Basically: You cannot make things dark. You can't use LCD type technology to block part of the lenses for the same reason you can't see scratches on your lenses: you're not focused on your lenses! You're focused in the distance. That means the ideal of having crisp, readable text overlays is mostly a fantasy. You can prove this yourself: get a piece of tape, write some text on it, and put it over your lenses. You can't read the text. You can even use clear tape to simulate what a real overlay would look like.
The second problem is power. Oculus makes sense because you're tethered to a high-end workhorse. When you're using glasses, you have neither a video card nor a huge source of power at your disposal. That means battery technology would need to improve by a massive amount just to make the glasses not die within hours.
The third issue is target recognition. The real world is instant. That means any overlay you use would need to attain 90FPS to appear seamless. If you're going to do the most basic thing with an AR device -- drawing an arrow on the ground, indicating where to walk, a'la Google Maps -- you'll need to recognize the ground + render the overlay within 11 milliseconds. This might work for a ground-based overlay, since you can assume it's a flat plane in a big city, as long as your gyroscope is accurate enough. But where would you even stick a gyroscope with that kind of precision inside a glasses form factor?
The best form factor for this would probably be some kind of "bandana", where the meat of the device sits on the back of your head. That'd look dorky and it'd be uncomfortable, but you might be able to pull it off.
The only way this might work is if you use LCD tech to dim the entirety of both lenses enough to then project light on top of them. But again, your focusing out in the distance, so trying to show anything readable (text messages) seems kind of a fantasy.
Here's proof: Pull out your phone and hold it up to your eye where lenses of glasses would normally be. I'm not sure if you could even render it in a way that would account for the focus blur.
Have you used any optical see-through HMDs? There's heaps, and they all have no problem displaying text. If you look at the spec sheet for the BT-300 for example [0] you'll see "Virtual screen size: 80" support (virtual viewing distance 5m)", or in other words you need to focus your eyes five meters away to see the display sharply, at which point it will appear to take up 80in of diagonal space in the real world (don't know why they mixed the units like this though). I'm pretty sure most people could read text off a 720p 80" screen 5m away.
As for not making things dark, yes that's true. You can of course tint the glasses or use an extra layer to darken the outside world, or you can really crank the brightness causes the user's iris to contract creating the impression of higher contrast in the dark regions. Best option is to combine the approaches with a camera observing the environment, which can even be done on a per-pixel basis as in the paper I helped write [1]
As for power/performance, it's all a matter of desired form factor. If you don't mind a cable to an arm or hip attached android device, you can have everything you need right now. Alternatively there's the Hololens style huge headset. But if you want slimline glasses yes you'll be waiting a few years.
A lot a people are bearish on the future of AR and I can appreciate that viewpoint, but the arguments you've brought for it are at best half true and I hope readers don't see this and think any of these problems are unsolvable (or indeed, haven't already been solved)
Whilst I am certainly supportive of open access science in general, and in the case of my own writing am glad to see it reach a wider audience, I'm surprised by the brazenness of linking to a pirated article whilst in discussion with its author. I don't know whether to be offended or impressed
Do you see a reason that, as an author, you wouldn't want your article to be pirated as widely as possible? You get no actual benefit from IEEE making money off of your article, right?
This is speculation, but perhaps there's a game-theory issue here. If IEEE loses money by people pirating (and you can apply the usual counter-arguments about whether the pirateer would have gone on to purchase it anyway) rather than paying then publishing costs increase to cover that loss, and therefore hurting the authors who must cover that increase in future if they want to publish in the journal again.
A better situation is where the preprint or open access version is available.
> then publishing costs increase to cover that loss, and therefore hurting the authors who must cover that increase in future
WHAT?! Does that mean it's the authors who pay publishers to get their work published? Is that really true? And then publishers also charge readers for what they were paid to publish in the first place?
I think the university insulates the author from the costs you mention. This also assumes that the market for articles is fair and rational, when it's anything but.
I think that science should be your ultimate goal as a scientist. Science is an inclusive goal, not an exclusive clique. You can certainly argue that your article was meant only for students and viewers who have access to IEEE's walled garden. But then you must also admit that your work isn't meant to benefit anyone else, except indirectly.
It also fosters discussion, since we can view your work and see whether it stands on its own merits, or whether it's related to the discussion past the title and the abstract.
The IEEE site also shows that your article has been viewed 238 times. My comment linking to the unimprisoned version was upvoted 5 times, and a good rule of thumb is that 4*upvotes = view count on a clickable link. (I've measured this in the past.) So your readership probably increased by at least 10%.
Lastly, I'm hopeful that IEEE and other institutions will eventually yield to the forces of technology, as the RIAA did. Attempting to stop technology by force doesn't have a good track record of success. It will be an unpleasant ride until then, but the world will be better off.
I think my manner of speaking was too wry. To be clear: I'm perfectly happy for people to read my research by whatever means they need to take. I too hope that scientific publishing gets its act together and moves away from hoarding knowledge behind paid subscriptions, and guerilla efforts to bring that closer to reality are appreciated. It was a tongue in cheek way of saying "I'm right here, you could have asked :)" (I would have said 'go for it')
To answer some of the sibling comments:
rspeer > You get no actual benefit from IEEE making money off of your article, right?
I'm well aware, although as an IEEE member I don't think it's fair to say no benefit, although it is very hard to quantify.
klibertp > what difference it makes for you personally, as an author and researcher, whether people access your work via a paywall or not?
None. As said above, I think my tenor wasn't communicated well enough. I was more just surprised that, given that sillysaurus3 and I have never communicated before, someone would take an action that could have potentially been quite upsetting (I know a few researchers where that would have definitely been the case).
klibertp > Does that mean it's the authors who pay publishers to get their work published? Is that really true? And then publishers also charge readers for what they were paid to publish in the first place?
Absolutely not in our field. As I understand it, in some fields there is a charge for very long form writing to cover the work required in reviewing and editing (though perversely the actual person who does that work is often unpaid), but generally pay-to-publish would indicate a low quality journal and would be an outlet of last resort.
I know next to nothing about the process of writing and publishing scientific papers, so forgive me if the question is stupid, but: what difference it makes for you personally, as an author and researcher, whether people access your work via a paywall or not?
The "crisp, readable text" thing is doable. Just put lenses in front of the text screen to make the focus appear in the distance and combine with incoming light using partially mirrored glass or a prism.
11 milliseconds on a modern CPU is a long time for image processing. You need a lot of grunt to do it, but it would work. It makes sense that the glasses should just be a dumb screen with some odometry, connected to a much more powerful data source.
I have no idea how they work but the Hololens can display text. If you piece of tape thought experiment was relevant it would apply equally to VR headsets but they also provide readable text and other forms of detail.
Regarding the battery problem, this is solved right now by having a cable run to your pocket. You can carry a 5000mAh battery in your pocket no problem.
You realize that the entire thesis of your statements are given the lie by the fact that Google Cardboard exists, right? It's literally holding the phone up to your face, and using the camera to take photos out the side. I've used it to filter reality
"AR" doesn't mean "projecting clear text onto the real world". It means retransmitting the real world from a camera onto an LCD with things added.
We literally already have the technology built into phones. You add some lenses and a battery pack and you're good.
That's hardly going to comfortable to keep your eyes focused on short distance for a long period of time. That's the problem with "holding the phone up to your face".
Your eyes are not focused on a short distance. The whole point is that the lenses make it so that your eyes focus to infinity while the device is very close.
Unlikely. I've elaborated on this in the past.
Basically: You cannot make things dark. You can't use LCD type technology to block part of the lenses for the same reason you can't see scratches on your lenses: you're not focused on your lenses! You're focused in the distance. That means the ideal of having crisp, readable text overlays is mostly a fantasy. You can prove this yourself: get a piece of tape, write some text on it, and put it over your lenses. You can't read the text. You can even use clear tape to simulate what a real overlay would look like.
The second problem is power. Oculus makes sense because you're tethered to a high-end workhorse. When you're using glasses, you have neither a video card nor a huge source of power at your disposal. That means battery technology would need to improve by a massive amount just to make the glasses not die within hours.
The third issue is target recognition. The real world is instant. That means any overlay you use would need to attain 90FPS to appear seamless. If you're going to do the most basic thing with an AR device -- drawing an arrow on the ground, indicating where to walk, a'la Google Maps -- you'll need to recognize the ground + render the overlay within 11 milliseconds. This might work for a ground-based overlay, since you can assume it's a flat plane in a big city, as long as your gyroscope is accurate enough. But where would you even stick a gyroscope with that kind of precision inside a glasses form factor?
The best form factor for this would probably be some kind of "bandana", where the meat of the device sits on the back of your head. That'd look dorky and it'd be uncomfortable, but you might be able to pull it off.
The only way this might work is if you use LCD tech to dim the entirety of both lenses enough to then project light on top of them. But again, your focusing out in the distance, so trying to show anything readable (text messages) seems kind of a fantasy.
Here's proof: Pull out your phone and hold it up to your eye where lenses of glasses would normally be. I'm not sure if you could even render it in a way that would account for the focus blur.