Miguel de Icaza On Usability and Openness 349
doperative points out comments from Miguel de Icaza on the struggle for usability in many software products:
"De Icaza uses OpenSUSE as his main desktop (with the GNOME interface, of course), says he likes Linux better than Windows, and says the Linux kernel is also 'superior' to the MacOS kernel. 'Having the source code for the system is fabulous. Being able to extend the system is fabulous,' he says. But he notes that proprietary systems have advantages — such as video and audio systems that rarely break. 'I spent so many years battling with Linux and something new is broken every time,' he says. 'We as an open source community, we don't seem to get our act together when it comes to understanding the needs of end users on the desktop.'"
Hasn't used RealTek (Score:4, Insightful)
Closed source audio can break too. My last motherboard had onboard RealTek audio. Worked perfectly in Linux. Under XP, it crackled endlessly. Ended up buying a discrete sound card.
By nerds for nerds ... (Score:5, Insightful)
Re:Windows is popular because it works. (Score:4, Insightful)
Windows 'works' largely because it comes pre-installed. Try taking any random PC, wiping the disk and installing Windows on it from an official Microsoft install CD and you'll find it at least as hard to get working as Linux.
Though personally the last few times I've installed Linux I just stuck the CD in the drive, selected a few install options and half an hour later I had a working system sitting at the logon prompt. Finding, downloading and installing all the correct updated drivers for a fresh Windows install would probably take longer than that.
Re:By nerds for nerds ... (Score:3, Insightful)
You see this come in to play with de Icaza himself. Just look above in the comments. He's a "sell out" even though he probably has written more software for Linux than anyone here. So here he is trying to make Linux better and he's been cast as an outsider because he wants to make it mainstream and now works for a company that took money to make that happen. These guys must eat ideology cereal or something.
Re:Windows is popular because it works. (Score:4, Insightful)
For using the computer more as a commodity tool for email/word processing/video watching etc Windows still is better presented and more importantly doesnt break grotesquely with every new update that appears like Ubuntu does (and yes I'm looking at 9.10) Until Linux, or more strictly I suppose GNOME/KDE etc get over this then I suspect that further adoption of linux on the desktop will stall.
That's something that a lot of people seem to miss.
If you need to get at internals, Linux is the choice. If you want a workhorse back-end system, Linux is the choice. If you want a desktop with great cutting-edge features, Linux is the way to go - KDE betas are best for that ;). On the other hand -- if you want a desktop system that stays out of your way, Just Works, and requires little maintenance beyond letting an auto updater do its thing... Windows or OSX are your only real options.
When I use my computer to get a task done, my time is valuable - and I increasingly resent time I am forced to spend fixing or working around issues that are not immediately germane to the task at hand. That task might be browsing the web, editing a document, writing code, watching a video, debugging, etc. I have consistently found that I can't simply do that on the various flavors of linux - there's always something that seems to need adjusting, or stops working correctly, or doesn't work at all.
The problem is that people will often start blaming at this point, when they hear these statements. They'll say, "It's nvidia's fault for not doing X" or "it's your fault because you didn't do Y" or "it's the upstream maintainer's fault because he didn't do Z". Which is, unfortunately, completely missing the point: when you are using a system to get a task done, fault does not matter.
I, as a user of a product, want to simply use the product -- and spend zero time hunting down answers that I shouldn't need to concern myself with. As a developer and a tinkerer I understand why doing this is necessary, and can even enjoy it sometimes. But as an end user, I experience a ridiculous level of frustration and exasperation when I need to devote MY time to working around somebody ELSE's issue - no matter whether we're talking about operating system, development tool chain, pc games, the amazingly badly designed FIOS TV interface, or anything else.
In recent years, I also find that the desktops are experimenting with increasingly weird crap - things that are both fun and frustrating. (Fun because they look like good concepts. Frustrating because the deviation from the familiar means less time Just Working even as I enjoy playing with them.) I will keep trying every few months, and I certainly have more than my fair share of back end linux servers and dual boot desktop systems - but for the forseeable future, Linux just isn't there.
Re:Bigger Question (Score:5, Insightful)
You know Miguel works for the GNOME project, right?!
Joking aside, it's perfectly possible thanks to open source's inherently modular structure. Someone makes an idiot-proof GUI, distro X bundles it as the default and only option. Someone makes a uberhacker GUI, distro Y bundles it as the default and only option. Distro Z prides itself on being able to switch from newbie to expert and back again in less than three seconds.
IMHO, GNOME tries too hard to lower itself to the lowest common denominator jack of all trades - look at the recent decision to remove the "minimise" button from the taskbar because it's apparently not useful and not optimised for touchscreens. But neither is the rest of GNOME, or all the apps it's going to run. Sorry, if it's touchscreen users you're after then I'm sure GTK is perfectly capable of having a new UI constructed from the same frameworks.
Similarly, KDE often gets flak for having too many confusing options. It's personally the UI I prefer (after I've spent forever configuring it) in *nix but it's not without its own share of problems either, and much like GNOME they seem to have some project heads who are entirely convinced that theirs is the One True Way of doing it. KDE remains more usable to me because of its configuration flexibility though, but it can be baffling if you don't already know your way around, and they make fewer stupid choices than GNOME.
The problem with both KDE and GNOME's approaches (and windows as well for that matter) is people who are convinced that one tool can be everything to everybody (this goes for almost every DE I've seen in the PC world), and that the inherent differences between, say, a 5" touchscreen and a 60" TV warrant completely different approaches. So to answer your question: yes, Linux can (and does) cater to computer novices (I'm not aware of anyone needing to use the CLI in ubuntu for example, but I could be wrong) and still leave all the juicy stuff available to geeks like me. I'm no fan of apple, but when they released a phone they were smart enough to realise it would need a brand new interface, not a badly screwed fork of their desktop OS as MS did with WinCE. This supposedly revolutionary idea has netted them billions because it's the only approach that makes sense. Tightly coupled with the need to have differentiated UI's for different purposes is the attitude some people take is that theirs is the only way to do something, anyone not doing it their way must be stupid. This is tragically false - everyone has a different way of working, and what works for one person doesn't work for another. For instance, I can't live without focus-follows-mouse, despite the fact it took a lot of effort to get working in windows 7, but almost everyone else hates it. Some people just don't want the options to be there because they don't think they're important, and this stops people from finding tricks and tweaks that may help them work better; some bury the config panels with boxes and the user often doesn't have a clue what options to start with.
Off my high horse now. All YMMV, IANAL, IMHO, etc. I just think all these "there is one best way" arguments are detrimental to the computer experience as a whole.
Re:More FUD (Score:4, Insightful)
At this point one might point out that if you can't watch Hulu or Netflix (we're talking about OpenSUSE here), cannot put in a credit card number to buy or rent a movie from Amazon Unbox or iTunes, and must install separate pieces of software in order to watch DVDs, this OS may not be "broken" but it might not really be meeting modern consumer expectations.
Of course you could argue they shouldn't be paying money for content, and that the DRM is illiberal or something, but you're still keeping the customer from doing what they want to do and what other platforms don't think twice about forbidding for what are essentially elitist moral reasons.
I have to wonder (Score:2, Insightful)
Re:Hasn't used RealTek (Score:5, Insightful)
Re:So is this the year? (Score:3, Insightful)
You know, I'm pretty fed up with OSS attitude toward usability. Apparently you just don't get it.
There needs to be a way to use the software on my machine that doesn't require me to open a MAN page and edit a config file. There's a simple reason for this; people do not have TIME to do these things. The utopian world of thousands of sweaty, Cheetos-encrusted Metallica T-Shirt-wearing geeks the world over writing code that will break the Microsoft monopoly is permanently doomed to failure because you all think that design is making a Mac OSX Metacity theme.
User Interface design has nothing to do with making things "pretty"-- it has to do with making things usable. This is something where nearly all F/OSS fails. Miserably. Making software that does work cleverly is good. Making it intuitive and powerful is excellent. That's not dumbing it down; you'll find that making a user interface that works well, and designing software to do things right, quickly, is significantly harder than writing good, clean code. Shifting the blame of not being able to design an interface well to users being "stupid" is shameful-- don't blame your inadequacies on anyone but yourself.
There will never be a year of the Linux desktop because geeks will never get that.
Re:By nerds for nerds ... (Score:3, Insightful)
It's impossible he did it because he truly believes that .net is a superior development environment?
Linux is supposed to be open, yes? What's *wrong* with taking something Microsoft invented and using it in Linux? As long as that thing is good, and as long as Microsoft is ok with it.
Re:More FUD (Score:4, Insightful)
There goes the lower TCO I guess.
I think you exemplify the fundamental open source attitude, namely that only people who know how to code deserve to have a working computer, and everyone else has to pay through the nose that the coders may deign to help them. The fantasy is world where the IT dweeb becomes the overpaid fat-cat and the right to compute is really a privilege delegated by a priesthood.
Re:By nerds for nerds ... (Score:3, Insightful)
>Linux is supposed to be open, yes? What's *wrong* with taking something Microsoft invented and using it in Linux?
According to Microsoft, EVERYTHING.
Microsoft does not play well with its competitors. It doesn't even play well with its partners. Where the hell have you been for the past 20 years? Eh?
Seriously. Did you not notice any of the threats from Microsoft about patents over the last decade? Did you not notice Microsoft funding its lawsuit-by-proxy against IBM through SCO?
Incorporating Microsoft IP into Linux is the most dangerous thing anyone could do to Linux and Miguel has been shoving it in as hard as he possibly can with Mono.. And Miguel wants us to believe it's all rainbows, unicorns, and blue skies dealing with Microsoft.
No. Not in a million years. Microsoft cannot be trusted. Ever. To trust Microsoft as a competitor means you are denying history and just being stupid. The lions always go after the weak ones. You can't ever be weak in front of Microsoft. Not even as a partner.
You are either grossly naive or truly disingenuous and a troll. You pick.
--
BMO
Re:By nerds for nerds ... (Score:5, Insightful)
What's wrong? You answered it yourself:
> and as long as Microsoft is ok with it.
We don't need anyone to be ok with anything (technical) we would like to do - especially since there's no telling if they're still going to be ok when what we do actually becomes a success and starts to threaten their business.
It's not just my knee-jerk reaction. It's history. But if you didn't know that by now, you're never going to know it.
Re:More FUD (Score:4, Insightful)
My argument is that I want a product that has one predictable price, and once paid it works like any other tool. The open source business model is about selling services to make products that only work well enough to keep you buying more services -- Shuttleworth can engage in all the altruism he pleases but eventually someone needs to pay their bills, and for devs services on Linux are the only option. I don't want a serf, but if you decide a priori that shrinkwrapped software is forbidden, it becomes impossible to retail a "just works" solution; you're stuck paying the $100/hour guy who rolls his eyes at you all the time. I mean, this is your pitch for consumer Linux: it's free but your costs for support of X that Windows and OSX have will either cost you $unknown or $MAX_INT, if the feature is in forbidden by "stupid laws." Why would anyone take that deal? If you cannot yourself code, the continuing free-as-in-freedom benefits of Linux are meaningless.
Everybody here assumes I'm using Windows, which is interesting. I've never used a Windows PC outside of a Kinkos, let alone owned one...
Don't complain about poor mainstream adoption (Score:4, Insightful)
I thought the discussion was about mainstream adoption of open source software.
Re:So is this the year? (Score:4, Insightful)
It's not "his personal problem" with open source. It's A LOT of people's problem with open source. Plus, anytime someone actually dares to say some interface is, shall we say, less than optimal, someone like you comes out of the woodwork to say "don't you dare tell me what to spend my time on! If you want it fixed, why don't you lead the effort to fix it yourself!"
Therein lies the problem. You want desperately for Linux to succeed, but you don't want to actually spend the time and effort working on the things that ordinary users care about.