The problem is that Unity doesn't have much in the way of deviation from the standard configuration. They're not just changing defaults, they're removing everything else.
This is a problem not for people who are already developers, because they will just install something else and move on, but for new users who won't have anything to discover.
The article said "In fact I think Linux has a tendency to encourage average computer users to become power users once they spend some time with it."
This may not be the best thing for everyone, but that's what the article was getting at.
>>> "In fact I think Linux has a tendency to encourage average computer users to become power users once they spend some time with it."
I think this is an important general point. I've got no real opinion on the evolution of Ubuntu, but we must stop thinking about beginners and expert users as rigid, separated categories. First, there are many degrees and variations on the tech savviness scale. Second, people can learn. Granted, many people won't, since they don't care about the tools they use. But others will learn, if you give them computers that don't push toward laziness and reduce the steepness of the learning curve.
I think Mac OS, in some aspects, is a good example of empowering. For years, Preview and iPhoto had basic tools for editing pictures that Windows lacked. You can adjust saturation, levels, and so on. You can realize that these tinkerings can be not hugely complicated. And if, as a newbie, you end up stuck in Photoshop someday, you will be less scared and hopeless than if you had never used Preview.
The vast vast majority of people want to click on the Firefox icon and go to Facebook. People have lives that are busy, and the computer is a tool to get a job done. How many people "explore" the engine of their cars? Not many.
What about something as simple as changing you screensaver? My few years working as a sysadmin in a maintenance department at a University (read low tech knowledge) tells me people who care if they couldn't change something as simple as that yet you can't do that in Ubuntu 11.10. Changing fonts is another thing that's pretty annoying in this release but not as big a miss as the screen saver to most people. I hit those two things in the first few minutes of using the new release.
They are removing so many simple things like that pretty much everyone expects to be there that it's becoming ridiculous. The fact that the option to change some config exists does not in anyway prevent someone from clicking on the firefox icon and going to Facebook. (and I really hate the attitude that most assume people just want to go to facebook. The fact is most people do a whole lot more with their machine than most simplists are assuming with no data to back it up.)
I'm all for sane defaults but don't remove every configuration knob.
I couldn't agree more. Lots of people are panicking about Unity because it's not as configurable as Gnome use to be, but they forget that we're in a transition period. Unity is brand spanking new, and I'm sure it'll have more configuration options in the future.
That might be a good point if it weren't already so simple to build user interfaces that are simultaneously obvious to newbies while at the same time being ridiculously extensible. You just build it in a language that exposes a repl. Mozilla doesn't even do a particularly good job at this (given how they bury the replness), and it's still fostered a plugin ecosystem that has kept it competitive despite competitors eating its lunch on performance grounds.
Having a REPL and being extensible seem like totally orthogonal concepts to me. You can have a REPL and be hardly extensible. You can lack a REPL and be extremely extensible. If anything, the REPL is going to expose interfaces that enable extensibility; it's those APIs that are important, not the existance of an REPL.
Technically they are orthogonal, but in practice if you have a repl implementation that's well done, it will be used by the developers, and the developers will ensure that the functionality they need to implement the project is accessible through it. The developers can still screw things up, but it's a factor that encourages them to do the right thing through basic dogfooding.
Obviously it's implied that the repl be implemented in the same language as the rest of the codebase. So adding one to a project written in C++ would be a waste of time.
> What about something as simple as changing you screensaver?
In over a dozen years, I've never heard my wife ask me: "how do I change my screen saver?". In over 30 years, I've never heard either of my parents say: "how do I change my screen saver?" I've also never heard this question from aunts, uncles, sisters, brothers, cousins, nieces nor nephews. It's not for lack of questions. At holiday events I wear the T-Shirt that says: "Yes. I'm a software Engineer. No. I will not fix your computer".
But do the mainstream buy it? What's their selling point aimed at them?
Everybody who I know who uses or tries ubuntu is either:
- A software developer of some kind.
- An advanced or else adventurous user who wants perspective.
- A family member of one of the above.
Neither of those is "mainstream" in that sense. And making product mainly used by X while keeping Z in mind is a troublesome tactic.
Perhaps true, perhaps not. Some people don't want to invest time and energy to save time and energy later. Others do. A good Linux desktop should be able to cater to both types of user.
"Some people don't want to invest time and energy to save time and energy later."
I believe the phrasing you were looking for is "the vast majority".
Apple concentrates on providing a good experience for that vast majority. Everyone else can use OS X's CLI or tools like Automator, AppleScript, Xcode, Dashcode and MacRuby (all for free, albeit not all of it is 'free software').
That's simply not true. People pay thousands of dollars to attend Excel classes which show them all the little 99.99%-of-people-will-never-need-this-feature features, and they then use these features to be more productive in their work. If they were programmers, they wouldn't need those features, because they are simple enough to implement. But they're somewhere between programmers and users: they can't program, but they can learn a lot about how to use software. To make those users productive, you have to give them a lot of features.
Yes, the average person that visits YouTube and Facebook doesn't need a featureful desktop. But many people want to use their $2000 computer for more than that.
This is a problem not for people who are already developers, because they will just install something else and move on, but for new users who won't have anything to discover.
The article said "In fact I think Linux has a tendency to encourage average computer users to become power users once they spend some time with it."
This may not be the best thing for everyone, but that's what the article was getting at.