Realize that is who Ubuntu is aiming at, converts. I know Unix has many superior features and I'm for it, but the Ubuntu people want to make the transition easy and have already made a lot of changes from other distros. If they are going to go in that direction, they should go all the way. Then if those people get interested in trying slackware or something else more Unixy the difference is su vs sudo not oh my god what do you mean the command line?

I think it's worthwhile for Ubuntu to aim at converts from Windows. I also think this experiment was useful. I just thought the significance of the result should have been stated differently. Making Ubuntu friendly for computer-illiterate people is a separate, but also worthwhile goal, and would call for different experimental subjects.

"make the experience a lot more friendly for people who use Microsoft Windows extensively"

should be

"make the experience a lot more friendly for the 99% of the population who use Microsoft Windows extensively"

Did you read the article? The subject of the experiment had, for example, "downloaded music using uTorrent before in Windows" and that affected the results of the experiment. She might be representative of a high percentage of friends of the news.yc community, but there's no way she's representative of 99% of the population.

You're right, she sounds somewhat more savvy than the average user, so her troubles should be even more worrisome. If she cannot do something, it is likely that >90% of other users also cannot.

I think the error would go both ways. Average users might fail where she succeeds, e.g. she recognized the Firefox icon immediately. They might succeed where she fails, e.g. they won't search the web for .exe files.

Just to be clear, I do think this experiment was worthwhile. It's important for Ubuntu to be usable by more savvy users. I just think a separate experiment would be needed to truly measure usability for computer-illiterate users.

Well, learning is too hard for most people, so apparently Linux should be exactly the same as Windows.

If you want to avoid the $300 Vista upgrade, along with the DRM and general suckage of Windows, you might have to learn to use the Free alternative. Why is that such a terrible thing?

Firefox started its ascension by systematically eliminating any reasons users had for using IE, until it was similar to IE, but with more power, less suckage, and a better interface.

It certainly doesn't mean Linux should jettison things that rock, like throwing out package management in favor of Windows-style installer chaos. But at the very least, we should expect to help users coming from different paradigms understand what package management means and how it's different. Any novices who are remotely sane will love the little Add/Remove Programs widget in Ubuntu.

Exactly, his complaints are mostly "Why didn't they make this little thing slightly clearer for the uninitiated" i.e. call it Brasero CD/DVD burner instead of just Brasero. That is hardly throwing out the baby with the bathwater and copying Windows.

But Ubuntu and Linux want these users. Why is helping casual computers out with a few simple, easy to implement tips such a terrible thing?

Every hour spent developing these features is an hour not-spent developing real features that differentiate your application from its competitors.

Besides, if the features are so easy to implement, surely you have a patch ready for submission.

Wrong, every hour that helps fix a show stopper bug for 90% of people is better than implementing a new feature for 10% or less of the people. If something you did every day didn't work under Ubuntu you wouldn't switch. If something that Windows didn't have, didn't exist in Ubuntu either, you wouldn't notice. A lot of these are not simple fixes either, you are not going to change the DVD codec legality with an hour of hacking. Most of the remaining sticky issues seem to be either polishing some things off or legal/commercial reasons. You'd think with Dell getting behind Ubuntu more would have changed, but so far not so much.

Sure, but I don't know that I'd characterize the lack of tool-tip bubbles as a show-stopping bug.

For you, it's not (I'm talking usability for novice to intermediate level users, not just specifically tool-tips). If you're looking for wide adoption of this technology by current Windows users, it's showstopper.

Lack of usability for novices and intermediate users should not be considered a bug unless ease of use for novices and intermediate users was a planned or expected function of the software. If an author writes software for experts, one shouldn't expect the software to cater to the needs of novices.

Usability tweaks to ease use for a class of user should be feature requests. So a hypothetical developer of open source software has to choose between spending her time developing software features that have no fundamental use to her, or spending it developing features that are exciting, interesting, or are otherwise fulfilling.

I don't think you should be voted down (unfair!), because I see where you're coming from. A developer wants to build neat stuff that he or she finds useful. Developers don't want to spend their time doing routine, boring, or trivial stuff, and some of the usability requirements for non-expert users can seem to be just that.

The rejoinder is that wide adoption requires that kind of usability. It must be an integral feature of your application, or your app will remain ghetto-zed and obscure. That may be ok for some developers, but for most, usability must be just as key as functionality.

You're right. Unfortunately, due to the distributed nature of open-source application development, making that sort of shallow usability integral to the whole system is very difficult. Until there exists an incentive for open-source developers to provide those sort of features, you either won't see them, or you'll get the patchwork quilt that the author was complaining about from Ubuntu.

the perfect measure of UI success, is based on the total efficiency of the system. If 10 million users have a problem with a driver that could be easily fixed with one command line call. you still have:

10 million * (30 min on google + 1 hr reading docs + 1 min typing the command line)

that is a 15.0166 Million person hour loss.

Making the process more efficient, does not involve making it easier to Fix the problem for the user, it involves making the problem not existent.

The number 1 rule of design is "if a user does not need to make a decision, then you should make it for them."

When I was in Japan, you could put your subway ticket in any direction, and "IT WOULD JUST WORK," the machine did not spit it back out at you and FORCE the USER TO DO IT THE RIGHT WAY, because it was a waste of time to do so. Instead the machines are incredibly more complicated to account for someone putting the ticket in wrong. It made the whole subway experience work much more smoothly. Now people who can not read, make a mistake, or are completely foreign to the system, can JUST USE IT. They do not have to read docs, to do one simple thing, pay for the subway ride that should be fast, simple and easy.

The whole software industry could use a little bit of time learning industrial design.

The fun part is that you can put your ticket in upside down but it comes out right-side up. Very amusing.

It's almost unfortunate that everything is contactless now.

Yup - particularly "search for website and download app" versus "open package manager and search for app".