Uncanny Valley? What Uncanny Valley?

I have to tell you, I'm at the point where physics rendering engines are getting ridiculously scary good.  In the YouTube video below, the room is unbelievably real.  But more importantly, it's not the Miku model that looks so real, but the shadows she's casting! 

Wow.  Just.... wow.

http://youtu.be/57EiJJdrlbU

The Elusive Master Wallet

Ever since Google made their announcement that they would be doing a wallet for their Android OS, It seems that everyone has been jostling for a piece of the pie. Google, Paypal, Amazon, MasterCard and Visa have all pushed to be the premier "go to" service.  But not one of them has been able to see beyond using "coin of the realm" for making purchases.  Which is really limiting themselves.  Why should I use any of their wallet systems, when all I need to do is swipe a piece of plastic?

The world is changing.  Anyone that doesn't realize that fiat money is all digital really needs to educate themselves.  Especially now that FinCEN has declared that true digital currencies earned from gaming can indeed be tracked, regulated, and most importantly, without emphasizing it, taxed.

What *I* need and want, not what some financial conglomerate *thinks* I want, is an app that I can put my Gaia Online, Second Life, IMVU, Bitcoin, and whatever other digital points and coupons into.  I want to use those values, because FinCEN says they're values.  AS fiat money!  That app needs to go to whatever international financial center that says "On this day, that measurement of value is worth such and such. Go spend it!"  I shouldn't have to go to a third party exchange and turn L$ into US$.  Now I have nothing against third party exchanges.  They're the ironic link between the virtual and the physical.  But the tech exists to do this automatically.  Especially when the exchanges themselves track the  values daily.

Until these guys stop bickering over whether NFC or Bluetooth 4 is the best carrier and working with FinCEN or whichever agency to get those digital values, I just don't see the practicality of pulling out a wireless device, picking whichever service I choose to use, all of which have my credit card numbers, btw, and "tapping" it.  It's quicker to grab the trusty leather pouch and swiping that thin plastic card directly.

The Contextual Web

I learned of an approaching paradigm shift this week: the contextual web.   Normally, I would file this bit of fluff in the back of my head as "that's nice", and forget it about 10 minutes later.  But I decided to do a quick Google for the terms "conceptual internet", and got a link from scobleizer.com.  Normally, I file anything from Robert Scoble in the back of my head in the same manner.  But this time it made me raise my eyebrows.

Mr Scoble is one of those personalities that you either accept anything he does with open arms, ignore anything he does, or keep him at arms length because of the manner he does things.  I've learned to follow him on the odd chance that he will turn up some interesting tidbit of tech, or a stinker I want to stay away from.  Case in point; I watched an NBC Nightly News report where Robert was praising Pinterest.com.  I tweeted that just because Scoble was into Pinterest, I was not going to flock with the masses.  Or something to that effect.

I read his blog, and was thoroughly sold on the contextual web.  It is a concept that absorbs "social networking", "ubiquitous computing", "cloud", and the "semantic web", and attempts to make sense of them all.  We seriously need to pay more attention to this in the upcoming months.  Not years.  Months.

And BTW, I've recently opened a Pinterest account.  I found something there that I'd been looking for for months.  Not because of Robert Scoble.  Because of Google.  But I had to at least acknowledge that Robert was right on this one, too, even if I fought it tooth and nail.

Test, Test, Test...

It's time I dust this off and get back to work, I think.

Avoiding the Uncanny Valley

I've been pondering how realistic the models are in many of the Miku Muku Dance videos I watch on YouTube.  Some movements are painstakingly crafted in the program.  And many are motion data that was created by a live dancer using capture software.  Many are so pretty that you can seriously forget that you're watching a computer program.

It can be argued that the Japanese have been leading the way with robotics and artificial life.  The illusion of life and realism seems to be their driving force. So how far can they go before you get to a point where your machines and programs become creepy?  Quite far, actually.  MIT did a report a few years back just to understand the creepy factor.  It says that for a robot to be considered in the uncanny valley, facial features need to be just a bit "not right" (my quotes).

So how does Japan avoid the creepy factor?  In my opinion, it's in using manga and anime facial features.  Stay with me on this.  A human body can be created physically or virtually to a standard of perfection, and everyone 'oohs & ahhs'.  The chest needs to be larger, or the bottom smaller, or the male parts more endowed.  Or even included!  Put a human head on it, and it scares the bejesus out of them.  If that same perfect body has an anime head on it, with it's cartoon eyes and neutral facial features, those same people have no problem whatsoever with accepting their new best friend.  I know I don't.  There have been some videos on YouTube that I've watched dozens of times just to see the body movements in dance.

Are we desensitizing ourselves from the foundations of 'true humanity'?  Not really.  We're just broadening our concepts of what is humanly acceptable.  Isaac Asimov had seen this pattern years ago when he wrote his robot and foundation novels in the 1940s and 1950s.  I think the acceptance of anime characters in games and videos is paving the way for a broader acceptance of robotics and AI in general.  I've been in the uncanny valley a few times now.  And anything that bridges that valley can only be a good thing!

2013: The Year That AR Goes Mainstream.

Augmented Reality has reached critical mass.  There are dozens of apps that offer just about anything you can imagine.  I can lie in bed, half asleep, and boot Junaio, and see stuff surrounding me.  The only thing keeping it subdued is the lack of an affordable head mounted device.  Google is posed to bring that cost down in the future.

I've always daydreamed of having a pair of glasses that allow me to interact with virtual objects and people on the fly.  While having an HMD will facilitate that, it isn't the only piece of hardware involved. A separate camera in the room, tied to geospacial coordinates to observe your position will be necessary, I believe.  The best piece of hardware I can think of is the Kinect.  Your visitors can be virtual or physical, in the room with you.  Anyone in a virtual reality can invite you in a location and your recorded image becomes your avatar.  The best thing about that is you will feel as though you are physically in that location. And it all happens in real time, 24/7, without having to start an app.  It just works.

Granted this is just January. Realistically, prices and hardware won't be widely available until mid to late 2013.  But I'm willing to go on a limb and be firm on this being THE year!

Committing to HTML 5

I've been really intrigued with HTML 5.  It promises to do a lot for the web, and sometime in the future, the "internet of things".  But the catch is that most people are comfortable working in HTML 4.  We all have our templates set in 4, and we just don't want to take the time to recode all of that stuff.  That, and we're not nagged to enter metadata when we're using 4.  But there is just one thing that's wrong with this kind of thinking.  Sooner or later, the people that are working in 5 are gonna leave us all behind like dinos.

Now I'm not saying we all have to do everything in XML.  That would be like trying to write in binary, when using C++ is so much easier.  But even the most dense of us are starting to notice that the web is heading towards HTML 5.  And I'm cool with that.  I've done a lot of switch over from HTML 4 using XHTML.  And even XHTML is getting long in the tooth.


I was just reviewing some files I have in my 3d worlds folder. As it turns out, much of that work is bits and pieces of VRML coding.  I was reminded of just how challenging VRML is.  It was created back when HTML 3 was in use.  It's one of the reasons I've taken an interest in X3D for doing small projects.  It's designed to work with HTML 5.

There are websites like Second Life, and programs like Open Sim that you don't need to worry about the magical incantations to get a world up and running.  And many works I've seen are just stunning.  But there are just a couple of tiny problems.  A world needs to load directly from the servers for me, even though it supposedly is cached somewhere on my machine.  And the realism shatters whenever I see an incompatibility in prims.  The little glitch line that is constantly pixellating when you least expect it.  You know what I'm talking about.  I've heard it called the "Denno Coil" effect.

Now one thing that VRML/X3D is great at, is eliminating pixellation.  That, and caching a world on your machine, rather than calling it from the cloud.  Of course you need a plug in.  I've come to accept that.  Eventually most plug ins will become apps.  The one I prefer is BS Contact, from Bitmanagement.de.  I discovered recently that many web based games are written in VRML and X3D. Of course there is some Java and Javascript involved as well for that, but I'm already babbling for too long on this.

What I'm trying to say, is that I'm going to recommit to upgrading my templates to HTML 5, and take another shot at learning X3D.  While I'm not going to become a master of X3D, I hope to be able to learn enough of it to do a few projects that have been in the back of my mind.  So we'll see what happens!