I've been thinking quite a lot about Linux's future, here's what I've come up with, and why I'm worried.

The majority of Linux's community is technologically aware, we generally know how to operate our machines, and in many cases repair them if they go wrong (or at least start to understand how to repair them if they go wrong) but what of the future? Linux's marketshare is growing, it's no longer something nobody has heard of, but something that somebody who is anybody knows about and probably, uses. What if we get a continual influx of stupid users?

Granted, computers are becoming more and more esstional to modern life and more people (especially kids and teens) know how to set up a wireless Internet connection, yet they themselves can hardly be called geeks. (On an interesting side note, the old sociological model of geeks vs. non-geeks is slowly dying, there was a time when you were either strong, or smart. It's both, now, and everyone is becoming a geek)

So, the problem of stupid users may well go away by itself (much to the celebration of tech-support folks)

But what of another problem? One that Linux can only play second violin to?

Hardware.

Linux is not centrelized, we have no hub of software development, no multi-community app store...and no hardware manufacting devices, we cannot, for example, create computers to run with Linux on the same scale as Apple does, but there is an even bigger problem here;

Project NATAL.

If your a gamer, especially if your an Xbox360 gamer, then you should know what project NATAL is, if you don't, read the next paragraph.

Project NATAL is bit of hardware that will attach (It's not released yet) to Xbox 360's and Windows computers. It can recognize emotional expression, gestures, voice....and it uses a special form of motion capture, though you wont have to wear ping-pong balls or have multiple cameras. It only uses one. You could talk to a character in a game, watch them cower in fear when you look angry, or better yet, play a racing game without needing to go out and spend £40 on an extra controller shaped like a steering wheel.

Gesture-based computing is a very important technology. it saves on handsets for starters and has a particularly large degree of freedom with games, photo-modification, information analysis, medical computing...yet Linux will not be able to match it to that extent. I'm sure, if it proves popular, or if someone gets really excited about the idea, then it could be done, but whilst we can give out software under FOSS, it dosent quite work for hardware, does it? *sigh*

On the other hand, touch-based computing has a lot of possibilities (maybe even more then gesture computing) and I don't think gesture based, whilst good for a lot of things, can replace a keyboard (I'm particually looking forward to mainstream touch-based computers, it'll save me hours of cleaning the keyboard...)


Going back to NATAL, as I said, it can understand speech. We have no non-experimental means to use speech. (Yes, there is Simon, but that, whilst good, is very basic in comparison to most systems out there, there is also CMU Sphinx, but that is nowhere near the amount of user-friendliness it needs) Speech tech is very important, we only now have the processing power to do it properly and accurately. Imagine speaking out an email, or document, when I was on Windows, I did that all the time. Very useful technology, even more useful to disabled users...

It's my own personal opinion that until the community can make a push forward (and NEPOMUK is going in the right direction) for newer technology; to acutally INNOVATE and INVENT...rather then clone or copy...then Linux will always lag behind.


Another example; When has anyone tried to produce a fully featured music production suite? Don't point me towards Audacity, thats a toy compared to the more professional stuff, what about Apple's Logic? Have we anything for that? Nope.


That concludes my dire warning of the far future. Goodnight.

-Dan

Hello there,

I have a similar observation, but a different take on the outcome. Survivability isn't determined by acting on new technologies. Not in nature and not in computing. Linux's migration into new technologies is slower, this is true. But the development process, while slow, will outlast all proprietary development and "new" or disruptive technology.

I see your point of view. But that is on an immediate time-line. Think longer term. Wars. Catastrophe. Governments and nations rising and falling. Corporations going bankrupt. While Internet infrastructure isn't new tech or even the best tech, it is modular/stable and can be run by anyone anywhere. Nukes can wipe out California and Washington state, and Linux and the Internet would go on as a whole. Apple and Microsoft would not.

From another perspective, the server side of things have moved towards Linux. Whether it's game servers, or web servers, or database servers. Linux isn't the exception, it's become the rule.

And I think there are more important new technologies happening now with Linux than other platforms. Take ksplice for instance. Once it evolves, the ability to sustain uptime for hand helds, desktops, servers, and storage systems will improve dramatically.

In any case, I know that it looks bleak from the end users point of view. But that will change over time as the numbers start to tip the scales on the desktop. My opinion is that it will happen, although I think Mac is slowing the progress due to competition(thanks BSD). It may take 10 years. It may take 5. But once the saturation point is reached, the applications will be there.

I also wouldn't worry too much about, "stupid users". The repository system ought to become the single, major location to install software (free and proprietary pay-to-use), and therefore only trusted software would be included in it.

There is real innovation going on in Linux, it is just in different directions from other proprietary companies. I do have to agree that there are a number of technologies that Linux adopted from Microsoft and Apple, but there are others where Linux leads the charge.
Of course your note is about hardware; Microsoft and Apple are both looking to make money with their products, and they both see $$$ in producing innovative hardware. That's fine, they can do that all they want, and when they release it Linux will adopt it too. Just google "Wii Linux" or "Wii remote Linux" for a recent exampe of Linux adapting to new hardware. Linux does this all the time; granted it is playing catch-up in this space because it is waiting for products to be released so that it can produce the required drivers.
There is one simple reason why Linux won't go the direction of driving hardware innovation: Linux is an Operating System - a piece of software which runs on a piece of hardware. Linux doesn't care about pushing innovation in hardware because that's not its purpose/goal. Its goal is to drive innovation in the operating system space. The Linux perspective on hardware is: does the operating system run well on/with that piece of hardware, or how can we make the most of the harware that we have.
On the other hand, there is certainly plenty of room for companies/groups/whoever to enter the realm of hardware design and develop innovative new hardware. If Red Hat or Novell or Canonical or some other Linux-based company wants to go that direction, that's fine and they will certainly get plenty of support for it. Of course there are other companies besides Microsoft and Apple driving hardware innovation, for instance HP, Sony, Visio, to name a few.

Possibly, I agree that from an end-user point of view, there is little innovation...sure, we may have a very robust filesystem, Ksplice, Plasma, NEPOMUK...but we are still seriously lacking in those fields which are far more relevant to the end user.

Look at GNOME, for instance, how many of it's popular software is not directly copied from Apple?


I also agree that Linux's FOSS development model is better over closed source; the closed source model requires not only a good idea and motivation to make that idea reailty, it also requires money, popularity, and a good public image. In the FOSS world an idea can go beyond it's creators life. This model may prove more difficult to hold, the larger scale projects are getting harder to reprogram and build (look at NEPOMUK, that will require vast amounts of developers, time and to a good extent; money, before it can do anything. Such projects can only get larger and grander, and indeed, more difficult to make.


We ARE fragmented, though, this is another problem; we have RPMs, debs, tar balls....no easy way to convert between the former two (which seriously affects how much software is avaible to the community) and the latter requires people to have more knowledge then just how to use a keyboard (why oh why it's compilation can't be automated I have no idea)


I've got a friend who is in music production technology, they all use macs; why? Because Windows has a far higher TOC, and Linux dosent have the applications for music production.

As for hardware, yes, thats true. My main concern is, however, that we can't really pull strings for hardware, we have to work really hard to get Linux to function on it, or hope whoever designed it made it Linux-friendly.
My note there was that the Linux community can never really sell, say, a multimedia home server, or something like Project NATAL, sure we can provide instructions on how to build a computer

Same with video production, DJ software, etc etc.

Just leave your desktop and expand your horizons. Whatever happens to the desktop, Linux is here to stay on supercomputers, as part of the infrastructure of the Internet, in digital televisions, in satellite navigation systems and in mobile 'phones. There would not have been the huge increase in contributions to the kernel if companies and individuals did not think that Linux has a future.

More fundamentally Linux is part of a shift in the way development takes place. In the 80s people learnt how to program on home computers; in the 90s people, often reluctantly, moved to DOS because that was where most games development was taking place. But the FOSS way of development gives any young programmer the chance to create code, have it peer reviewed and if it is good enough used by millions of people some of whom will give feedback. Because your name is on the code that is your reference for a job. You simply cannot get that sort of experience with a proprietary company and you risk losing your network unless you join a company that supports FOSS.

Proprietary companies will find it harder and harder to get first class programmers; the quality of their products will decline relative to FOSS products and they will only be able to survive either by adopting the FOSS approach of by retreating to a niche which is too small or specialised to gain the interest of FOSS programmers.

I know little of the programmers perspective;

Under a closed-source base, any code written will belong to the company (yet it's faults are yours to correct, odd, no?) you may not speak about the code's development, it's function, purpose...in many cases, it is not improved upon, simply scrapped and rewritten by someone else five years later. Needlessly reinventing the wheel. Delaying progress time and time again.

I am also aware that the idea of FOSS is spreading. It has been a great misfortune to work with two people who felt FOSS was, less then good.

One of my projects involved assisting someone with a chatterbot. However, he refused to use any FOSS technologies because of the GPL agreement, you must give back any changes.
He also refused to use others work (such as OpenCYC, or DBPedia) , even if it used a more lax license, preferring to do it all himself. I gave up on him after he moaned about how much work he has to do, how I'm not helping him (Incorrect, I had offered help many times, he refused each time)...and his unwillingness to use services, like Twitter, that would of helped his chatterbot build up it's database.

Another was a programmer in SecondLife, a virtual world. He told me that due to changes in SecondLife, he would be very shortly leaving it. This man had a large amount of customers (he built props from a science-fiction show) and was the 'premium' creator of his content. I asked him to release his code and textures used for the creations. He refused, simply because he wouldn't make money from it. I pointed out that if he left, he would also stop making money. He ended the discussion there. *sigh*


There was a book that was quite popular about a year ago, called 'After the Software Wars', it was written by an ex-Microsoft programmer, who foretold that Microsoft was doomed (mainly because of it's antiquated codebase and increasing unfriendly tactics) and that Linux was the future.


In 30 years time, I would like to live with technology whose creators are not squabbling over patents, pushing the price tag up or seeking to actively defile and destroy each of their products.



But, back to the point; yes, I see where you are going, supercomputers down to DVR's can (and most run) Linux. On a related note I'd like to get a reference to Linux's server-share to settle an argument with a friend.
Regardless, however, it is my feeling that we are doing something wrong if we have a higher TOC then Apple computers.
It is also my feeling that we can never have a modern computer without sorting out the whole mess with packages and tar balls (I've spent almost a year on Linux, and I STILL havent learnt on how to install a tarball)

Dante Ashton wrote:Regardless, however, it is my feeling that we are doing something wrong if we have a higher TOC then Apple computers.

Not sure where you get this idea. Please explain, I'm curious.

Dante Ashton wrote:It is also my feeling that we can never have a modern computer without sorting out the whole mess with packages and tar balls (I've spent almost a year on Linux, and I STILL havent learnt on how to install a tarball)

Well, the mess with packages and tar balls Has been sorted out for the most part. On any modern distro for a large number of apps, you can just open that distro's software manager, search for the package, install, done. If the app isn't in distribution's repositories then you need to resort to installing the tar ball, which generally requires unzipping it and following the instructions in the INSTALL file (or readme if there is no install file), which is usually (in a terminal) configure, make, (as super user) make install.
I know of only one graphical app designed to do all this for you called Kompile, but it was for KDE 3.5 and I don't think it has made it into the world of KDE 4, yet. There may be others, and there may be a way to set up a service menu to do it, but that's as far as my working knowledge goes.

Of course, there is always the question of why there are tar balls and debs and rpms and not just one single unified 'package', and that gets a little complicated. I think that ideally the tar ball should be that single unified package since that is how most software is distributed (the developers put together an app and bundle up the source into the tar ball for general distribution with the expectation that distributions will create the packages necessary for the individual distributions, which is a process in and of itself). Having the source tar ball be that one format is that they lack a number of things including dependency management and binary packaging (which means that a package can just have its contents extracted to specific locations and doesn't need to be pre-compiled before use).
Another ideal would be for some distribution(s) to provide repositories with all packages, but that would be an immense effort and incredibly difficult to provide support for. The biggest distributions are about as close as it gets and they are constantly growing their package base, but the number of projects out there is also constantly growing making it impossible.
Yet another ideal would be for the various app developers to make ready packages for a number of major distributions, but that's never going to happen since those developers are already putting all the effort they can spare into those projects, expecting them to create (and support) individual distributions' packages would cause development on many projects to grind to a halt; not to mention the duplicate efforts which are streamlined by having the distributions create those packages.
So, for the time being we have what we have and we learn to deal with it. For a vast majority of users what's in the repos is sufficient.

Dante Ashton wrote:I've been thinking quite a lot about Linux's future, here's what I've come up with, and why I'm worried.

Well, i am both agree and disagree with you. My points are:

1. You described some things that involve some politics. Well, it may be good to get more power to linux, but it will cut freedom. Every platform/desktop is targeting it's own user base. Apple got design and graphics, M$ got gamers and offices. Linux has developers and advanced users. Will linux benefit from coming of another group into it? Most surely gamers and office workers will not make linux advance, they'll require to cut off any advanced features, sticking with the most basic set just to keep working without confusing themselves. It's better to take a look why linux users switching to Mac/Windows or what they have to virtualize with Wine/VirtualBox/VMWare. For instance, tablet support and Photoshop. Those are people who already like linux itself, but for some reason it's not enough for them. So I think there is no sense in stealing users from other platforms, it's better to make linux fans more happy. Compare it with religion people are stuck with their religion, they will hardly switch. But they can or cannot be happy with it.

2. Linux is just good kernel for my desktop. I need KDE and it's feature set for every day computer usage, and linux kernel is just best currently available platform for that. If something will happen to linux - i'll try GNU Hurd or BSD or something. KDE is something that gives you user experience, nut the kernel. KDE is the most advanced desktop currently available, and no single feature like voice control will change it.

3. And I am worried a lot about future of KDE. It's naturaly that KDE team want to make it popular, get wider user base. But to do it - they'll have to make it idiot-ready. Like GNOME. Being advanced user I am worried a lot about future of advanced software, and I am afraid of the day when KDE will give up with it's advanced features and will follow GNOME with it's idiot-proof user experience.

As for me being on the second place is not a bad thing, it's great thing! To be first mostly means to be dumb. For instance, Windows/linux, IE/firefox, C#/Java. Companies like canonical care about market share and developers are making idiot-ready software for money there. But that is the reason to love KDE - developers are making it for themselves, making it efficient, advanced, feature-rich, while employers of companies mosly never use software they're working on. Can you imagine the day KDE guys will remove virtual desktops because it's too confusing for lame users? If that day will come - it will be my last day with KDE. And I will not follow Linus path, I'll switch to Windows - you can make it a bit more advanced using shareware, but you can't do it with GNOME any way. As example - take a look at Amarok. Amarok from KDE3 was great music player, with advanced playlist configuration and great collection feature. What they did with Amarok 2? Great context feature that I never used before and never will, poor playlist organisation (until 2.1), and not configurable transparent indicator. It looks more like some music studio for music critics, but not for people who just keeps their music in background and want to same network speed for things other than lyrics and wiki pages. I almost started to write my own music "playlist with play button" in QtJambi, it's great that Amarok 2.1 arrived.

PS: For GNOME fans, just some points why I think KDE is more advanced than GNOME:
- global keyboard shortcuts. Yes, all my KDE apps are using F2 for Save.
- slideshow on wallpapers.
- I have 4 buttons in my desktop configuration: OK, Cancel, Apply, Defaults. You have only Close.
- I can switch between my 4 desktops by my own shorcuts, Alt+1,2,3,4.
- And of course I have Klipper, KWrite, Krusader and Kompare. Do you?

airdrik; I know three musicians, all of them have Macs. Whilst they know how to play a guitar, piano and have damn good singing voices they have less technical competence then a rather stupid bumblebee. I've seen attempts to get Linux to operate as a music production platform, it takes a large amount of technical know-how to do so, and often bypasses most sound servers and goes straight to JACK, hence the higher TOC.

Aekoid:
My main concern was the lack of innovation on the desktop, we have many features but in comparison many few that the end user experience that the community has not cloned from other OSes.

Dante Ashton wrote:My main concern was the lack of innovation on the desktop, we have many features but in comparison many few that the end user experience that the community has not cloned from other OSes.

I understand that, but I can't understand what kind of innovation do you expect? Most of those concepts like gestures or voice control are not a big step forward, it's just a minor feature that requires big development. You talked a bit about different kinds of professional software, and that i found important enough to care.

While it is not a bad thing to think about the future, I can't understand your reasoning.

Hardware: yes, there are companies which sell hardware and software, but also ones that sell either and still have no problem, e.g. Microsoft.

Then the example about some external device for XBox or Windows PCs. Why would it be impossible to use this with a Linux PC?
Microsoft most likely won't support it, but no support from the hardware vendor hasn't stopped other devices from becoming supported by other sources.

Media Servers: actually several companies have Linux based media devices, e.g. set-top boxes, digital video recorders, streaming servers, etc.

Speech recognition: I wasn't even aware of Microsoft doing their own speech recognition thingy. Most often this has been in context of some other company doing work there, e.g. IBM's DragonSpeak.

Cheers,
_

Ahh, I'm saying we cannot create and distribute hardware, we can only make use of what others have created and distributed. As of this, we have a disadvantage.

Speech recognition has been in Windows since XP, it was refined in Vista and is more or less fully functional in Windows 7.

SIMON is an example of a KDE app that uses speech recognition, but it cannot transcribe things; you could not get it to write down what you write. Rather, you'd be presented with a choice of a few options and you'd say which one you'd like.


I would certinatly like to see this mess with audio sorted out, I'd love to see Kompile make a comeback. I'd love to see a proper audio tool like Apple's LOGIC...

I'd love to be able to sit down with KOffice or OpenOffice and have it understand what I'm saying to it...

Dante Ashton wrote:Ahh, I'm saying we cannot create and distribute hardware, we can only make use of what others have created and distributed. As of this, we have a disadvantage.

Who is we?
Or why would a company be able to create and distribute hardware with Windows but not with Linux?

Dante Ashton wrote:Speech recognition has been in Windows since XP, it was refined in Vista and is more or less fully functional in Windows 7.

Interesting, didn't know that. It certainly wasn't advertised very well.

Dante Ashton wrote:SIMON is an example of a KDE app that uses speech recognition, but it cannot transcribe things

Right. As the maintainer stated in the interview, its main purpose is to aid people in controlling software and that dictation has different requirements.

Cheers,
_

'We' refers to the Linux community.

I'm not referring to a company making hardware then putting Windows or Linux onto it. I'm talking about hardware like the Xbox 360's Project NATAL, an add-on bit of hardware running their software that they will be selling under their name. We do not have the capacity to do a similar thing, Canocial could not easily come out with a Ubuntu-ready webcam, for instance, as they do not have the capacity to manufacture hardware, nor deals with hardware manufacturers. The community does not seem to have any power in the hardware market, rather, it adds the ability to work with said hardware and relies upon manufactures to not screw about it.

The speech recognition in Windows XP and Vista (and, presumbly, Windows 7) requires training. Nuance Communications came out with DragonNaturallySpeaking 10, a program of the same ilk but one that is so well developed it requires no training to function to it's fullest potential.

I would love to get a speech recognition system that could operate here, as a writer it is far easier to talk out loud my stories then it is to type them.
There are two or three FOSS speech recongition systems, CMUSphinx was one I tried; it can tell when someone is speaking vs. background noise, but cannot tell you exactly what words are being said