I am working on an UI right now that is inspired by Apple, and yes--it's based on Linux.

The long term goal is to put together an operating system. The short term goal is to create a intuitive, zooming user interace.

http://brevityos.blogspot.com/

Then the question would rather be if there are going to be any non-web applications for Chrome OS.

I would like to use Gmail just like I'm using any dekstop application. The favicon is already available, and the HTML title-tag says at least something about the title of the application. However, the favicon is often too small to appear in an application launcher that renders the icons around 48x48 pixels size.

One solution is to make a database of web applications containing the address, the title "Gmail" and also a high quality icon. This is how an user would make an application available:

* Right click a 3x3-icon app launcher, it flips around, and the rest of the screen dims down
* Available applications are displayed on the sides of the screen and can be dragged onto the launcher--who needs the app store?

I have two conserns for web applications at this point; privacy and low-latency access to the hardware for professional audio and video editing. While I'd love to use a web app to do something creative, I'd love it even more if I had complete control of my data. An encrypted cloud could solve this-or-it could be made possible to save the data locally.

Recording audio requires direct access to the hardware, and that can't be done with your typical JavaScript, HTML and Ajax programming at this time. Maybe with Silverlight, but that's out of the question for obvious reasons. We will hopefully get to that day where will be possible, and I think streaming, caching and even distributing applications by using BitTorrent might eliminate the need for software upgrades.

I see no reasons to separate how web applications of today and desktop applications are presented on the screen. Instead of living inside a tab in the browser, web applications should be treated like native desktop applications.

That's what I'm working on right now with the prototype that are using Cairo and Xorg.

To move the applications around on the screen, I need to convert from "pixel input" to vector points, and when I've managed to do that at desktop level, it shouldn't be hard to do the same with the content of the apps.

I've thought about showing an overlay with important actions when an application is to small to be controlled directly. When the mouse is over the app, a media player for instance can use 1/3 of the bottom of the app to show buttons for play, pause, next etc. But that's just an idea. If it doesn't feel intuitive, then it's a scrapped idea. :)

Feel free to mod me down if this is regarded as spam, but I'm working on a UI called Brevity, which will be based on Clutter.

The idea is to create a user interface that is like a huge wall that can be zoomed in and out of. You'll have apps that run where you start them, and you start them by clicking on an empty space on the wall. A 3x3 app launcher will pop up around the mouse pointer with the browser in the middle.

It's still in the early development, but you can follow my blog at http://brevityos.blogspot.com/

[start of senseless whining]

You got modded insightful for this? It bugs me a lot when there's a post like this in a discussion, it gets modded high, and people swallow it without even thinking. Here's your very own arguments, rephrased:

Pros:
You only have to choose one.

Cons:
End users need several GUI toolkits installed.

Discussion:
If there's a standard, you only have to choose one as well. Your only valid point is about the end users that needs several GUI toolkits installed, which you categorized as a downside--which it is!

a) What does these words even have to do with the discussion?
b) Could be used as an example on Wikipedia about the logical fallacy that is "two wrongs make a right"

[end of senseless whining]

Thanks!

I registered after having lurked here for half of my life. I'll share my thoughts, and it would be interesting to see what response I might get.

I've been thinking about that maybe it's time to let Linux be what it is, and start fresh with the goal to make an open source desktop OS that doesn't get in the way. The world is different now and when Linux was started in 1991. It might actually be a good idea to rewrite the OS every fifteen year or so.

I've got some ideas for a new OS.

It consists of a kernel, an UI, a browser and some basic applications. This part of the OS is open source. Then there's an app store or something similar where the user can buy applications, games, professionally designed typefaces, proprietary codecs etc.

It will be the first OS that is resolution independent. We wouldn't even need anti-aliased fonts on the screen if monitor vendors started to increase the DPI. So that means that there will be vector graphics from the first pixel drawn after the boot loader. The default settings should favor what's intuitive and non-intrusive for the average Joe.

I can imagine an UI with blues and grays, light gradients, mostly a flat look and some light shadows. Caching, timing and redrawing will get a lot attention, to keep the feel of the UI rock solid. Hot spots should be made the most of.

I think history has shown that there's nothing wrong with the start-like menu, task list, clock and a desktop in the background. OS X and Linux have got the tray right for the most part. It's for wireless networks, volume changes and similar. The user can install a program by clicking a link on a website or by using the app store-like program. The app store (or whatever it might be called) should keep track of the updates. Distribution of packages could get done by BitTorrent or similar technology. The death of mirrors and package maintainers.

I think the most important part is that you shouldn't think about that you're using an OS. Microsoft have experienced with browser integrated into the desktop. Most people didn't like it. KDE 4 placed the desktop icons in a box. Most people didn't like it. Let's draw from operating system experience since the beginning, and use what worked. Throw away what didn't.

And then, let's talk about how it should be organized. One centralized website with an unified look. It should be easy for people to suggest ideas and comment on them. The ones who are making decisions will have to favor the public opinion rather their own. The feedback from the user should be taken very seriously. A release schedule will be set up that in the best way benefits the whole system. The Scrum process might be used as a model for the development in general.

The challenge is to convince people that this is a good idea. I will donate my time for free and lead the project, if there's any interest. I've waited for an usable Linux desktop for over a decade, and I'm done waiting. That's not to say that I don't respect the work that has been done in Linux-land. A new OS would benefit from a lot of the Linux kernel source that has been written, and all that the world has learned about Linux. And Linux won't die. This will be just another experiment, but with different organization and goals.