The AlpineInker : E-Book

The Microsoft Research Codex: Are Dual Screens the Future of Mobile Devices?

Ken Hinckley — Wed, 01 Oct 2008 10:17:00 GMT

Never buy one of anything. That's advice you should stand by when you're buying unusual gadgets. The advice was good when Randy Pausch offered it to me some 15 years ago, and it's still good now.

Of course, with 18 month old twin girls at home, this has become second nature to me. Two boxes of diapers. Two gallons of milk. Two Elmo plush dolls.

Oh, and yes, of course. Two screens for my tablet computer.

Dual-screen devices have become the subject of increasingly elaborate internet fantasies. Now, rumors about an Apple "Brick" device have stirred up dreams of future tablets, such as the alluring One Laptop Per Child v2.0 concept photos, which now orbit the internet once again.

My interest in dual-screen devices goes back a lot further, though, and was really spurred on by the University of Maryland dual-screen e-book reader project. That effort is led by Francois Guimbretiere, who is a long time collaborator and friend. I had some ideas to build on what his team had done, but also to take things in a different direction.

First of all, I wasn't interested in an ebook reader. I wanted a device that was all about writing. Sure, reading and writing go hand in hand - you encounter cool ideas and search out reference material on the web-but what I wanted to build was a tool for thought. To me that means a tool with writing, sketching, and annotating as the core of the experience.

I'd been thinking for a long time about picking up an OQO Model 02 computer. My team has an extensive code base for pen-and-tablet functionality resulting from InkSeine, and the OQO runs it out of the box. It's got an active digitizer for high fidelity pen input, and it's the smallest slate Tablet PC that money can buy. That's a pretty good start on a small form-factor tool for thought.

So when it came down to it, just how many of those OQO Model 02's do you think I purchased?

Project "Codex" was born.

Twice the Screen at Half the Size

Here's the Codex packed up for mobility. It folds up quite nicely and has a moleskine-style knitted elastic strap to hold it securely shut. There's a loop for the pen and a mesh pocket so you won't lose small accessories, business cards, or receipts that you collect in your travels.

Sure, it's a bit of a brick at present. The whole thing weighs just over 2 pounds. The OQO's are considerably thicker than I'd like. But my goal is to prototype the future as quickly as possible and start living it. The OQO offers a handsome, jet black time-travel machine in that regard.

Don't underestimate the ability to quickly pack up with the screens protected. Folding the Codex in half makes it comfortable to carry and easier to stuff into my gadget bag. It's a self contained kit for ultra-mobility.

The Codex is Not a Container for Dead Trees

A codex is just an archaic term for a bound book. But this Codex is unlike any book that you've ever read. It's not a long linear text that you flip through. To me, there's no use in going to all the trouble to build a dual-screen tablet prototype and write elaborate software just to mimic a traditional book. This is the 21^st century and it's about time we moved past containers for dead trees.

We have lots of information sources. We have multi-tasking. We have hyperlinks. We have split attention. We have a left brain and a right brain and we rarely do one thing at a time any more.

A Separation of Concerns

The Codex has two screens, it's designed to be used that way, and you won't find any half-apologetic demos that try to mash them back together into one big screen. Instead, it's all about the intelligent partitioning of tasks and interface elements across the screens.

Above is one example where I'm working on a blog post that I've had planned for a while. On the left I have a whole bunch of cool photos that I found tagged with moleskine on Flickr. I was browsing through these as inspiration for our InkSeine digital note-taking software. On the right I'm organizing bits and pieces from these photos along various themes. So I just take a snapshot from the collection on the left screen and it appears in my notes on the right screen, where I can arrange it and mark it up as I see fit. I can scroll back and forth on the left screen to find a photo that meets my current needs, while the page that I am authoring on the right screen always remains visible. The two screens are invaluable because I always have the reference material in the context of what I am working on, instead of feverishly flipping between them on a single screen.

Can I do this on a single large screen? Well, sure, I could monkey with the window placements and get everything arranged just so. But that takes a lot of effort and the temptation to expand windows to take over the full screen is hard to resist if I have to expend effort to do so. A dual-screen device that understands the partition between the screens gives a much simpler experience where I don't have to constantly manage the set-up of the windows.

Navigate without Losing the Big Picture

Here's another example. I'm further along with authoring my blog post now, and I have a bunch of material floating around in my notes. I create a page that is a Table of Contents, with links to several themes that I've identified in my Flickr moleskine investigations. If I open a link, such as my Creative Collage page, it opens on the opposite page. I don't even need a "back" command to return to where I was - I still have my navigational structure on the left, side-by-side with my content page on the right.

Unlike a traditional dead-tree book, I have no physical restriction that forces me to view consecutive pages - The Codex lets me follow links or flip through the screens separately to view any two pages together.

Working Big and Small

We can draw some inspiration from traditional media as well. In magazines and books, sidebars are a distinct section of a page that augments the main text with auxiliary information. Well, the Codex has sidebars on steroids. I can take any chunk of my notes, make it into a sidebar, and then arrange a bunch of these on a page. Here, I've made a storyboard page consisting of six sidebars. When I tap on a sidebar in one of the storyboard cells, it expands to full size on the opposite screen. I can plot out the broad sweep of my story on one screen, while maintaining full access to the zoomed-in details on the other.

Your Mother Was Right - Posture is Important!

Perhaps the coolest property of a dual-screen device is that the compact, mobile form-factor encourages shifting the device around. I can orient the screens with respect to one another, stand the device up, look at the screens in portrait or landscape, and so forth. In fact, the Codex supports about a dozen different configurations. We call these postures. (A colleague, Michael Miller, coined this term).

Add two cups of accelerometers, a dash of flex sensors, and bake with some simple software to fuse it all together. Out of the oven pops an intelligent dual-screen display system that configures itself depending on how you arrange it.

For example, if I want to work in landscape, I can flip the device into the laptop posture, as opposed to the book posture that I've been showing so far. Now I have one screen that's angled for easy reading, plus another screen that's horizontal for easy writing.

Or I can use the laptop posture to hook up one of my screens to a projector. I put the public part of my presentation on the top screen, while the controls to drive the presentation and my private notes are confined to the bottom screen.

Now perhaps I meet a friend in a café and I want to show him what I'm up to. I can just lift up on the binding of the Codex to angle the screens so that one is facing me and one is facing my friend (below, left). We call this the battleship posture - as in "You sank my Battleship!" Each person has one screen with a private view. The Codex automatically configures the software for shared whiteboarding so the collaborators can mark up the screens, pass notes back and forth, and other such foolishness.

Next I drop the binding back flat. Now both of us can view the screens, but one screen is oriented towards each person.

Don't Get Too Attached

Something that turned out to be surprisingly useful is that the Codex allows me to pull each screen right out of the binding. A firm pull pops it out, a firm press pops it back in. This is really handy for laying out the screens to suit my work, or to review a video while I jot down some notes. I can even hand the screen to another person to show them something - without actually giving them an electronic copy of the information.

What? You Still Have a Desktop Computer?!?

The Codex is just another wireless device so of course I can connect it to my desktop computer. Now, any screen capture that I take from my desktop screen shows up in my Codex notes right where I left off. For web pages, the snapshot comes across with a hyperlink back to the source, so I can easily revisit it later when I'm reviewing my notes on my Codex. This makes it the perfect companion no matter how I'm working.

A Parting Shot at Text Entry

The Codex is designed around doing most things with a pen, but I'm no idiot. When it comes to text entry, a keyboard is a good thing.

And that, in part, is why I'm puzzled that so many of the dual-display concept designs that I've seen take up a whole screen with a virtual keyboard. I'm sure people in focus groups ask for this. But I still think it's a really bad idea. What is the point of having two screens if you are instantly going to cover one of them with a picture of a keyboard? Your hands immediately occlude it from sight anyway. At that point I'm basically using a single screen again, so I might as well just grab my laptop to get a decent keyboard.

The Codex offers a solution for text entry that I much prefer, by virtue of the OQO Model 02's mechanical keyboard. I just slide the screen up and that reveals the keyboard when I need it. Works great. No dorky touchscreen keyboard. I still enjoy the full benefits of partitioning my work between the two screens. That's the way I like it.

When Can I Buy One?

The Codex is a prototype-- and a rather flaky, cobbled-together one at that. But it uses off-the-shelf devices, and there's nothing magical about the software. So the crass answer is that you can have one now if you are willing to spend some dollars, build yourself a custom binder, and write a little bit of code. That's how I started. My first prototype was a repurposed day-planner with Velcro holding the screens in there. Install a shared clipboard utility and you can start copying and pasting between screens. That will give you just enough of the experience that you will hunger for more.

What would you want out of a dual-screen device? What capabilities would make it most useful to you? How do you see dividing your own work between two screens?

In the meantime, the trajectory of ultra low-power "e-ink" displays bears watching. Check out the recent Plastic Logic device, for example. Right now e-ink is an abomination for anything interactive, but eventually some display technology will get where we want it to go. Low-cost, low-power screens are crucial to make dual display devices a practical consumer device, rather than a research lab curiosity.

The sheer number of concept devices that have popped up in the last 6 months suggests that dual-screen devices are poised to take off in the near future. My hope is that our research on the Codex can help in some small way to unearth the full promise of such devices. I, for one, am convinced that dual-display devices have a well-motivated role to play in a future ecosystem of mobile devices.

In my office, at least, that future is already here. It's just not evenly distributed.

Acknowledgements

I would like to thank the following key contributors to this project:

Morgan Dixon
Raman Sarin
Francois Guimbretiere
Ravin Balakrishnan

Research Frontiers: New Stuff Coming in Pen & Multi-Touch Interfaces

Ken Hinckley — Sat, 05 Jan 2008 09:53:00 GMT

The annual Computer-Human Interaction (CHI) conference is fast approaching (April 5) and the advance program has been published. That means AlpineInker can scan the list of accepted papers and see what looks cool - but the actual papers aren't available yet, so it's a fun guessing game to read between the lines and try to figure out what the research community has been up to.

CHI is considered by many to be the flagship conference of the broad discipline of human-computer interaction, so it's important to be aware of what developments are coming. The papers are all rigorously reviewed and it's difficult to publish stuff there - typically only about 20% of the submitted papers get accepted in any given year.

I also learned last week that I'll be the technical program chair for the CHI 2009 conference. That means I just got dramatically busier (as if 11 month old twins weren't keeping me busy enough) and I probably will be posting to this blog a little less frequently than I originally planned to.

Anyway, here are my picks for what look to be some cool devices and techniques that we'll all hope to see become more commonly available as soon as possible.

Navigation Techniques for Dual-Display E-Book Readers

Authors: Nicholas Chen, Francois Guimbretiere, Morgan Dixon, Maneesh Agrawala (University of California, Berkeley - University of Maryland, College Park. Full disclosure: Francois, Nicholas, Morgan and I have co-authored papers together; Maneesh is a former Microsoft Researcher who has also co-authored papers on pen stuff with me.

Imagine the Amazon Kindle meets the Nintendo DS - only much cooler. This is a project that Francois' lab has had going for a while, and from earlier demos they've done, I know it's a device with a pair of displays that are joined together by a hinge. So you can read stuff just like you would with a real book that has two pages facing one another. But you can also flip the two displays back-to-back so you have a two-sided UMPC type of device, or you can even separate the two screens at the hinge so you have two independent devices that you can use like separate sheets of paper, with different documents on them, and they stay synchronized using a BlueTooth connection. They've hacked up each of the displays with a motion sensor that lets you flip through pages much like you would leaf through the pages of a paper document. I would love to have a dual-screen UMPC like this that I could ink on - it would be the ultimate digital moleskine notebook. It will be very interesting to see what new developments they have come up with for this paper.

Update: Their web server has been having problems, but since the original post, I've been able to access their site briefly. They have a page about the dual-display ebook reader project and a video (WMV) showing how the motion sensing works.

I also found a few images. Here is what their original prototype looked like in the dual display configuration:

And here it is with the pages pulled apart:

Finally, here is a tantalizing shot that apparently shows the new version of their reader. It looks like they have gotten it dramatically thinner and lighter. I'm also excited to see the pencil next to it and what look to be small writing styluses (or is that styli?) attached to the bottom of each screen - maybe that implies we will be able to do inking on this thing some day!

In fact, they say that they prototyped this using a 3D printer to make parts of the case and a carbon-fiber composite bezel to really keep the weight down. It's amazing what a few clever university students can pull together these days.

Update #2: This seems to be an extremely timely bit of research given the announcement today of the e-detail Dual Display Tablet PC:

The e-detail product suggests another great scenario for the University of Maryland dual-display e-Book: set it up like a picture frame so one display faces each person in a face-to-face meeting. That would be fantastic for a highly strategic one-on-one meeting, such as playing "battleship" .

K-Sketch: A "Kinetic" Sketch Pad for Novice Animators

Authors: Richard Davis, University of California, Berkeley; James Landay, University of Washington, Intel Research Seattle. Full disclosure: I've co-authored with James Landay in the past.

Wouldn't it be cool to be able to set your ink and sketches on your tablet PC in motion? In theory it's possible to do this using ink with PowerPoint's animation features, but in practice it can really difficult to throw something together quickly. Well K-Sketch has some nifty ideas and techniques embodied in a fairly clean and easy-to-use sketching interface that makes it possible to do this super fast -three times faster than you can do similar things in PowerPoint. Now that sounds like fun.

This is one that you can soon play with yourself, according to k-sketch.org:

1/3/2008: We are very close to finishing our initial release of K-Sketch, and we expect it to be ready within a month.

K-Sketch will allow ordinary computer users to create informal animations from sketches. [...] Our design allows the most important types of motion to be defined with pen gestures, and gives visual feedback for coordination of events.

A video of of K-Sketch (WMV format) shows some of what it can do. The presentation is tailored for the academic audience, but the capabilities are very cool. Be sure to keep watching until the ski jumper double flip animation - try to imagine how tough it would be to do that in PowerPoint!

Indirect Mappings of Multi-Touch Input Using One and Two Hands

Tomer Moscovich, University of Toronto , Brown University; John Hughes, Brown University.

I don't have any inside scoop on exactly what this paper is about, but Tomer has done some very cool multi-touch interaction techniques in the past. He's also the person who came up with the "virtual scroll ring" technique that is the basis of the circle-the-pen-to-scroll gesture that has been a big hit with the tool ring in InkSeine. So I'm very keen to see what clever new things Tomer has devised in the multi-touch space.

Multi-Flick: An Evaluation of Flick-Based Scrolling Techniques for Pen Interfaces

Dzimity Aliakseyeu, Philips Research Eindhoven, Netherlands; Pourang Irani, University of Manitoba, Canada; Andres Lucero, Eindhoven University of Technology, Netherlands; Sriram Subramanian, University of Bristol, U.K

Scrolling with a pen, virtual scroll ring notwithstanding, is something that begs for cool gestures that let you easily express different granularities and speeds of how to flip through your documents. From the title I'm inferring that these researchers have come up with a few different kinds of flicking gestures that let you navigate through stuff quickly and easily. They don't have the paper posted on their web site yet, but it looks like they are implementing it using a pen on a tabletop computer (kind of like the Microsoft Surface). They have about a half-dozen other papers - one with a video - of other interesting work they've done on tabletop devices.

"It's on my other computer": Computing with Multiple Devices

David Dearman, University of Toronto; Jeff Pierce, IBM. Full disclosure: Jeff Pierce and I have published together before, and are both former students of Randy Pausch, who is waging a now-famous losing battle with pancreatic cancer. But he's hanging in there so far and still making great contributions to society.

All of you mobile enthusiasts out there have struggled with keeping your mobile devices and computers in sync, only to have some key bit of information elude you at exactly the moment you need it. Jeff Pierce had some cool projects going when he was a professor at Georgia Tech on "Personal Information Environments" where his vision was that you could walk around with your cell phone or notebook computer and use opportunistic annexing to dynamically associate your device with displays, touch-screens, keyboards, or other devices you encountered. My inference is that he's made some nice progress in this space and has come up with some techniques to maintain easy access to all your stuff across multiple devices.

PieCursor: Merging Pointing and Command Selection for Rapid In-Place Tool Switching

George Fitzmaurice, Justin Matejka, Azam Khan, Mike Glueck, Gordon Kurtenbach, Autodesk

This is the same crew of guys who came up with all the interesting user interface stuff in the Alias Sketchbook application, which I still use to this day for sketching on my Tablet PC. It looks like they've come up with a spiffy way to switch between different tools with your pen on your Tablet PC without having to move the pen to the edge of the screen and tap on some silly icon or bring up a menu. There's an intriguing thumbnail image of the technique on Azam's web site:

Thumbnail of PieCursor from Azam's web site - can you guess how it works?

My best guess is that it is some kind of a floating menu that follows your pen. Then you can activate a radial menu on it to pick tools (like a Zoom function), but you can continue the same stroke to specify how much to zoom. This is a trick known as "merging" in some other papers, hence my guess based on the title of their paper. This could be something really cool to let you pick different tools and commands much faster on your Tablet PC.

Those are the highlights from my short list of what caught my eye, but there's well over 100 papers to appear at the CHI conference and I'm sure there's many other gems in there. For example, there will be a whole session on image search interfaces, including CueFlik from Microsoft Research. I'm not involved in that work but I've seen that demo in the labs here and it's pretty cool. Image Search is one area where I think Microsoft Live Search has a clearly superior user interface to Google: for example, try a Live image search for Tablet PC. CueFlik is something that could make this even cooler and more useful.