Who says the open source and Free software world only copy but never innovate? Over at Canonical’s design blog – you know, the company which does nothing for the Linux world *cough* – someone named Christian Giordano has shared some ideas about how Ubuntu – and therefore, the rest of the Linux world – could make use of hardware sensors to better serve its users. Pretty interesting stuff.
The general idea is that Ubuntu could be more aware of its physical surroundings, with the user interface responding to how the user behaves in front of his display. This would need to make use of either cameras or proximity sensors (or both), and already, the guys at Canonical have come up with some interesting use cases.
For instance, a video playing in a windowed state could switch into a fullscreen state when the user leans back from his display. You could use a simple hand gesture or nod to show the launcher, or, go a bit further into the future, and tilt windows when users aren’t properly facing the screen (when using 3D displays, obviously).
The most interesting use case, as far as I’m concerned, is to automatically display notifications using a large font (fullscreen, preferably) when the user is not sitting in front of his computer – maybe he’s sitting on the couch or walking around. This way, notifications can be read from afar. Giordano already experimented with such a feature back in 2006.
A quick prototype was created for all this.
Parallax and fullscreen interaction via webcam from Canonical Design on Vimeo.
Interesting stuff this.
These days most webcams, especially those integrated into screen frames come with an activity light. It could become quite annoying watching a movie with a bright blue or green led flashing right in the eye.
Presumably the LEDs could be suppressed in software if need be. I’m not sure how they’d deal with unintentional input though.
I suspect that might not be the case. One of the functions of the led is the notify you that nobody is spying on you or that your webcam is active. If one could turn it off via software that could present a risk.
But then again there are those anti burglar programs that record videos based on movement…
Presumably also a piece of tape of appropriate color can help with the issue (or an official iBandAid for fruit-lovers 😉 )
Logitech camera I have software controllable activity light (I have pro 9000) and UVC software (guvc IIRC) can control it. I assume that this is a requirement of UVC standard.
as long as there is a way to turn it off, go nuts!
it looks annoying to me.
Agreed – computers being helpful is good. Computers failing at being helpful, bad. And I’m pretty sure that having the computer using a camera to try to guess whether I want windows full-screen or not falls under the latter…
The full screen notifications when you’re away seem cool. Some of these things are only ideas for what one would do with sensory input on computers. The truly groundbreaking uses, or more realistically, the “killer app” will probably come after the implementation. We’ve already “solved” the problem of how to interact with computers, but why not toy with a better solution? Sometimes reinventing the wheel is the right thing to do. You wouldn’t want a great big wooden wheel from a horse drawn carriage to be used on a train. A better example, who could possibly want an mp3 player? People can listen to music on the go with a cassette walkman.
Agreed – computers being helpful is good. Computers failing at being helpful, bad.
I can well imagine ripping my hair out in frustration with this thing if it were not possible to turn off. I don’t just sit in front of my computer neatly all the time, I turn around to play with my cats, I turn to the side to chat with my girlfriend, sometimes I lean back, sometimes forward, sometimes I am like the effing hunchback of Notredame..oh, and there’s kitchen right behind me.
That’s a lot of stuff that is in no way intended to be recorded and used to control anything but would undoubtedly be mistaken for such. Tracking eyes doesn’t really help either: plenty of people do not have their webcam sitting right in front of them but instead on f.ex. a shelf somewhere nearby. And eye-glasses, make-up, various kinds of eye-related diseases and defects and so on could easily fool the system, especially if the webcam isn’t top-notch.
Sure, this thing makes for cool videos and yes, it IS indeed interesting research. But I just can’t see it being useful in real life.
Indeed. Also, think about the additional eyestrain added by this. Positioning that’s not pixel-perfect makes things blurry, and blurry text is hard to read. If it moves, it becomes even worse.
So, if your blurry windows constantly move as you move your head, you’ll have a hard time reading what’s actually written inside them, inducing much eyestrain when you could have simply grabbed your mouse and clicked the window…
Cool feature, but useless as is, and has a high annoyance potential.
Edited 2010-09-21 16:27 UTC
edited
Edited 2010-09-23 19:16 UTC
If I can now spin the cube by waving my hand, Jedi-style, I’m all for it.
Now that’s what will bring Linux to the masses. Canonical hits the nail on the head once more
Oh, I thought getting a Linux distribution to actually properly run in the first place is the prerequisite for mass-adoption.
While all mainstream Linux distributions have great out-of-the-box support for many hardware combinations, GPU support is not the shining example for that.
It almost seems as if the three big GPU companies (Intel, AMD, and NVidia) sabotage Linux on purpose. While Intel on one hand is a big Linux contributor through MeeGo, Intel also makes it next to impossible to properly support Intel’s Poulsbo graphics out-of-the-box by Linux distributions. Before Mandriva fired almost all its personnel, the company at least maintained an easy installable package for Poulsbo with the binary blob extracted from some Dell OEM installation or so.
And even Intel’s FOSS drivers can cause problems as seen with the recent problems with KWin.
AMD needlessly fragments its own driver space by developing a closed source driver and a FOSS driver at the same time. Problem is, with its closed driver AMD is kinda slow supporting newer Xorg releases and it also can’t ship pre-installed on Linux distributions legally.
AMD’s FOSS driver OTOH supports the latest Xorg releases but then does not support the latest hardware. If Phoronix articles were correct, it takes AMD way longer than one year to add support for its GPU models (eg. support for the 2009 ‘Evergreen’ GPUs just starts to appear as experimental feature).
NVidia only develops a closed source driver. That again makes difficult to ship it with Linux distributions in a legal way. It has to be somehow installed afterwards.
If you’re lucky, your distributor has the driver is the software repositories but that still requires an active internet connection.
If you are out of luck, you need to install the driver by hand. And no, that’s not a simple double-click on an icon: Kill the X-Server, log in via terminal, change the install file’s permissions to executable, and then run it via command line. Great!
Novell at least helped in the past fixing the GPU troubles by developing radeonhd but financial troubles meant the end in that field.
Mandriva’s death means the end for the Poulsbo package.
The only commercial entity left trying to fix the GPU issue is Red Hat with its involvement in the Nouveau project but that one is still experimental for anything related to OpenGL-accelerated graphics.
There are a bunch of distributors out there with pretty deep pockets. Since it fits with this news item, Canonical obviously comes to mind.
I firmly think that spending corporate money on improving the FOSS GPU driver situation is way more helpful getting the operating systems on people’s PCs than some funny tricks with webcams.
In all fairness, I don’t think that Canonical is the only “guilty” distributor. For starters Orcale even ships two OSes that could use improved drivers: Oracle Enterprise Linux and Solaris (as both use Xorg). Linpus also comes to mind, several other Asian distributors gathered around Asianux (Red Flag,…) as well. God knows what Xandros does these days but certainly not contributing to FOSS GPU drivers.
Re-distributing and selling services around Debian is also a a big business but so far I haven’t seen anyone of those (with HP being the most prominent example) contributing anything to Linux GPU drivers…
Edited 2010-09-21 00:41 UTC
Anyone else think the fullscreen video guy is watching porn? 🙂
Yup ^^
Instead of toying with ever more experimental features (that typically introduce more problems then they’re worth) I wish Canonical would take just one development cycle and fix, once and for all, the myriad bugs, glitches and regressions that have crept into its Ubuntu operating system over the past few years. I know its user base would be eternally grateful.
http://labs.opera.com/news/2009/04/01/
Those who don’t understand Microsoft are doomed to reinvent Clippy, poorly. (With apologies to Henry Spencer).
How about instead of playing around with toy apps for video-based interaction, to do their jobs as OS vendors and provide robust and stable APIs for video device sharing and the like, and get them merged upstream, with broad support from other distro vendors.
Then, once that is done, and popular third party apps and video frameworks e.g. Skype etc. are using the new standard video APIs instead of attempting to simply open /dev/videoX in exclusive mode, then maybe it would be a good idea introduce the ability to do jedi moves.
Because if its done the other way (Lets write some stuff that doesn’t interoperate and breaks other apps that run on our platform), its going to be immensely painful for people who actually use their webcams. PulseAudio was bad enough, does anyone really want to go through all that again?
Its one thing for a developer to come up with a cool way to interact with the computer, but if the framework isn’t there to allow that interaction to happen smoothly and to share the resources required, its going to be an awful user experience.
Awesome. Now I can use my webcam to wake my laptop from hibernate.
Well after they fix both of those, maybe.
Only people who don’t know what they’re talking about.
I don’t want my windows move around when I lean to take a coffee cup. Well, the idea is good, but it must be made in useful and not distracting ways. The community could make an idea pool site for example.
Another thing is the laptop battery life. Video recognizing and camera will raise CPU usage, so the battery will drain faster.
“at Canonical … you know, the company which does nothing for the Linux world *cough* – someone named Christian Giordano has shared some ideas”
ideas are cheap. With Canonical’s track record I am not impressed unless I see ~ideas~ turned into working code
Well, here’s one Ubuntu idea (albeit, a bit old one now) that I wouldn’t like to see turned into reality:
“Canonical believes that in the near future, GNU/Linux users that are currently spread among a plethora of distributions will be concentrated around 2 or 3 main distributions.”
http://www.ubuntulinux.org/news/guadalinex
That’s the vision about the future of GNU/Linux that Canonical representatives have communicated to one of their derivative distros about half a decade ago. Maybe Canonical have changed their vision in the last five years but, unfortunately, I don’t see too many signs that they have.