Nah dude, \pr0n is a new escape character.

Fairly intuitive??? It took me about 20 minutes just to figure out where the fucking bookmarks were. The minimalist interface implied quite firmly that there was simply no bookmark functionality. Oh I have to hit Ctrl-B to show them? What the hell is this, emacs? How is that intuitive?

No real Adblock Plus. (The Chrome plugin doesn't actually block anything, just hides it. And no, I don't want to roll my own hosts file. It's a pain in the ass, and not nearly as versatile.) No CookieCuller. The tabs bar doesn't even scroll, it just makes existing tabs smaller and smaller as you add more. Unfortunately Google doesn't seem to care about fixing any of the things people have complained about.

But... but that would leave the poor man penniless!

Wait a sec... I think I recognize the background. Is that shot from inside Steward Mirror Lab?

Quite often the write-once nature of optical disks is a positive. Burn an OS install disc once, finalize the disc, and run a verification on a separate computer. You can be reasonably certain the install is "clean", and it can't really be tampered with. But if you instead put that installer on flash media, what's to keep compromised software from later rewriting the bootloader or modifying the installer in some way?

Systems do get rooted, and sometimes reinstalls from known clean media are necessary. If you reinstall from compromised media, you are actually worse off, because you get a false sense of security. Unfortunately most USB drives still don't have a read/write physical switch. Even if they did, I'd be reluctant to use them in some CYA environments; I can prove that my burned DVD of Ubuntu LTS could not have been modified after the disc was finalized. Can you say the same for your USB media? Same goes for backups. I love USB flash for its convenience, but it is actually a disadvantage in this situation.

>USB drive then?

If you're going to do that, then you might as well just make an intelligent crypto token that generates a sequence of numbers according to some known algorithm. The device should have a set of buttons (akin to a small PIN pad) where the user enters a known sequence of buttons on the device itself. Online bank software either queries the device directly as USB (which may introduce other security issues) or has the user enter a set of numbers from an onboard display, in addition to their username and password. A single PIN entry allows a single login session. For extra security have the user press a "confirm" button on the device and perform another verification every time money is transferred or other sensitive operations take place.

Prevents access via software keyloggers, because the buttons are on the device itself. Provides two-factor authentication, making phishing attacks a little bit tougher if done correctly. Should be reasonably cheap. And it's a lot more convenient than booting into another OS to do your banking.

> 1. You say each LED is not collimated or directional but then you mention a microlens system. What does this microlens do, if not collimate?

Think. Why are lasers of such importance? Why can't we just use LEDs with mirrors and lenses to accomplish the same thing as lasers in optical drives? The reasons here are very similar. There will be leakage, there will be diffraction, and the light won't focus cleanly on a single region of the retina.

> 2. Contact lenses move with the eye.

That's exactly the problem. When your eyes move the patterns from the outside world "move" across the retina, and the visual-optical response system can function properly. This set of lights is stuck to the front of your eyeball, so the light emitted by the LED array does not move. The way to solve this is to have some very intelligent circuitry that can pan the LED patterns on the display along with the eye movements.

Normal contact lenses do not produce light. They act as a surface to modify the shape of the cornea in order to fix aberrations in the lens system. (The lens inside your eye is one source of refraction, but the boundary of your cornea with air is the other major one. This is why refractive eye surgery can correct your vision.)

Does this make more sense?

There are several difficulties with this type of system that have prevented it from becoming a reality. Here are a few:

1. This is too close to the eye to be able to resolve focus in most situations. The light isn't collimated or directional (it appears to be focused with some sort of "microlens" system), so one LED turned on can spread out to stimulate a wide patch of retinal cells. With any regular LED system you'd just see a big blur. For information requiring a single light this isn't a problem (flash an LED on/off under certain conditions, or change the color) but anything more will require something which can project cleanly onto the retina. This is not a trivial problem.
2. The detail-oriented part of your retina is near the center, in a part called the fovea. While you think your vision is equally clear across a wide range, this is actually a trick of your brain. Your eyes are quite sensitive to rapid movement (low latency) on the edges, and more sensitive to detail in the center. When observing fine detail such as text, your eye actually "scans" an area and forms a larger, detailed image from the composite. Even if you could project the light cleanly 1:1 onto the retina, for any textual/HUD information you'd have to track eye motion very precisely and provide the information that the brain "expects" to see at each point. And again, the light has to be projected onto a very small part of the retina.
3. Retinal cells can get easily overstimulated, much like the burn-in on old CRTs. Even when looking at one object of normal intensity for any period of time longer than a few seconds, your eye will "jitter" back and forth. This involuntary movement is called nystagmus, and your brain compensates for it. (The rhythm changes when alcohol or drugs are ingested, which is why nystagmus tests are part of a DUI test.) Lab tests have shown that when the eye is physically restrained from moving in this way, objects effectively become invisible to the subject. So any 1:1 projection would also have to track nystagmus and then "jitter" in the same way as the eye, or the conveyed information would also become invisible.

Excellent reminder. I wish I had mod points right now.

Wow. Do not post when preview is broken.

Why don't you just drive there?

... ohhhhhh nevermind, I see! :)

Is it only me that I care for functionality and not decoration?

Two points:
1. You're not the average user.
2. Good UI design exposes functionality unobtrusively; it's not about airbrushed icons, and it's not just aesthetic.

Icon choice is only one minor facet of UI design. When you do File->Open in a native Windows application, it will probably look like this. The appearance may vary slightly, but for a given operating system revision the buttons are always the same size and in the same place. That is to say, the open file dialog is consistent across all applications, and a user familiar with the OS immediately recognizes it. Familiarity encourages your users by giving them a successful outcome essentially for free. "OK, I can do this." GTK+ on Windows ignores these conventions, giving the user something like this, or something like this. Though you might think they are stupid, 9 out of 10 users will be utterly baffled by the disparity here. They expect to see a Windows file open dialog, and they see... something completely new. They don't hate it because it doesn't look pretty, though I'm sure at some level that comes into play. More importantly they're thinking, "Oh crap, new interface! Man, if even opening a file takes this much thought and effort I hate to think what the rest of it is like." If they don't give up on the program outright, they'll have to spend an extra minute or two learning something they should already know how to do. Those are minutes they could have spent learning your program! Sit those same people in front of the same dialog in an Ubuntu context and they will chug right through it, mostly because they expect there to be a bit of a learning curve. But something as simple and ubiquitous as an open file dialog shouldn't present any learning curve. Do you see why?

If you're the only person using it, then you can design it using whatever interface you want. By definition you become familiar with it as you code it, and your interface will be designed around your logical conception of how the program works. I write most of my internal tools for CLI, just because it's easier. But if you ever expect non-programmers to use it, don't be surprised when they hate your program (and you) for making them jump through unnecessary hoops. "Why didn't the developers spend the time to do it right?" Things like whether or not the menus function as expected can make or break a project, and even an introductory UI and usability textbook should make this clear. Learning a proper cross platform kit takes a few hours at most, and makes everything feel loads more polished. And it will make your programs look a lot more professional if you ever decide to port them.

I'm struggling to think of a car analogy, but GTK+ applications under Windows remind me of a beat up car with primer gray doors, except that the seat belts pull up from the waist and buckle at the shoulder, and the ignition key turns the opposite direction. This is a trend we desperately need to change. The question isn't, "Does it function?" The question is, "Can the user operate it with minimal effort?"

So in other words they're being smug assholes to people who use one OS versus another.

Classy.

And the reasons for this are political, with a side of history. The GIMP developers invented GTK+, and now they're tied to it.

Know what? GTK+ is great under X11. It looks and behaves like crap under everything else, regardless of what theme you select. Basic UI principles say that your applications should be consistent with the OS, and that means using standard widgets (menu bars, icons, buttons, file open dialogs, and other things that match the look and feel of your OS). When GIMP was first released there WERE no standard widgets for Linux, so the devs hacked some together and released them as a separate library. A couple other devs saw the work that had been done, and figured that GTK could be used to save work on their own projects, and before long a ton of apps and window managers used it. Some of those app developers wanted to port their apps to Windows and Mac OS, and so GTK+ was ported as well. But because GTK+ manually renders all the buttons and widgets and so on, the ports look out of place. Strike that. They look godawful. Really this isn't GTK's fault, it's just not the right tool for the job. It's not just that they look bad, but users have to learn things like how to interact with a new file manager. It's unprofessional, it robs the user of time and effort, and it makes ported open source software seem inferior to native apps.

Recognizing the inherent problem, several other developers made toolkits so that apps would look normal again. A wxWidgets program compiled for Linux uses GTK+ to draw the dialogs and menus, but calls the native widget functions under Windows. The result is a program that looks like it was designed specifically for each OS on which it runs. Take a look at screenshots of Audacity for a great example.

They could design the UI in wxWidgets or Qt to make it actually look decent under every OS, but they won't-- and really at this point they can't, because the former would be seen as pandering, and the latter would be seen as abandoning GTK+. But to everyone outside the community, it looks like the GIMP devs are rallying to prove the superiority of GTK+ using the flagship Linux graphics app, at the expense of the open source movement. It only pisses off those of us who are trying to ease migrations to a free OS by gradually replacing existing non-free apps with free alternatives like OpenOffice. OO is a drop in replacement for MS Office in many cases, and behaves almost exactly like a native app under Windows. On the other hand, Inkscape is a great program with a decent UI, but I can't wholeheartedly recommend it as an Illustrator replacement to Windows users because it doesn't look or act like what they're used to. And if I can't get my mother weaned off the crippled photo editor that came bundled with her camera, I'm never going to get her to switch completely.

Face it, folks, GTK+ is cross platform only by the loosest possible interpretation. I realize a lot of time and effort has gone into the 2.7/2.8 "redesign", but these UI changes to GIMP are too little, too late. At this point the only thing which is really going to save The GIMP on other platforms is a complete UI redesign using something other than GTK+. If GEGL is ever finished this shouldn't be too hard. Conveniently this would also allow us to change the cringe-inducing name as well. The result would be a Photoshop replacement that would look like it wasn't cobbled together by two bearded guys in a basement.

Wrong summary. I emailed the editors a minute too late I guess.

Slashdot has editors??? As in, people who look at the stories and fix errors before (or after) hitting the submit button? You must be new here.