, i.e. one not mainly put together by a single company for its own purposes.
There may never be a “year of” but that doesn’t mean linux doesn’t slowly gain traction – after all it does tend to keep getting better. The one massive missed opportunity was Netbooks – there was the ideal platform for launching linux, and its failure shows up all the points you made: no matter how great people thought their OS was (because it did everything they wanted, because they contributed the features they wanted) there was, in fact, not enough to make it compelling to end users.
Which is why Ubuntu detractors come across as kinda silly – they complain about how ubuntu is doing things, but at the end of the day its Ubuntu thats actually shipping on Dell machines, that has an attractive netbook interface that could genuinely go somewhere, and that most people would recommend to new linux users. So as a matter of observation, whatever Ubuntu is doing is better than whatever anyone else was doing to bring linux to end users.
I’m using “Linux” here as shorthand for “Linux-based operating systems using pretty much the standard complement of community-developed components”. That is to say, what the freest of the free want to insist that you call “GNU/Linux”. Weirdly—and after I’ve complained about the silliness of it all these years—it actually turns out that the term “GNU/Linux” has some actual utility, but it’s not at all what the FSF had in mind.
They imagined that it was necessary to deal with what would surely be a rising tide of various versions of GNU: GNU/Linux, GNU/HURD, etc. It’s not the “GNU” part that’s wound up being the common element, though, it’s the Linux kernel that keeps getting put under alternative user spaces. So GNU/Linux is distinct from Android/Linux which is different than webOS/Linux. The only one you mention that I believe would qualify as (mainstream-ish) “Linux” is MeeGo. Given the competition to MeeGo, its (apparently) being supported by just a single handset vendor, and the timing around actually getting MeeGo devices into people’s hands, I’m not all that optimistic about how it’s likely to do.
Mostly I agree with you, but there’s one major issue with your analysis.
First off, note that I agree with your entirely on your *starting point*: Linux on the consumer desktop is a tiny minority operating system and that’s unlikely to change before ‘desktops’ stop being mainstream devices again (about 9 years from now, I figure).
However, in explaining it you sadly fall over the hoary old ‘most people aren’t coders so software freedom doesn’t matter’ fallacy. Really? Surely you’ve been around long enough not to hit that one. Software freedom is important to everyone whether they’re coders or not, and whether they *know* it or not. Well, I should qualify that: if you’ve ever had any kind of bug in your software, or ever thought ‘well, this software doesn’t meet my needs perfectly in every way’, then software freedom is important to you.
The benefits of software freedom affect you even if you never exercise them yourself, because they make it effectively impossible for the developer of the software to act in a way that is good for them but bad for you, and because it massively increases the ability of those who didn’t write the code to enhance it and fix problems in it. Whether any particular end user happens to be able to do that really doesn’t matter.
After that, though, you’re mostly right; but your argument suffers horribly from getting up on your old freedom-bashing hobby horse. Software freedom is mostly orthogonal to this, which is the point most people in the debate – whichever side they’re on – can’t get their heads around. Linux on the desktop is a tiny minority interest for the same reason *every* OS but Windows on the desktop is a minority interest, whether they’re free or not – it’s not enough better than Windows to make it worth not using Windows unless you’re quite dedicated. (As a variant, you have to be quite dedicated to configure things such that Linux is better enough than Windows to make it worthwhile).
By the way, you don’t think a CLI on a phone is useful or interesting? That’ll be news to Google and HP, then. Both Android and WebOS have universal search, which is a basic CLI. In WebOS 2.0, you can type something into the universal search box, and you’re given options to publish it to Twitter or whatever else is contextually relevant to what you typed. What’s that if not a CLI? It’s just a question of how it’s presented. Sleight of hand is very useful in interface design…
I’d agree with most of your other points, but with some different emphases. I don’t think there’s anything particularly unique or interesting about ‘hacker culture’ or the observations you make on it. It’s a symptom as much as it’s a cause. Go and find *any* small community of enthusiasts and you’ll see all the same stuff. Go find people who buy obscure high-quality audio hardware from Taiwanese manufacturers or American boutiques and ask them what they think of Bose (or Monster Cable). Go ask an indie nerd what s/he thinks of the ‘big labels’. Go ask someone who reads high literature what they think of chick lit. Go ask an sf fan what they think of high literature. It’s all cliques railing against whatever they perceive as the mainstream and believing that the world would wake up if they just *understood* and paid attention (but at the same time secretly enjoying being a little clique), they all display much the same characteristics as you describe above, and it’s all very boring and psychologically predictable. I don’t think any of this stuff really presents a problem if the clique in question happens across something which actually would appeal to the unwashed masses; the nerdiness of the sf/fantasy cliques never affected the success of Harry Potter or Avatar or the LOTR movies or whatever. (What tends to happen is whatever became a success gets written off as some kind of betrayal, pandering to uneducated mainstream values, and the clique gets right on with hating it too). If someone managed to write an incredibly awesome free software desktop it’d become a mainstream success and probably someone would come up with some reason why it’s not ‘really’ a proper F/OSS community thing and should be exiled, the clique will carry on as normal, and no-one else will care. What I’m getting at is that the typical small clique elitism bullshit you identify is mostly background noise.
I think free software developers who actually want to develop for mass market appeal understand what you’re saying, and the cases where it doesn’t are mostly due to a) insufficient resources or b) developers who do understand this and don’t care, they’re explicitly *not* developing for mass market appeal. They’re fine with being a small elitist clique and they’re writing for the clique. There’s nothing intrinsically wrong with this. I’d say the ‘free software desktop’ as a whole is mostly a case of a). I reckon that KDE and GNOME as projects, and most of the distributions which actually aim to be suitable for general public consumption, understand what you’re writing here. They just haven’t *succeeded* yet, mainly because they don’t as a whole have the resources for it, and I don’t see it happening in any sort of near-future time scale.
(BTW, I haven’t seen anyone refer to Opie or GPE for at least five years. Anyone who still does is terminally deluded and should probably be committed to some sort of institution. If you want a vaguely F/OSS phone environment you should be cheerleading very fucking hard for Meego, as it’s the only hope.)
“Software freedom is important to everyone whether they’re coders or not, and whether they *know* it or not. Well, I should qualify that: if you’ve ever had any kind of bug in your software, or ever thought ‘well, this software doesn’t meet my needs perfectly in every way’, then software freedom is important to you.”
Except that, as the end user stack level, open source software tends to be more prone to problems than most of the products surrounding us that use proprietary software.
At the very least, there are distinct gaps in both functionality and usability in many open source components versus their proprietary equivalents.
That’s a ridiculously broad observation and I’d be very surprised if you could substantiate it in any rigorous way. There are literally hundreds of thousands of both open source and proprietary codebases covering everything from fart apps to stock exchanges. It’s absurd on the face of it to imagine there’s much you can say that’s true about *all* open source, or *all* proprietary software. They’re just development methodologies, in the end.
The development methodology of community-driven projects has philosophical and political components that quite frequently drive projects to make counterproductive decisions. This in turn results in OK solutions that are rarely optimal. The Linux desktop experience, for example, is defined by a hodgepodge of technologies created by different groups of people that do not necessarily share the same vision or goals, and this shows in the end product in the form of lack of polish, suboptimal quality and unnecessary complexities.
koki: better effort, but still silly. Kicking the Linux desktop around is an easy game, but look at it this way: there’s exactly one company which has ever built a graphical desktop environment that is a ‘success’ by measurement of market share, and that’s Microsoft with Windows. And I think it’s pretty easy to argue that Windows is also subject to “lack of polish” (if you get to take an easy shot, so do I – Start button to shut down, anyone?), “suboptimal quality” (boy, howdy) and “unnecessary complexities” (how many device management screens does it have now?) As I said, it’s easy to kick around the Linux desktop, but who’s done any better? Microsoft built a just-about-good-enough product that came to dominate the market through various historical quirks and shady business practices, Apple built its own quirky walled garden, a bunch of others came, tried and failed (RIP, OS/2, NeXTStep, BeOS et al), Linux has a toehold with geeks and enthusiasts because no-one with enough resources has ever seen any reason to take a really concerted shot at Microsoft when there’s easier opportunities out there, and it’s been that that way for years. I think it’s pretty hard to take the desktop operating system area as a good place to compare software development methodologies.
To counter your easy examples, I cite the F/OSS world’s easy examples – Firefox and Apache. And then we go ‘well your example’s an exception’, ‘no YOUR example’s an exception’, and blah blah freaking blah forever. There’s zillions of reasons people write F/OSS, freeware, or paid proprietary software, there’s zillions of different approaches to writing each, and zillions of apps of each type out there, and we could probably cite good and bad examples of each type from here till doomsday. What would be the point? It’s perfectly possible to write good and bad software with any methodology, and I’m not sure it’s worth worrying about any grand headline comparisons beyond that.
While you’re not saying anything that’s factually incorrect here, Adam, what you are saying begs the question a bit.
There are plenty of “free software advocates” arguing that GNU/Linux should get more share on the desktop than it does, many of them blaming various nefarious conspiracies, more often than not orchestrated by Microsoft (or by companies that have been “infiltrated” by Microsoft “entryism”, in some of the most fantasy-laden versions) as an explanation why it’s failed to do so.
Apache’s not an excellent example of a user-facing F/OSS success: it’s a server room technologies, no normal end-user runs their own Apache server. Firefox can be largely viewed as an anomaly, or at least a strong argument could be made in that direction. From another angle, Firefox is one of the very, very few open source projects that has achieved a level of not only functionality, but “fit and finish”, that’s competitive with proprietary alternatives.
(And “quirky walled garden” comments aside, Apple makes a ton of money, and has a ton of happy customers, so they’re arguably doing something right. They also have a ton of pretty happy developers, many of whom make not just livings, but pretty sizable amounts of money from their trade.)
“There are plenty of “free software advocates” arguing that GNU/Linux should get more share on the desktop than it does, many of them blaming various nefarious conspiracies, more often than not orchestrated by Microsoft (or by companies that have been “infiltrated” by Microsoft “entryism”, in some of the most fantasy-laden versions) as an explanation why it’s failed to do so.”
There are nefarious conspiracies, and they’re part of the reason. It’s pretty hard to disagree if you do even cursory reading on what the ‘nefarious conspiracies’ actually entail. But they’re boring everyday conspiracies that happen in monopolized markets (of which there are many) every day. (They’re also not exciting dramatic cloak-and-dagger, burn-after-reading secret conspiracies, just dull middle manager types doing what dull middle manager types do, and things that – from their perspective – probably seem perfectly normal and utterly unobjectionable). Microsoft takes advantage of its dominant market position to impose all kinds of conditions on hardware suppliers that they’d never accept if they were in a strong bargaining position. This is clearly not ‘fair’, but then, neither is just about any market in the real world, this kind of abuse is pretty much inevitable, and generally dull. It’s not really worth worrying about, because you’re never going to be able to do anything particular about it. It’s important not to confuse this with there not *being* any market abuse, though, because there really is. Whether Linux ‘should’ have more desktop market share is a pointless counterfactual: there’s just no point arguing about what ought to happen in a fair market because there are no fair markets. It’s a waste of time worrying about it.
When it comes to the mobile market the conspiracy is the network operators; they have a very strong interest in restricting consumer freedom, and as the market is *designed* to be a cartel (by the way wireless frequency is allocated in most countries – by government auction to a limited number of bidders), it’s absolutely inevitable that the operators will ‘mysteriously’ all wind up imposing the same restrictions on the handsets they ship (shovelware by the bucketload, no tethering, approved marketplaces, no root access, all that crap). Again, this is dull and inevitable and not worth getting all frothy about, really, but that doesn’t mean it’s not *true*. The reason Meego is the only realistic chance for a decently free handset operating system is because it more or less has Nokia’s clout behind it, and Nokia’s clout is (for another couple of years at least) enough to get concessions from the networks.
I never said anything about ‘user facing’, either. That’s your issue, not mine. I don’t care about ‘user facing’. The bit of your analysis I was mostly interested in doesn’t read as being in any way specific to ‘user facing’ software. Also, when I cited Apache, I was replying to Koki, who was making a very general argument about community-developed software, and certainly didn’t restrict the terms to ‘user-facing’ software.
“Firefox can be largely viewed as an anomaly, or at least a strong argument could be made in that direction. From another angle, Firefox is one of the very, very few open source projects that has achieved a level of not only functionality, but “fit and finish”, that’s competitive with proprietary alternatives.”
See? Just as I predicted. Your example’s an exception, mine’s an exception, blahblah. =) I think one thing a lot of people fail to consider is that lots of proprietary software is truly fucking terrible as well. The rule that 90% of everything is crap applies strongly to software. Almost all software is crap, it’s no real surprise that that applies to open source as much as it does to proprietary software. I tend to find my computer-using experience enjoyable in direct inverse proportion to the amount of different programs I’m running, whether we’re talking open source *or* proprietary. I can cite terrible proprietary software all afternoon, if you like, but I’m not sure I need to. Most of the proprietary software that’s really popular is shit. Windows is shit. Office is terrible. Acrobat is hideous. Oracle is just about the worst thing in the history of things. As an industry the most remarkable thing about the software industry is that it’s come to be so fundamentally important to the world while being so, so, so horribly bad. I think this is something it’s very easy to lose track of when you work in the field. But to take the hoary old car analogy: millions of people actually *enjoy* driving their cars, even when their car is fundamentally a tool that allows them to do other things, such as get to work every day (or actually travel around while at work), and they’re certainly not ‘car geeks’. How many non-geeks do you know who actually enjoy using most software (proprietary or open source), versus grinning and bearing it? Yes, even the Apple users?
“And “quirky walled garden” comments aside, Apple makes a ton of money, and has a ton of happy customers, so they’re arguably doing something right. They also have a ton of pretty happy developers, many of whom make not just livings, but pretty sizable amounts of money from their trade.”
Yeah, what Apple does wrong is at a more fundamental level than economics. If you just judge by that yardstick – if it makes money, it must be right – then sure, Apple’s great. So are cigarette companies. I loathe Apple for the same reason I loathe cigarette companies: what they do makes lots of people happy in the short term, and makes Apple lots of money, but is fundamentally and in the long-term deeply negative for the general wellbeing of society. (Yes, I probably think about some things too hard.)
Adam: I use open source software and I agree that there are plenty of success stories. I did not mean otherwise (though I can see how it could have come across that way).
My replies were meant to address specifically the statement that I quoted on my first reply, the “if you ever had a bug in your software, then open source is important to you.” This mantra is flawed IMO, as there are many different ways to produce both good and bad software, and open source does not guarantee good quality as many advocates fervently predicate.
As an end user, I look at software as a tool and I am more concerned that the tool does the job properly and not whether its source is open or not. In my experience over the years — been using a very wide array of personal software since the 70s — closed source tends to be more refined, better suited and of better quality (not that all of it is).
When I brought up the hodgepodge argument, it was merely an example of how the techno-diversity of the Linux ecosystem can play against it in terms of coherency (lack of thereof) and a lack of overall direction and/or vision. Admittedly, that’s where I digressed.
All being said, I use Ubuntu, and I think it has come a very long way and is a good desktop. I still have to jump back into Windows when I need to get real work done, because the open source alternatives for the tools that I use are inadequate (specifically, Inkscape vs. Illustrator, GIMP vs. Photoshop, Scribus vs. InDesign, etc.).
You seam to miss the main point of the article. It is not saying that “freedom” is not important, is saying that as a “selling point” is not just enough. Open Office is cool (using 3.2 right now on a Ubuntu 10.04), it does ALMOST everything that MS Office 2003 does, so 9/10 people would probably use it instead of MSO and be just OK. BUT and this is a big BUT it DOES NOT do everything MSO does, specially when comparing with excel, and for some reason i don’t understand why, it looks outdated, I mean is really that hard to make it look a little more sleek?
So why switch? the product is not better (except is 100% free both as in beer and as in speech), it looks old and on top of that i have to learn it, when i already “know” the other. just look at the backlash that Office 2007 got with that silly ribbon.
First, there is and isn’t a conspiracy against Linux. it isn’t a conspiracy because it’s just Microsoft behind it så that defines it as competing, not a conspiracy. It is a conspiracy in the ‘behind the scenes’ kinda way. What Microsoft does is forcing companies like HP, Acer, Asus, Dell a.s.o. to keep away from Linux by using licensing ‘blackmail’. It’s simple. Microsoft say’s, -”If you offer computers with preinstalled Linux intended for the private sector we’ll charge you xx% more per Windows OEM-license” – and that’s the end of it. That’s why HP, Dell a.s.o. offer Linux on the business-segment but not on the consumer market.
Why MS don’t do the same in the business segment is simple, it’s about the customers. Today many states and organisations demand open source, sometimes it’s costbased, other times law – as som contries have it in Europe today. Not offering Linux to this market isn’t an option. But for consumer products, they’re doing it.
Too bad with the GPL3 restrictions though, if HP or Dell could I’d bet they would have made they’re own Linux-based ‘OS X’ a long time ago and I bet the market would look a bit different. But as it’s to hard to do it and/or make money on it, well, it just never happened – or will, I guess… So no breakthrough.
And I might be wrong about the above, but I don’t think so – having a few years in consumer computer sales behind me – and I know I wouldn’t have any problems selling Linux-based systems today… It’s all about covering needs. You do that and the customer is satisfied. As simple as.
Without knowing the specifics of your claim, it’s difficult to be sure, but that sounds like anticompetitive business practices to me. Do you have any concrete substantiation that Microsoft threatens to substantially raise prices on Windows if OEMs offer pre-installed Linux OS’s to consumers?
They used to do things like that plenty before Microsoft was sued as a monopoly in the late 90′s. I don’t know of any evidence that they’ve done it since the settlement. This could be pure FUD.