The Linux Desktop in 2007
8 January 2007
- 2006 has been a good year for the Linux Desktop, with a bright outlook
- for this new year. While other platforms will have to deal with upgrade issues involving Vista and OS X Jaguar, for the Linux desktop, 2007 promises to be a year of consolidation and expansion. Here are my predictions. If this blog survives to be three years old, then we can see how well I did! Below we cover 3D effects, a resurgence of Java in the open source world, improved wireless support, refugees from Windows Vista and whether Adobe will open source Flash.
1. Bling
- 2006 was the year that 3D effects became (almost) mainstream. In 2007,
- almost every distribution of a Linux operating systen ('distro') will have to think about how they incorporate 3D effects into their releases. Most distros will find they have to provide a default option or fall by the wayside. I expect a greater degree of user movement between distros as 3D effects become a key selling point and an expected feature.
In 2007, the Beryl Compositing window manager will continue to surge ahead of Compiz in features, although the corporate supported distributions will probably stick with Compiz for this year.
I expect no resolution of AIGLX vs XGL, although by the end of the year we should have a greater idea of what is going on because all the now-neutral players will have to had laid their cards on the table (see above). Novell and Redhat might even be able to get together to thrash a common solution out, but I seriously doubt they will.
Co-operation is especially unlikely considering Novell's current pariah status. In 2007, Novell will probably stop digging their Microsoft hole and start looking for a way back, although it is hard to predict what that could possibly be. Expect an all out public relations exercise (more bright green clothing).
2. Java
During 2006, the GNU project's attempt to replace the proprietary Java runtimes with GNU Compiler for Java and GNU Classpath started to bear fruit, and Sun realised that its longtime strategy to prevent incompatible implementations from proprietary software vendors was in fact causing incompatible implementations among the open-source world.
So during 2006, Sun allowed Linux distributions to distribute Java using their own package managers, leading to an uptake of usage of Sun's rather than Blackdown's JDK/JRE. Some proprietary software companies may make a lot of noise and throw their toys out of the pram, but in 2007 Sun will ignore them and make good on its promise to open source Java, it has to, otherwise the 'write once, run everywhere' concept will fall apart.
By the end of 2007, Linux distributions will begin to come with a Sun Java runtime by default. I expect Sun will try to find a way to bring the GNU tools developers into the Java mainstream, perhaps with Job offers or prominent roles within the newly formed Java open source project. Anyone who contributed a significant amount of code to it should start brushing up their Sunday shoes.
3. Flash
- 2006 was the year that Gnash became the leading open-source Flash
- movie player. Meanwhile, or possibly because of Gnash, the proprietary Linux Flash player finally caught up with other platforms.
Due to the extremely large installed base and the training requirements for Flash movie creators, there is an upper limit to how much the Adobe Flash team can make dramatic changes to their player - at most a new version every two years, much slower than Gnash can develop.
So Gnash fully supporting the latest Flash is a matter of time, possibly having partially working Alpha code by the end of 2007. By the end of 2007 the Gnash player may well have more full-time equivalent developers than the Adobe flash player itself.
At some point there will be a tipping point and Adobe will be in the same situation as Sun was with Java, they will have to either open-source the Flash player or accept that they have lost control of Flash on the Linux desktop, mobile devices and potentially other places too.
If Flash becomes a generic protocol then it gives less reason to buy Adobe's own Flash creation tools, hitting the bottom line.
I'm sure flash hacking is hard work, but there is no need to overestimate the task here, just look at the filesizes. Java is 60MB, while the Flash player is only a few thousand K at the very most.
So my prediction is that moments before Gnash player has full support for version 9, Adobe will buckle. Whether that comes in 2007 or 2008/9, I'm not sure, it will certainly all be over before Flash 10.
4. Wireless
For the first time in 2007, I predict Wireless will become a non-problem on Linux. The abstracted Wireless card support stack will mean the kernel will finally move as fast or faster than the manufacturers, while Dbus/Hal and Network Manager the service (but not necessarily the applet) will make wireless support standardized and simpler, although there will still be a healthy competition in providing front-ends to it.
5. Media codecs
I predict the situation will not get any better in 2007. Even more incompatible media formats, many of them involving DRM, will be foisted on consumers.
Indeed consumers will have so many incompatible devices that there will be an obvious market advantage in providing content in de facto standard formats such as MP3s, MPEGs, etc. If the content providers will not fulfill this demand then the hardware providers will. Expect home media devices that double as 'content safes', ripping everything to common future-proof formats and storing the files safely in a box that can be accessed by your ipod, computer, TV and so on.
Like the grim reaper stalking his quarry, the pain of incompatibility will slowly but surely overtake the advantages of high-definition. Many hastily bought electronic goods will end up in the second hand shops and the landfills.
6. Windows Vista
In the Windows world, 2007 will be the year of Vista. While some consider Vista a threat, I consider it to be a competitive opportunity. There are a lot of PCs with specifications that will either not run Vista at all or will run very slowly indeed. These are not necessarily old computers, many of them are budget PCs with integrated graphics and low memory. Many of these computers have a long usable life, just not a simple upgrade path. Upgrading to Linux may be the most environmentally friendly thing these people can do.
As well as those who cannot keep up the technical arms race, then there may be another wave of Vista refugees - people who lose their patience with DRM hassles in Vista. If you buy legitimate digital content and then are forced to buy it again in a different brand of DRM, then that gets annoying pretty fast.
From the perspective of those people who dual-boot as an easy introduction to Linux or because they are gamers, nothing will change in 2007. GRUB will boot Vista with no problems and despite an attempt to make a next generation filesystem (WinFS), Windows Vista will use NTFS all through 2007 so the user will still be able to access all the documents in Linux that he has created or downloaded within Windows.
Here were my predictions, what are yours?
1 Robert Haas says...
I predict that those unable to upgrade to Vista, and refugees who can't stand it, will end up using Windows XP. In fact, I predict that Windows XP will end up having the longest "tail" of any Windows release ever, with more people using it for longer after Microsoft wishes that they had upgraded than ever before. I predict that Microsoft will attempt to cut off support to force people to upgrade, but that massive user backlash will prevent them from doing so, and that there will eventually be XP SP3 and probably XP SP4 (though the latter will not happen in 2007).
I predict that Linux bling will continue to lag behind Windows bling, and that the importance of this "bling gap" will continue to be overstated by large numbers of people.
I can't tell you how happy I would be to see predictions (2) - (4) come true. That would rock.
Posted at 1:42 a.m. on January 22, 2007
2 Zeth says...
>I predict that Linux bling will continue to lag behind Windows bling
Hi Robert, I assume you have not yet had a chance to try Beryl in anger? I have also seen the Vista and Panther demos and they come nowhere near Beryl.
Posted at 8:36 p.m. on January 23, 2007
3 Z says...
agrees, beryl rocks... what vista bling ? one or two light aura effects are not bling
Posted at 5:48 p.m. on June 21, 2007
4 James C says...
You guys must be dense or something, it's Leopard not Jaguar. Jaguar was like 10.2, we're next on 10.5 and are currently on 10.4. Tiger aka 10.4 has been out since 2005, and has been in development since the release of 10.3. What was added was automater, spotlight, quartz changes, new core apis, and underlying stability and unification of the system.
Linux can't say the same, it's still not unified, it's still scaring people away with its communist breath. Beryl is cool, but it's alpha software. It doesn't work right on every card, as it is not abstracted, and it won't become stable ever because nothing with free software ever steps up to the plate and marks theirselves stable when they should be. Always in limbo, always cutting edge, always fighting, scaring away people, it's not gonna ever catch on, especially when people are used to paying for things.
I don't mind paying, as long as I get some cool features that WORK and don't break all the time. The only stable thing with GNU/Linux is the base, and then again, look at the kernel itself how much it changes, and its continued lack of standardization. I've used Linux, Beryl/Compiz/Fusion for a while, and it's not my preferred platform. A while meaning years, it's not suited for the desktop, Cedega isn't good enough and the breakage/discommunication between distros is just annoying. Always been more of a FreeBSD/Solaris fan for this reason, although the later is a tad annoying to work with.
Mac OS X 10.4 Tiger was a stabilization period where Apple's Kernel and Driver API's were finally marked 1.0. Vendors and people that don't live in their garage appreciate consistancy, and the FOSS nutjobs need to learn this. GPLv3 is scaring people away worse than GPLv2, it's unfair and unrealistic, and it's a big thing killing this new found 'shiny' GNU/Linux has this year.
Vista might be slow, non-innovative when stacked against the aging but still good Apple product, but it's pretty stable and Windows is still the more consistant platform to program for. Developers are what matters anyways, and GNU/Linux isn't that friendly unless you like keeping current for no reason all the time.
Posted at 11:57 a.m. on July 16, 2007
5 Zeth says...
Hi James, thanks for reading and commenting. I think you kind of knew I am going to disagree. No offence intended, but at points your arguments seem somewhat hard to substantiate or hard to follow, so you will probably want to reply if I am not getting it.
> You guys must be dense or something,
Please keep it technical, lets not attack each other.
> it's Leopard not Jaguar. Jaguar was like 10.2, we're next on 10.5 and are currently on 10.4.
Are you referring to a different post?
> Linux can't say the same, it's still not unified,
Sorry, what do you mean by that specifically? Linux is about freedom, therefore anyone can distribute it. You can also can put it together any way you like. This is a feature not a bug, the analogy is with Children's toys.
Action Man VS LEGO
Action man is a soldier, he does not do much else than that. Therefore Action man gets pretty dull after a while. LEGO is cooler because you can making it into many different types of things, you never get bored of LEGO.
Linux is like LEGO, the same codebase that runs almost all the top 500 supercomputers and Google's massive datacentres can also can run on a mobile phone, a satellite, a laptop, and so on.
Senator McCarthy not required
> it's still scaring people away with its communist breath.
We are the real believers in free markets. It is Microsoft and Apple that want to create a global monopoly of all technology and content, a new monarchy.
Have you ever heard of Eric Raymond? He was one of the early proponents of Linux but is also a prominent member of the national rifle association, no one can call him a communist.
Who has been scared away? Wall street all runs on Linux; the original capitalists in computing, IBM, have bet their business on Linux and make a lot of money from it. The biggest distribution, Red Hat, is a publically traded company.
On 'stability'
> Beryl is cool, but it's alpha software.
Well to start with, 'Alpha' and 'Beta' are stupid marketing terms that are unquantifiable. However ignoring that for a minute, you are certainly right that at the moment Beryl is still somewhat experimental yes; but in my limited experience I have found it a lot more stable and less resources hungry than Vista's effects. Effects aside, the benefits of Linux to me far out weigh a transparent window or a spinning cube.
> it won't become stable ever because nothing with free software ever steps up to the plate and marks theirselves stable when they should be.
Well we can't lie because our code is open. Others will come to market with a broken product and pretend that it is stable then send lots of updates to try and patch up the bugs. Are you saying Emacs is not stable? What about Firefox, that works pretty well. Evolution seems pretty solid too. GCC, binutils and all that seem fine. I'm not sure what else to look at.
> GNU/Linux isn't that friendly unless you like keeping current for no reason all the time.
What are you talking about? Windows XP and its service pack 2 changed the security model completely and broke very many applications which could not even start in non-administrative mode. Applications on Linux from ten years ago still work either no changes or simple recompilation. That is because they tended to be written in an abstracted and portable way in the first place.
Let's compare Apple, if we take an application from ten years ago, we are talking about Mac OS 8, these applications tended to be very integrated with the underlying implementation of the operating system. Since then, Apple have changed kernel, changed compiler, changed from big-endian to little-endian, changed graphical toolkits, changed almost everything. Good luck getting that OS 8 application to work on a Intel Mac.
Freedom of speech as well as free beer
> always fighting, scaring away people,
Democracy and freedom are noisy yes. The idea that we "we need a strong man to give us a quiet life" argument is the one that leads to Hitler. Who is fighting, who is scared? We are fighting Microsoft and other people who want to deny freedom for the many in favour of the few. And yes they are scared because we are slowly winning.
> it's not gonna ever catch on, especially when people are used to paying for things.
Well it has caught me, if other people want to use crap and overpriced products then I can't do much about it. You can bring a horse to water but you cannot make it drink.
Free is the new business model that has become dominant. Do you pay directly for TV? No you don't, it is just broadcast at you. Do you pay for search results? Do you pay for the newspapers that they give away on trains and buses? However, the point is not free beer, but free speech. Rights for the many, not the few. Control for the many, not the few. Let Freedom reign for everyone.
> I don't mind paying, as long as I get some cool features that WORK and don't break all the time.
If you buy Redhat, or buy the Canonical support for the long term release of Ubuntu, then you will have no problems then. It will be stable and any problems will be fixed for you. Likewise buy a Linux laptop from the Linux Emp. or wherever and it will be all done for you. If you do not want to pay then you figure it out yourself. It is the same with Apple or Windows.
> continued lack of standardization
Well the cycle goes innovation > standardisation | innovation > standardisation
Sometimes it is too early to standardise things, sometimes it is not. You only have problems when you have to interoperate with Microsoft software and the roadblocks that Microsoft specifically chooses to put in the way.
> Cedega isn't good enough
You are right, compatibility layers, while essential for now, are not enough. We need to break the MS monopoly then there will be more incentive for the game companies to create native ports.
> breakage/discommunication between distros is just annoying.
Then just use one. This is called competition. Competition is good.
> Vendors and people that don't live in their garage appreciate consistancy,
The core operating system is based on Unix which has been more or less the same for like 30 years, is that consistent enough? Sometimes in your argument you bash us for being not cutting edge enough and other times for being too cutting edge.
> GPLv3 is scaring people away worse than GPLv2,
GPLv3 has been fantastic. I'll write you a whole post to explain why. The GPLv2 era was only the start, in the GPlv3 era we are really gonna kick some arse.
Posted at 1:50 p.m. on July 16, 2007