some of us have been talking about that in kde for at least a year (probably more, but that's the first conversation i can remember having about it explicitly)... maybe once we're out of the kde4 woods some of us might actually start working on that.
but some of his thougts don't jive with my experience. such as this one about project portland, which is an effort to bring a simple set of desktop-neutral APIs to ISVs (those people who write third party software). in his missive, luis said:
Given two mediocre ISV platforms*, and a finite amount of time and money, why spend that time and money to fix one ISV platform to kick ass** when you can spend the time and money to add a third mediocre ISV platform instead? Yay!
first, this is an effort done in direct response to ISV requests. they want something simple to use now not n years from now. portland is a pragmatic project that most of us realize may well have a limited shelf life. and that's ok.
it's ok because it doesn't take a lot of effort, not compared to the work to "fix one ISV platform to kick ass". not that doing that would fix the issue at hand, which isn't "the platforms are mediocre" but rather "how do we do X" where "X" might be "install an icon on the desktop". now, i'm not trivializing the efforts people have put into portland, but it's pretty telling that the lsb is already looking at it after a very small number of people have put limited amounts of time into to produce something useful already.
did you know that google ships some of these portland tools with their google earth software?
now, creating a right-size solution at low cost that people are needing now may seem like the greatest idea since sliced bread. well, at least like an obvious idea that deserves to be done. but this obvious idea does not come at the expense of making kde's api kick ass (i'll leave comments on gnome's api to those involved with it). we're making kde's api kick ass right now even as portland continues. and note that the driving coding forces behind portland are mostly people who hail from the kde project, so one might expect that if portland was going to cost anyone major API kick-assery slow downs it would be kde. but then one would have a hard time explaining our moving to dbus, adoption of cmake, development of akonadi, solid, phonon, liveui, the k3m and trysil meetings, etc... these are all API level, ass-kicking-potential improvements and only represent a fraction of total activity in this area.
so while i respect the desire to keep an eye out for things that will detract from the long-term sustainability of the free software desktop, ISV and API issues among them, i don't feel portland represents such a risk while delivering real world value to our ISVs.
13 comments:
It's just too little, too late. So many historical mistakes regarding desktop linux and now you're stuck with them. Desktop linux still languishes at 2.5% of the worldwide market years and years after all these ridiculous predictions were made.
If Qt had a better license, and the Gnome project had never been started or been subsumed into a freedesktop that would have started earlier, where Gnome developers could have written low-level libraries in between the kernel and the desktop (all those libs on a windows system that are the supporting cast, but not gui related) then things would have turned out differently.
That didn't happen and now people are getting nervous. There's absolutely no end in sight to stagnating desktop linux numbers and those with vested interests better hope that Apple never cuts a deal with Dell or starts selling shrinkwrapped.
Mozilla made tens of millions of dollars by directing searches to Google. Searches entered in Konqueror's Google search box have "client=safari" in the URL. Here's an example:
http://www.google.com/search?client=safari&rls=en-us&q=test&ie=UTF-8&oe=UTF-8
So the revenue from searches by Konqueror users is going to Apple instead of KDE!
> That didn't happen and now people > are getting nervous.
the people getting nervous are those who never understood the market dynamics at play. those who did have never been surprised, nor have they made rediculous claims of "year of the desktop".
the malinformed becoming nervous is no indicator of impending failure. that's just part of punditry on any major venture, successful or not.
you may note that microsoft, who probably understands more about this space than most, grows increasingly active (both positively and negatively) in this area.
> There's absolutely no end in
> sight to stagnating desktop linux
> numbers
evidently you missed my last blog entry, because the numbers aren't stagnating. anything but.
while people such as yourself have been moaning this same crap line for years now (while your counterparts on the other end of the spectrum have been spewing the "year of the desktop" crap line), our numbers have been increasing and the rate of that growth has been growing.
Aaron, you're employed by Trolltech so even if you do get it, then there is no way you can publically acknowledge the history of mistakes that are the cause of the stagnation of the linux desktop. And yes, desktop Linux has been stagnating at around 2.5% for quite some time now.
The Portland Project is an attempt to rectify historical mistakes, but that's like putting a band-aid on a gaping wound.
The ISVs have needed a stable api target that can be ensured to be on the majority of linux desktops and they never got it. You've got no decent cross-distro installers and a billion distros that each do things a little differently.
Maybe once you're no longer employed by Trolltech you can be a bit more honest about the situation. Trolltech and their licenses is what caused the fragmentation in the first place and why the current marketshares of Gnome and KDE will just ensure more of the same.
Rick (or whatever your name is),
The Free & Open Source GNU General Public License (GPL) is perfect for a GUI framework such as Qt -- it maximizes the user's freedom, and that's what I care about. I commend Trolltech for staying true to the ideals of freedom and I encourage them to not let unscrupulous entities like you rip them off and take away freedom from users.
Did you notice Dell is selling Micro$oft-free computers? And look at all the other computer manufacturers that sell Linux pre-installed. I totally agree with Aaron that the Free Desktop is making tremendous progress and that the last thing we should do is sacrifice our freedom. Go GPL!
BTW: I can't wait till GPLv3 comes out with its anti-treacherous DRM provisions!
> you're employed by Trolltech
let's get something straight right here and now, rick:
i held these positions and opinions long before trolltech came around (less than a year ago).
moreover my arrangement with trolltech is a "hands-off" one. they sponsor my work but i'm not an employee, i don't work on trolltech technology and i don't have a manager there.
so i don't appreciate your accusation of corporate whoring. if you knew me you might understand how that sort of goes against my ethics.
oh, and i did notice that you didn't actually address the facts in the blog entry i pointed to. personal accusation is no substitute for rational argumentation.
so let's try that, shall we?
> The ISVs have needed a stable api
> target
this is something they've had with toolkits such as Qt (among others) for years now.
what we have not provided is desktop integration facilities without the ISV having to commit to a full desktop stack (which isn't entirely unreasonable either), and that is being addressed on multiple fronts now by projects such as portland.
as for things like workable installers, etc ... i do suggest grabbing google earth and seeing what they provide. between the LSB creeping forward and people writing the software needed, these issues are being taken down one at a time.
yes, it takes time. and yes, as we clear each successive hurdle our user base as well as the number of ISVs targeting the platform has grown.
> Trolltech and their licenses is
> what caused the fragmentation in
> the first place
what caused the fragmentation were people running off and creating GNOME and then everyone taking years to fix the resulting divide. that's mostly behind us at this point.
and so our user base is growing quite nicely. note that we don't have to have N% by a certain date; we can (and will) continue on for as long as it takes. if we were losing users or our user base was growing at or lower than the rate of the global desktop market i'd be concerned, but we're doing just fine there.
i suggest you go look at the history of linux (and include the bsd history for fun) on the server as an example of a very similar set of growth curves.
some i really agree with, like the idea that it's almost certainly possible to bring a non-trivial amount of monetary support into free software desktop projects through web based services that we offer access to from our desktop apps.
:-). Ahhhh. I see the recurring theme of how to sensibly fund open source projects keeps coming back. Seriously, Eazel and various other companies tried this. Where are they now?
In terms of desktop Linux and KDE usage, yes, it's getting marginally bigger. However, it's never going to reach any sort of critical mass and will never be anything tangible that an ISV or anyone else can target and say "Yes, this is desktop Linux" or "Yes, this is KDE".
In terms of Portland, I can't help but agree with Luis and others. It is merely a third desktop API to mask two other desktop APIs so that ISVs can have the one API that they need. It's an open source world problem, not anyone elses' or something that can be masked. For a lot of simple things like consistent menu locations, putting icons on the desktop etc. it can just about work. However, for it to be of any real use to ISVs it will have to be fleshed out to mask all the differences. And what about installing services on a system, and all the other things software installation tends to do in the Windows world? It's several country miles off what's required. I found the whole rational behind the RuDI compatibility thing a laugh. As if organisations are going to choose between KDE and XFCE!
I find it funny that Rick talks about Trolltech's licensing, because I'm now more convinced than ever that that is what's going to make desktop Linux get to any sort of critical mass. Going back to funding, and how companies like Eazel or Ubuntu would realistically make money to cover themselves, I'm convinced that funding open source software through development is the right way to go. Being able to give open source developers a great tool in Qt to produce great open source software, and charging proprietary developers to develop the software they want. Developers get good development tools and they are also contribute to the free software that they are using as well.
However, it looks if Trolltech will need to take this a step forward and start realistically looking at funding their own freely available development and operating system platform together with the open source community. Microsoft gets it's money through license fees for software, and KDE and Trolltech can generate their income from development - a far more sensible idea to my mind. It would even be investment that would pay off for Trolltech, as a platform that they support would become much more popular and they would get less of their money from people on Windows. Microsoft's long-term aim is to kill off other development tools on Windows like Qt or Java and companies like Trolltech need a get out.
Trust me. This is the only way desktop Linux is going to work. If Trolltech don't do this then absolutely nobody else can. They just don't have the right business, and Novell is even selling a distribution in SLED that they don't publicise and that they expect OEMs controlled by Microsoft to ship! Anything else will continue to fail and will never reach the critical mass needed. Somebody just needs to get a clue, and no one seems to have one.
as for things like workable installers, etc ... i do suggest grabbing google earth and seeing what they provide. between the LSB creeping forward and people writing the software needed, these issues are being taken down one at a time.
It's clunky, not straightforward, a hundred times more complex than it needs to be and I'm not sure anybody has any real idea just how far away the open source desktop is from addressing what's needed. I find the LSB silly to be honest. Did you know you can actually pass the LSB test suites just be using certain hardware, and fail it with other hardware people might use? The only safe way to guarantee compatibility is to have the same binaries and distribution, because you are inevitably going to get totally inexplicable differences that will perplex people. People don't like it, but that's the way it is.
and so our user base is growing quite nicely. note that we don't have to have N% by a certain date;
I let out a sigh when I saw that whole 10x10 thing.
we can (and will) continue on for as long as it takes.
Unfortunately, you probably won't have that much time. Technology will continue to evolve that will make it almost impossible for people to move away from Windows even if they want to, and desktop Linux needs to be more of a force for people to take notice or even care. Time is of the essence I'm afraid.
Trolltech and their licenses is what caused the fragmentation in the first place
Actually, it was people who wanted to impose their own idea of what they thought open source software should be.
You can see it on the mailing lists from the time, that even if KDE and Qt was under a license that was generally agreed upon, the FSF and various people would have either found some way of making it their own project or found some other excuse for making their own.
As open source developers have had to take on the heavy burden of not only developing an open source desktop and applications, but also a complete development platform as well, I think the decision to use Qt has been vindicated. It's become clear that GTK, Glib, the Gnome libraries and stuff like Mono are just about good enough for open source developers but are just not good enough for the wider world. Some serious investment would be needed to achieve that, and of course, that investment needs to come from somewhere.......... ;-).
Aaron, you work for Trolltech. That's fact. You worked on KDE before you worked for Trolltech. That is fact. So we can't say that you're an unbiased source.
About the numbers...The numbers you're throwing out rely on a non-growth market. Obviously, that's not happenning. The total number of desktop linux users continues to grow, but the market percentages aren't growing because the number of new windows installations dwarf the number of new linux desktop installations.
Now let's talk about the Qt license. Let's imagine a world where Xlib has the same licensing scheme as Qt. Obviously, that wouldn't fly and an alternative would have to be developed. Or if Glibc had the same license, or any other systems library. And Qt _is_ the systems library for KDE.
Nobody sells toolkits anymore. Even Microsoft is giving away their non-all-languages-integrated IDEs for free.
The best thing that could have happened is that Qt would have been released under a liberal license, Trolltech would have paid the bills by developing applications using Qt, the entire community would have gotten behind KDE and Qt, and we would have a unified platform, and not this fragmented mess that we have now.
>this is something they've had with toolkits such as Qt (among others) for years now.
No it isn't. The only omnipresent "toolkit" that we have right now is Xlib. There are omnipresent libraries that we can rely on being on a desktop linux system - such as Xlib, glibc. Qt isn't one of them.
>what we have not provided is desktop integration facilities without the ISV having to commit to a full desktop stack (which isn't entirely unreasonable either), and that is being addressed on multiple fronts now by projects such as portland.
There's a familiar theme here. We get all this self-congratulations from desktop developers and fanboys alike. It's the "We don't have it yet, so you don't need it" syndrome. But then a couple people come around and start freedesktop.org and portland project and suddenly its being addressed - but too late.
>as for things like workable installers, etc ... i do suggest grabbing google earth and seeing what they provide. between the LSB creeping forward and people writing the software needed, these issues are being taken down one at a time
We could have had a universal installer that handled fine grain versioning years ago, but that never happened. It's just another case of fragmentation. That Smart package manager is a step in the right direction, but too late.
>what caused the fragmentation were people running off and creating GNOME and then everyone taking years to fix the resulting divide. that's mostly behind us at this point.
obviously, because of your Trolltech and KDE affiliation, you are incapable of admitting the truth about what happened. Everybody knows it was the Qt license that was the cause of Gnome being created. And we're not even close to being "mostly behind us at this point"
>if we were losing users or our user base was growing at or lower than the rate of the global desktop market i'd be concerned, but we're doing just fine there.
And back to the first point that relative to global market rate it is not really growing or growing at such a small rate that it is insignificant. You better hope that Google never releases a desktop for Linux because KDE and Gnome, both with about 1% of the overall desktop market, are not entrenched enough.
The major problem that desktop linux faces is the inability for developers and those with desktop linux interests to see that they should be competing against Microsoft and not with each other.
We see the same sort of fragmentation happening in the handheld world right now! Just another disaster while Microsoft pushes their integrated stack. Oh well, some people never learn.
Rick, you're on crack!
Nobody sells toolkits anymore. Even Microsoft is giving away their non-all-languages-integrated IDEs for free.
You're pulling this out of your *ss. To develop for Microsoft platform, you have to pay for the Windows operating system and pay thousands of dollars for the Visual C++ studio. Same for Apple.
Everybody knows it was the Qt license that was the cause of Gnome being created.
Yeah, the people who started and are heading GNOME are sellouts, and KDE needs to make this clear to the average GNOME developer who cares about freedom (most do). I hope KDE does take (and publicize) the high moral ground, because the Qt/KDE framework certainly deserves it.
Nobody sells toolkits anymore. Even Microsoft is giving away their non-all-languages-integrated IDEs for free.
I'm afraid you're missing the point. Microsoft can afford to give away development tools because they sell Office, Windows and various other bits of proprietary software. Trolltech funds itself by selling something that people can make into software that they can sell or increase productivity, or something, which in turn can be used to make better free software..
To my mind, if you want free software then that's the only thing that's going to work.
Trolltech would have paid the bills by developing applications using Qt
And Qt would have never been that good, and they would have went bust like every other Gnome company who's tried this. Wonderful.
Everybody knows it was the Qt license that was the cause of Gnome being created.
Saying it doesn't make it true. If it wasn't the license it would have been some other excuse, like allowing prorietary developers to develop for free or something.
Post a Comment