Ubicomp
I’ve been meaning to write up my experiences with a couple of mobile phones for a while. I used a Nokia 6682 for a year or so and recently switched over to a Sony k790a.
My experience with the k790 is a great lesson in why product managers and designers continue to fight about whether features or experience should drive product development. I bought the phone primarily because I wanted a phone with a “real” camera and at the time it was (arguably) the best cameraphone out there (3.2 megapixels, auto-focus, xenon flash, etc.). I did lots of research and so on and bought it primarily on a features basis. After using the phone and camera for a few weeks, however, I realized that while the k790 has an amazing camera for a phone, especially for stills, it’s going to be a while before any cameraphone is going to hold a candle to a decent dedicated point-and-shoot camera (e.g. Canon Elph) let alone a digital SLR. So much for features.
So, you’d think I would have been disappointed but in fact I had already grown to love the phone based solely on the software interaction design. It’s among the most impressive pieces of small-device UI design I’ve encountered (and that includes the iPod which is obviously amazing, but has a much, much easier job description). It’s far from perfect, of course, and there are some patently silly things as with all phones (for example, when looking up contacts, you can only search by first name rather than first or last, so typing “JO” gives me “John Zapolski” but not “Oliver Johns”). But, there are so many places where the designers got it right. And, what I mean by getting it right is that they understood the context the phone would be used in, figured out the most likely thing or things that you’d want to do, and then designed the interface to make those things either extremely easy or automatic. I regularly find myself pleasantly surprised by the interface which is rare indeed, as we all know from countless bad experiences. And, as an interaction designer myself, I can tell you that it is incredibly difficult.
It’s the little things like when scrolling vertically through a list of your contacts you can then scroll horizontally through their different contact methods (home phone, work phone, email) and choose the one you want. It (mostly) just feels incredibly fluid.
By contrast, the Nokia suffers from having to support a lumbering, generalized operating system so that I have the option of downloading lunar calendars, golf handicapping systems, and all kinds of other pseudo-useful Symbian applications but at the expense of doing the things that matter with any degree of subtlety.
(Note: This was all written pre-iPhone, which generally out-fluids the Sony.)
A while back I was thinking about the overlapping layers of context within which mobile interactions take place (more on the word “mobile” in another post). I quickly sketched out a visual model to help me (and other designers working on situated experiences) keep these various layers in mind. Here’s the resulting paper (200k pdf), which my collaborator, Jared, will be presenting later this year at Mobile HCI in Singapore.
And, here’s the abstract:
Designers of mobile applications have long understood that mobile devices are operated within a context of significant constraints and environmental distractions. Despite this knowledge, however, many mobile applications are designed as if they were merely shrunken desktop or Web applications.
To encourage consideration of the specifics of context for mobile interactions and to highlight new user-meaningful opportunities latent in always-on, always-carried devices, this article describes a context model for mobile interaction and a set of design heuristics for successful mobile interactions.
Feedback welcomed!
Jared and Yue have written a nice overview of our ongoing Mobile China research for UIGarden. From their conclusion:
In contrast to the expectations of many of our technology clients, we have found in our research that youth discourse about telephones focuses much more heavily on emotions rather than technology and features. The implications for designers of mobile devices and services is to focus more on the human side of this emerging technology, new youth identities, and popular desire for entertainment, fashion and companionship.
It occurred to me the other day that Scott McCloud’s point about the web’s infinite canvas applies (and is vastly more important) to annotating physical space. Any given physical location needs both multiple categories of information (e.g. basic information, reviews, history, etc.) and also infinite instances of each. And it’s going to get real messy real fast, given both the variety of data (where’s the nearest bathroom vs. Mark Twain slept here vs. I love this place vs. danger: asbestos!) and the imprecision inherent in this kind of layering (did Mark Twain sleep here or over here or maybe right there?). I don’t see any way to do that other than to open it up and let the world have at it.
I’ve been bothered by the raft of mobile informational services that don’t take into account the real context of use. When would I ever want to find something as generic as “pizza” in a given location? OK, maybe ice cream, but still while walking around a city I either want something really specific (e.g. where’s the nearest Blue Bottle coffee) or something more general but of an editorially (or socially) vetted quality (e.g. show me a cool place to read for an hour near here). So, I have to give a shout out to MizPee, which is an old idea (and probably not implemented right), but at least it’s something immediately useful enough to get me to bother with the mobile web.
We’ve been up in Mendocino this past week at a cabin overlooking cliffs and white surf. Beautiful and rustic. And part of the beautiful, rustic cabin is a charming wood-burning stove in which a roaring fire was built. The trouble, of course, with charming wood-burning stoves in which one has built a roaring fire is that they get very hot and, especially when you (or, say, your darling son) are 2 years old or so, it is easy to forget this fact and, fascinated by the pretty, pretty flames, put your tiny hand on the corner of the stove. What ensues – screaming, parents running to the fridge for ice, more screaming, hugs, sobbing, etc. – is generally presented to the victim as a handy lesson (“now you know” etc. etc.), but the trouble is that next time he’s really just about as likely to put his hand back on the stove because (and this is the point) the stove doesn’t look hot. Hot is red (like fire). Hot is glowing (like embers) or waving (like air over a highway). But the stove is just dull and black. Oh yeah, and by the way, it’s 400-freaking degrees!
So, what’s my point. Well, this is a case where it makes sense to me to add in some information technology. To augment reality. To ubiquitously compute. How straightforward it would be if the stove (or, say, that scalding pot handle you were about to grab because, well, handles afford grabbing) was glowing bright red.
Which leads to my First Law of Information Technology for Objects: it makes sense to augment objects with information technology when doing so enables the objects to communicate useful information about themselves which would otherwise be difficult or harmful to detect. (There are other reasons, of course, such as to enable objects to make use of information about other objects, etc. but I only felt like making up one law).
Of course, there are issues that arise when you start augmenting things. You have to think about what you might be covering up with your informational “overlay.” For example, there was another rustic cabin I spent a night in, this one up in Humboldt, where we built a roaring fire in a wood-burning stove. No one touched this stove, despite the presence of certain vegetable inebriants, partly because we let the fire get so hot that the stove did start glowing red (all by itself). And, if some technological augmentation had turned it red at, say 250 degrees, maybe we wouldn’t have noticed it creep up to 750 (or wherever iron starts to melt) and it would have burnt the cabin down). This is probably a solvable design problem (multiple shades from yellow to white as it gets hotter, for example), but the point remains: careful what you subtract when you start adding.
Postscript: There’s a companion Law of Minimal Ugliness in Ubicomp, something to the effect that we need to consider the aesthetics of all information manifestation to avoid augmenting ourselves into ugliness (otherwise known as the Phoenix Principle).
So, the “first feature-length film shot on a mobile phone” looks about how you’d imagine a feature film shot on a mobile to look. But it doesn’t do so gratuitously (well, ok, no way am I going to watch the whole thing, but you quickly get the aesthetic point: voyeuristic, degraded, subjective, paranoid). I always like these moments of low-res injected into the endless march to higher definition, when the first people use the limitations of a transitional medium for art. (thanks, Tina)
WineM is a nicely-put-together “technology sketch” (basically a video of an idea for a product) from Mike Kuniavsky’s new company, ThingM. WineM is a winerack which basically uses RFID to associate bottles with UPCs to layer information on the collection of bottles it contains and then allows you to “browse” your collection (with admirable immediacy) along the standard wine facets (varietal, vintage, region, etc.). I link to it not because Mike’s a friend, but because people are always asking me “what’s all this about ubicomp… is that like surfing the web from your fridge?” and WineM is a perfect example. It’s cool, but it’s also way useful, technology enhancing the experience rather than technology becoming the experience. Well done, Mike and Tod.
I am fairly dismayed at the lengths mobile network operators will go to cajole, trick, or force consumers to use higher paying services, all in fealty to the cruel god ARPU. For example, Cingular modified the OS on its Nokia 6682 so that after you take a picture you no longer get a menu option to send via email as a way of trying to get people to send more (high-revenue) MMS messages. I suspect the result is that most people just curse under their breath as they go through the 10 steps it now takes to send via email or else they install a 3rd-party app such as the excellent Shozu. Either way, Cingular gets nothing.
And there’s now some data to back up the assumption that the “walled garden” is a misguided strategy: MobileCrunch recently pointed to an Opera report showing that “in direct contradiction of operator’s fears, the use of the Opera browser - which circumvents the “walled garden” that many operators attempt to maintain as a means of further monetizing data customers - is clearly increasing the ARPU that operators receive based upon customer data usage.” Hah!
Remember how you stopped using italics because they looked so bad on web pages, but then when the operating systems started anti-aliasing you started using them again? Well, think again, my friend. They’re worse than ever on the small screen:
A team at Nokia is working on putting a web server (specifically Apache) into mobile phones. My first thought is that they ought to work on getting the client side working before they move on to the server and there are lots of obvious security and performance issues, but it’s mind-bending nonetheless and certainly ups the world’s “ubiquity quotient” in a hurry.
Exhibit #324 presented as evidence that we are completely doomed. Just to put an image to Mike’s thoughtful observation.
Semapedia makes it possible to attach Wikipedia entries to physical objects. Very cool. However, it really underscores the need for GPS+Phone (or some other access to phone location info) to break out. So very silly to be pasting up flyers when the phone knows where it is. Still, this works now with real phones. (via)
Adam has penned (keyed?) a well-written, thoughtful discussion of the importance of ethical, user-centered design for ubiquitous computing.
You’ve probably seen this already: In-Room Chat as a Social Tool. It’s a thoughtful write-up of the social consequences of an experiment in parallel verbal/chat meeting. (thanks, tomalak)
Cranky Nadav is wondering: why do I still have to re-enter all my phone numbers into a new cell phone? I find it completely absurd that new phones don’t come with a USB cable and some basic ability to sync with a PC. Yeah, I know Bluetooth, blah, blah, blah, but really, a simple cable would do just fine. If we must make it whizzy and wireless, why not have the wireless provider offer a service to backup and restore your numbers right to/from your phone? I bet folks would even pay for it. On a related(ly cranky) note, my most recent phone is shaped so that it is almost impossible to cradle. Isn’t that a pretty basic usage scenario that you’d expect a reasonably competent industrial designer to consider?
Too cool: warchalking. Now if we’re really emulating hobo signs, we’ll need to start including passwords for closed networks as well. And there probably needs to be an AvantGo channel or some other easily-synced PDA file (or maybe Vindigo should just include access points along with it’s restaurant listings).
CHI-Web: Summary: Designing for VoiceXML (and telephone interfaces in general)
On pointandgrunt.net, humans are talking about interfaces for small devices. 10 Big Fingers is an informal overview of input devices and their appropriateness for PDAs and the like. And there’s a link to The Event Horizon User Interface Model for Small Devices proposes a new interface for arranging objects on small screens (radial expansion/contraction), but also includes an overview of WIMP alternatives (e.g. zooming, focus+context) and some discusssion of screen space issues.
.. and for more wireless stuff, check out David Anderson’s new dashdot log.
If you’re still interested in wireless web stuff (that’s so Summer 2000), this article gives a really good overview and comparison of the US, Europe, and Japan.
I was just reminded by xblog’s link to the i3 of the last time I visited Italy (gotta pay my respects to my “peeps”): I was lucky enough to get a demo of this rad HIPS project they were doing at the University of Siena (pdf). The idea was that there’s an information space that’s layered on top of physical space, not in the sort of McLuhanesque infosphere way, but where every physical location has layers and layers of information specific to it (a very Italian concept!).
But, even better than the idea is the product, a PDA tourguide that tailored its information to where you were, what you were looking at, and—most impressively—what you had been doing in the past. So, if it recognized that you had just returned to a painting you had already seen, it would give you deeper information, and might remember that you had been asking for more historical data about other paintings and so would float that to the top. It all sounds neat, but it’s nothing compared to actually using the thing (even in the buggy, prototypical version I got to see).
This is what pervasive computing is all about.
ps—the HIPS site is interesting, but pretty unstable/shocked.
pps—there’s also a Wired article about the project.
Why WAP matters, a good overview of the problems and potential of wireless web implementations, including a comparison of WAP and iMode. (thanks, alex)
IBM’s pervasive computing glossary.. I like the term “pervasive computing” (which I first heard in the title of John Thackara’s presentation). I like it because it describes the experience rather than the technology or—feh—the protocols being used. (via noise)
More WAPlash: “The WAP Trap”.. OK, enough with the silly names.
I’ve been talking a bit with Alex about mobile applications. One of the good points he made is that voice is in some ways a much more natural interface for many mobile apps (the most obvious example is driving directions). Here’s a related link to the W3C VoiceXML spec.
Oh, and I forgot to add this: Two Approaches to Bringing Internet Services to WAP Devices which vincent linked to a while back.
So, it turns out using the “web” on cell phones just totally sucks. Jakob makes some good points. The key here is that it’s a big hardware problem: WAP/WML was designed for browsing on leeetle tiny 4-line screens, so it doesn’t have any layout controls, which sounds reasonable on the surface. But, shit, you can’t do much of anything on a screen that small (I’m picturing like a 30-screen shopping cart right now). In other words, WAP/WML is an appropriate solution to absolutely the wrong problem.
There are some WAP resources at w3schools.com (thanks owen). Have I yet mentioned that WML is a big pain in the butt? Well, it is. So there!
Looky here, it’s an online WAP tester. In general, anywhereyougo.com seems to be a pretty good resource for mobile development.
Thanks, Vincent, for the link to the First Workshop on Human Computer Interaction with Mobile Devices. I also just stumbled on the Second Workshop on Human Computer Interaction with Mobile Devices (in postscript, for pete’s sake).
Speaking of compatibility problems, I’ve been starting to look at WML and think about wireless apps. I really like the idea of context-specific design which wireless enables, but—my god!—do you know how many web-enabled devices there are out there already (all with slightly different hardware and software interfaces)? If you thought simultaneously designing for both 640x480 and 1024x480 screens was hard, think about designing for both 20x30 and 200x300 (or whatever). Owww, my head hurts.
Cam points to a fairly in-depth critique of WAP.
Taylor on non-business uses of the wireless future.. Right on! .. the inability/unwillingness to think beyond commerce and business uses drives me crazy. That’s why I refuse to adopt the “customer experience” meme.
I would like to have been at this talk on The design challenge of pervasive computing:
Perhaps we might think about the design of pervasive computing as a new kind of street theatre.
And now they’ve put up a full archive of Doors of Perception 7: FLow