Hacker News Reader

Comment by jcromartie | Hacker News Reader

The answer to the desktop Linux usability problem used to be: "we should look more like Windows!"

Now it's: "we should look more like OS X!"

Where's the innovation? Why is an exciting new desktop just another imitation?

> /opt? /bin? /sys? dll files? What is all that?

Most users have no reason to even wonder, since they are already tucked away from them. Stuff like this just shows that the project is out of touch right from the start. Linux desktop projects constantly misjudge their audience. This one is no exception.

View on HN

  Where's the innovation? Why is an exciting
  new desktop just another imitation?

Where does this obsession with innovation comes from? I will choose the thing that just works without getting into way over the innovation for the sake of innovation anytime.

I'd argue, that there is little left to innovate within desktop, there is only room for relatively small improvements — and that's natural course of thing in maturing media.

Take paper books: there was almost no innovation for hundreds of years: I say that's because their current form is optimal within constrains imposed by media.

With computers we have less mature media and different set of constrains, but eventually it will settle down and the only innovation will be possible outside the desktop, like some next big thing that will replace it.

Heck, Windows looks more like OS X nowadays. And even in the mid-90s, Linux desktops wanted to look like Nextstep, so the Mac envy seems a logical succession (let's better forget the CDE phase).

But the whole design space is really, really small. You're mostly making a distinction by a subtly different look and some procedural differences. This is true for every desktop since the 80s, whether it's System 6, OS X, RISC OS, KDE, GNOME or WPS.

That is what the audience wants. The amount of developers isn't that influential anymore, and even they are most likely to use Windows 95/Nextstep style interfaces since they started out using computers. The old Unix greybeards (and their spiritual successors) populate a very small island with weird palm tress.

There still are about as many, if not more people who use their DE as a glorified terminal multiplexer (or Emacs life support). But for them, "Linux on the Desktop" never was an issue in the first place.

I seriously doubt that anything really new will come on the desktop. Everything that tried failed (mostly integrated systems like Oberon, Smalltalk, Plan 9) for the "common user", or got integrated in the normal desktop (web apps, terminal emulators). Everything new seems to happen on different devices (mostly mobile), and even there I don't see a different frontier, it's mostly a mash-up of traditional desktop GUIs and web apps, adjusted for the limitations on the devices.

This is not innovation; it is an improvement. If you want innovation, look at Gnome-do, rhythmbox, gimp, and any upstream project. The purpose of elementary is simply to improve usability and add small usability-related features.

Isn't Gnome-Do just a Quicksilver clone? Rythmnbox just a take on iTunes for Linux? Both are great applications, and arguably an improvement on the projects that inspired them, but I wouldn't say they great examples of innovation.

Yes, and The GIMP is Photoshop for Linux. Great projects and all, but none are especially innovative on the surface.

I should say though, I think people put far, far too much weight on "innovation" and "originality" (not mentioned here but along the same lines).

Which apps do you think are innovative then?

The netbook remix UI

XMonad?

In what way?

Tiling window managers existed long before xmonad.

True. Actually one of the most impressive things about XMonad is as a proof of concept of both Haskell in general but also Haskell as a configuration scripting language.

That said, it was more important for me that it was a tiling window manager, not that it was XMonad (it's simply the one I use). Dwm would probably work just as well.

Innovation is really risky. If they aren't even competent enough to clone OS X, trying to innovate past it sounds like a recipe for failure.

Maybe. It depends on if we're talking about individual applications or just the desktop experience in general.

The thing about people using applications like Photoshop and Word is that they have established workflows that its users have come to expect. From a design perspective, they're pidgeonholed because any significant UI change means that they would have to retrain the vast majority of its users. So Adobe and MS make minor tweaks to its UI instead. Admittedly, MS made a big change with the ribbons ui element, but you get the picture.

Open source stuff has the advantage (I guess you could call it that) of having a much larger volume of new users and a much smaller volume of old users, so there's more room to experiment with different UIs to solve the problem of getting people to interact meaningfully with the programs. I'd say for the vast majority of the projects, the UI is an afterthought so it ends up being poorly designed. Even so, the opportunity is there.

The desktop is a little different though, there's already a load of user expectations that you have to live up to. The space for proposing solutions is a lot smaller. Riding on the coat tails of OSX may not be a bad idea so long as the developers understand the reasons why Mac made the design decisions that they did and adjust accordingly.

At least that's how I feel. Trying to innovate isn't a bad thing, it has to lead to failure first before you get it right.

Actually, I think the trick and the challenge is finding places where open source can innovate effectively.

An application like MS Office or Photoshop or OS X are huge and monolithic and thus don't lend themselves to dispersed development.

What would be good is finding innovative interfaces that allow presentation and data to be separate and thus developed separately - #1 huge example, the web. Making the desktop like the web makes huge sense.

You should look into Gnome Shell. It's not copying anything I've heard of.

I'd be ok with a linux based desktop that was functionally identical to Windows 7 or OS X. I suspect a lot of people would be.

Reply on news.ycombinator.com