Dealing with humans is harder than dealing with the kernel.
And yet, cars and airplanes and bicycles and a lot of other tools (some more complicated, some less) are still being engineered by engineers.
I don't see a link here, sorry.
Dealing how they interact with a computer is quite a bit different. It is an entire research subject.
Dealing with (some) corporate egg headed's wishes on how people should use a computer is that it's quite different.
Common People are using computers since the 80's. Some metaphors sucked a lot, some others was successful (and some, besides sucking). The break even point was simple: I work/play/whatever better now that did yesterday?
Cars use steering wheels and pedals since Ford's era, besides a lot of research on new ways to drive a car. Did you ever considered why?
GNOME 3's deployment could have gone better. It would have been better to allow people to parallel install both and let them move when they were comfortable
That would had helped a lot, indeed.
The design itself continues to be a successful work in progress. Next iteration, I will likely make sure that we don't do it this way.
Successful to whom? You see, people are not all alike.
Let me a bit more succinct. I'm saying that the people who were building Linux based desktops took a lot of their design from windows 95, amigaos, and various other desktops that were there before. All the examples you've stated doesn't invalidate what I said. We took our designs from someone else. The most popular ones are based on windows 95, GNOME and KDE.
Now I see.
Perhaps this happens because is considered BAD ENGINEERING creating something out of the blue just for the sake of it.
There's a good reason cars are still using steering wheels and pedals nowadays - people don't have to be trained and reissued a new driver's license when they sold the old model and buy a new one.
Any nowadays stereo are still using the same metaphors from decades ago (Play, Stop, Next, Previous, Volume Up and Down - even the Eject are still used), and the remote control from my new shining OLED TV has the same buttons my old, ultrasonic, mechanical remote control had (but granted, it has that buttons and a lot more).
Of course modern remote control are a lot more sophisticated and useful that hat old craps, but the basic metaphors are still there - the enhancements were introduced rationally, without breaking current usability.
Why? Because this is considered *GOOD ENGINEERING*.
Nopes. I just asking Linus to step up and LEAD a Desktop project.
The result will probably be not shiny and new and full of [insert your favorite insult here], but it will be usable, and it will works, and more importantly, it will get the job done without hassle - that what matters when the month ends and I have to pay my bills.
Why? What gives him the expertise to run a desktop project, exactly?
Because it's already proven that he is capable of handling successfully huge and complex open source projects.
Just because they are both software projects doesn't mean that he has the ability to lead a group of people working on UI and middleware projects.
Just because he never did it, doesn't mean that he doesn't have the skills neither. However, he's a proven open source leader that delivers solid artifacts, with a nice (but granted, far from perfect) historic of consistency between releases. I prefer to do not make comparisons with some other Open Source projects - there's no need for a jihad here.
I'm not hiring the guy to write code or specify usability requirements. But you can bet your arse I would hire him to lead a critical software project anytime (if I would be on such position, what can be a good idea to state now that I'm not).
I don't think he's remotely interested in running it.
I agree 100% with you here. :-)
Perhaps when I run into him at some conference I'll ask him. Perhaps you can ask him?
Weirdly enough, this never crossed my mind. Perhaps linking this conversation to him?
In any case, I find it amusing that people think Linus is some kind of uber geek.
Me too. The guy is a god damned good engineer with a good leadership skills on Open Source projects. I don't see how his geekness can influence on it - but granted, I'm not entitled to say otherwise.
By the way, I'm not saying he's a polite, nice leader - au contraire, the man can be an ass when pissed on. But again, the guy stills delivers solid artifacts - thanks or besides his assness, I can't say.
Given the personal feedback he's given me, I know that his requirements are fairly modest compared to some.
Being that the(or one of them) reason you stated before that the Kernel is less complicated and easier to keep than a Desktop?