You have organised this conference. You have an audience of millions of Windows users. It's likely once-in-a-lifetime for the majority of them. You are going to give them a hands-on presentation detailing the value of free software and Linux, you will have their undivided attention for a week.
One problem: the organiser fell through with the chairs. There are unassembled ones ready to go, but the company who made those parts require that their guys assemble the chairs. You have some real smart fellas around that figured out how to assemble the chairs, only it's not perfect and there's a bit sticking into your butt.
So you start off the event with everyone standing their respective venues, no chairs in sight. You start off like this:
"We had a problem with the chairs. If you don't want to sit on the dirty floor you'll need to use one of these chairs one of our guys put together." (Not telling them about the bit sticking into their butt, which you don't)
"We can get the manufacturer to build the chairs for you, however, because we don't know how they did it we don't know if the chairs will break. If it breaks you could snap your spine, turning you into a paraplegic."
"Or you could remain standing, or sit on the floor."
How many people will walk out right there and then? Would you?
Some people, who no have reason to trust you over the manufacturer (keep that in mind), might use the ones your guys made. After a day sitting with a pokey bit in your butt, what would you do?
If you had a fully functional chair how long would you stick around for? Who provided that chair?
What have you just done to your potential audience?
The message is important, maybe one day people will understand but never if this hardline approach is maintained.
An honest appraisal of each option is needed for users, up-front during the installation process. They must not go in search of it:
- Open source drivers: poor performance, very stable, likely to not crash.
- Proprietary drivers: high performance, we don't know if you should trust them, if something goes wrong there is nothing we can do.
> Change always comes bearing gifts. ~Price Pritchett
> An honest appraisal of each option is needed for users, up-front during the installation process. They must not go in search of it
The better solution is there should never be a need to search for it. AMD already knows their Mesa driver is all around a better user experience than Catalyst, and are making steps to depreciate work done on the Catalyst side by using a free kernel driver. Hopefully soon they will depreciate it entirely and dedicate all their Linux developers to making Gallium great.
Because Gallium is not just about playing games anymore. It is the basis for every other free software graphics driver besides Intels. It has the potential to revolutionize not just the Linux ecosystem but all operating systems by providing a well architected way to support dozens of GPUs with one system API and one driver model. It has a state tracker model that lets you run OpenGL, DirectX, OpenMAX, EGL, GLES, etc all on top of one driver implementation.
And Nvidia should get in on that, but those who hold the reins over there (business suits) are incompetent twat assholes who don't want to play ball with free software and damn all of us for it, and cripple the industry and any momentum they have because nonsense like "SteamOS only works on Nvidia cards".
OpenGL 4.2 will land soon™. A lot of stuff up through 4.5 after that is already done. Once they catch up on OpenGL, hopefully they can fix the performance.
I completely agree with you on this. "Linux culture" has an attitude problem that WILL prevent it from overtaking Windows if things don't change. Unfortunately, calling out this particular problem (especially among a Linux-ey community) tends to attract the worst of that attitude problem attacking those who see the problem.
> Open source drivers: poor performance, very stable, likely to not crash.
Actually, in my experience the open-source drivers for all my NVIDIA GPUs has been abysmally unstable. I couldn't even boot or install Ubuntu with the graphical interface, because nouveau kept freezing.
You could blame my uncommon but perfectly valid SLI GTX 580 setup for confusing the open source drivers, but the end result remains: the open-source drivers are inferior in every way except perhaps for the "open source" part (which is kind of meaningless if they don't work).
The fear, I feel, is in making Linux follow Windows's user experience, rather than doing its own thing. Meaning, any call for "make it more user-friendly" is heard like "give me windows". It might make sense to rally behind Ubuntu as a user-friendly distro and push that, only mentioning Linux the way "MacOS X is built on Darwin".
Another big problem is packaging for third-party software. Personally I would love to see a system where software isn't distributed as final binaries, but pre-built object files, together with a few very light shim source-files that then get built and linked on my machine, and rebuilt and relinked when the libs they depend on get updated.
You have organised this conference. You have an audience of millions of Windows users. It's likely once-in-a-lifetime for the majority of them. You are going to give them a hands-on presentation detailing the value of free software and Linux, you will have their undivided attention for a week.
One problem: the organiser fell through with the chairs. There are unassembled ones ready to go, but the company who made those parts require that their guys assemble the chairs. You have some real smart fellas around that figured out how to assemble the chairs, only it's not perfect and there's a bit sticking into your butt.
So you start off the event with everyone standing their respective venues, no chairs in sight. You start off like this:
"We had a problem with the chairs. If you don't want to sit on the dirty floor you'll need to use one of these chairs one of our guys put together." (Not telling them about the bit sticking into their butt, which you don't)
"We can get the manufacturer to build the chairs for you, however, because we don't know how they did it we don't know if the chairs will break. If it breaks you could snap your spine, turning you into a paraplegic."
"Or you could remain standing, or sit on the floor."
How many people will walk out right there and then? Would you?
Some people, who no have reason to trust you over the manufacturer (keep that in mind), might use the ones your guys made. After a day sitting with a pokey bit in your butt, what would you do?
If you had a fully functional chair how long would you stick around for? Who provided that chair?
What have you just done to your potential audience?
The message is important, maybe one day people will understand but never if this hardline approach is maintained.
An honest appraisal of each option is needed for users, up-front during the installation process. They must not go in search of it:
- Open source drivers: poor performance, very stable, likely to not crash.
- Proprietary drivers: high performance, we don't know if you should trust them, if something goes wrong there is nothing we can do.
> Change always comes bearing gifts. ~Price Pritchett