I came across the phrase “Uncanny Valley” for the first time at South by Southwest last year. It came up in reference to the creepy not-realism of the then-recent Disney production “Mars Needs Moms.”

And to get up to speed on what that means, the phrase comes from robotics:

things that look somewhat human, but are clearly not — such as C-3PO (in Star Wars) or a Golem — produce an accepting reaction, while things that are very nearly human, but just a little strange — such as a child’s doll, aventriloquist’s dummy, or a clown — produce a negative response.

- TV Tropes & Idioms

So that’s uncanny valley. If you plotted this on an X and Y access, X for how human normal-like and Y for how cool you are with that, the valley refers to that acceptance drop– in other words, how creeped out you are by things that look like humans but are not.

In comics, we understand that the greater the detail of the image, the harder it is to relate to personally. Scott McCloud talks about that in Understanding Comics (still one of my favorites):

Uncanny valley specifically refers to visual feedback, but I think it goes further, and that it implies to a repulsion for technology that we designers can run the risk of touching on by building in too many personal details.

There’s something icky about systems that don’t appreciate that they have waaaay more than enough personal data about you and me at their disposal. Still, I sometimes get wrapped up in what it’d be cool to do, forgetting that I have an above-average understanding and expectation of technology and what it can know about us and what it can do with that information. So I run the risk of making a Chuck E. Cheese experience (if you had a birthday party as a child in North America during the 80s or 90s, you know what I mean. I swear; I still have nightmares) when what I was going for was Wall-E.

Recommend, don’t infer

When I first discovered Pandora and the Music Genome Project, I was super excited. I was so excited, in fact, that I was inspired to go back to school and learn about what interaction design was and how I could do it too. Recommender systems are great! I am still amazed by the magic of watching a smart system get smarter and get quickly to the heart of your taste preferences with more precision than an old friend.

There’s something lost in that, though. Recommender systems are pretty arrogant, and they’re wrong almost as often as they’re right. We’ve learned to ignore the static that’s an Amazon book or music recommendation for something we’ve already purchased somewhere else.

Respect for boundaries

Everybody has a different threshold for how much technology belongs in their lives. When we create without a healthy respect for those boundaries, we create uncanny interfaces. When I put on my researcher hat, I want to know more about the themes in those boundaries. For me, location based services frequently toe the line of freaky, even though I should be well used to them now. I imagine the boundaries vary and change by culture.

As promoters of technology, we should be aware of those boundaries, and respect them. Unfortunately, the ability to construct and deploy complex systems comes with a lot of power. Respect for boundaries is not frequently enough at the heart of business decisions.

When accessorizing…

It’s one of my mental go-to quotes that I picked up somewhere from an unlikely source (I’d prefer a Timbuk2 messenger bag to Chanel any day). At least, I believe the quote to be attributed to Coco Chanel, basically, that “when accessorizing always take off the last thing you put on.” We have to get to know the things that we like. What I’m thinking is that when I have the impulse to automate, I need to take a step back and think about the reciprocity of the system and the user. Probably one less detail than the one I’ve added is right. I think it’s good to establish familiarity, but in steps over time.

Ok, that’s it. I end with a random google image search for “Awesome Accessories” (this):

Trust and design. I’ve been chewing on that for the past few months. Since we started the contextual inquiry for our CHI project, actually. I’ve started to think about how there’s an element of trust to every well designed object, product or service.

Like these scissors.

I think about trust a lot.

CHI, Location & Technology for Mediating Trust

Last fall, as part of the CHI Student Design Competition, we started to ask questions about the necessity of meeting strangers in public spaces for the purpose of going for a walk. The line of questioning came from interesting, and often conflicted, anecdotes that walkers had about encountering strangers in public spaces.

To paraphrase a main point of our design (Night Beacon, an app that would empower people to walk with confidence at night), the app would lose its purpose without trust. It’s one reason that though we were approached to develop this app during and following CHI, we held off. I think we chose instead to use our project to chip at some underlying assumptions about the role that technology can play in building trustful relationships between humans, especially in cases where a smart phone with network connectivity can be the most readily available tool in a potentially dangerous or threatening situation.

This project challenge perceptions of danger and threats to safety. Like the rest of my teammates, I found myself thinking constantly about perceived versus real threats as the competition drew near. Who would want to design or build a system that would put someone in harm’s way? And we met people who passionately warned us that our idea would not work. We also met people who were willing to accept the possibility that it could.

As a CHI side note, I checked in on previous CHI SDC winner Noah Liebman’s website, and he has a great post on competing at CHI. His first piece of advice: Trust the Process. His teammate, Deb Lauterbach, also has some sage advice for competitors. These posts both incidentally fell in line with this post, but it might have something to do with the fact that the SDC, for competitors, becomes a big part of your life. And if done well (ie the way Noah and Deb describe), you will be intellectually challenged to the point where you may not really feel aptly prepared to reflect on it until things die down a little bit.

And one more piece of advice for the competition: know your project’s weaknesses better than anyone else. Poke your own holes so you can patch ‘em up (a messy metaphor, but you know what I mean).

Trust & Automobiles

I think a lot about the future of cars now. Trust is an essential element of getting in a giant hunk of metal that we propel, with the assistance of the machine itself, at speeds approaching 100 miles per hour. Since I work at GM now, I think about that every day. I think about the effective application of information systems to the driving experience. When you routinely trust that a car will work, that helps you get back in a car every time you get in a car. I bet people who have been in major accidents have much different relationships with their vehicles than those of us who haven’t [yet].

Data Packets: Trusting it Will Get There, What Should We Do Now?

And on another, somewhat more abstract level, I was just thinking about how I think that our trust level has reached a fair saturation point in the few years or so, at least in places of relative advantage. For a system predicated on sending a bunch of packets of data, on repeat if necessary, until they reach their source, we certainly have every expectation that the system will work quickly and seamlessly.

In his book, Emotional Design, Don Norman points out

Much of modern technology is really the technology of social interaction: it is the technology of trust and emotional bonds. But neither social interaction nor trust were designed into the technology or even thought through: they came about through happenstance, through the accidental byproducts of deployment (157).

Wired editor Chris Anderson proclaimed last month that The Web is Dead.  Whether or not this statement was meant to provoke and/or bait, Anderson discussed the shift from ‘the world wide web’ to API-driven applications, that is, applications away from a browser (like Netflix Streaming on your TV). The academic in me has me thinking about how deep-rooted our trust has gotten: enough that we are willing to get our data from device and application-specific filters and views instead of from a trusty old browser window.

The designer in me is picturing a game of catch. It’s a game of catch between you and your closest friends, and some of your friends who you aren’t that close with, but who you like to play catch with anyway. The passing back and forth of pieces of data between yourself and the people you trust, that is a big part of day-to-day life that I see coming. So what might that look like? What can we do with that?

Trust is something I’ve repeated 20 times over the course of this post because it speaks to a larger ethics of design that’s easy enough to overlook. But I think a basic principle is that good design doesn’t hide its intentions. So as far as CHI goes, I am so pleased that Noah says ‘trust the process’ around the time I’m thinking about the role of trust in design. So as far as your or my desire to be a creative professional, let’s keep thinking about this trust thing, okay?

photo courtesy of erin borreson

Question: what do interaction designers in the automotive industry do for fun? Answer: pile into a coupe and drive 8 hours until they reach Two Rivers, WI, home of the Hamilton Wood Type Museum. Or at least, that’s what I did last weekend.

The website Quite Strong, run by some of my workshop peers, describes the experience and has some fantastic photos. I regret that I didn’t have much of a chance to chat with the ladies of Quite Strong… I was a tag-along, but my attention was largely deeply focused on the experience setting and printing with wood type.

My professional world is largely lived in wireframes. And for this reason, the day-long workshop was an essential reminder of the materiality of spatial relationships. For wood typesetting (if you’ve never done it), you have to brace every section of your type, either with magnets, lead or wood. They call this stuff “furniture.” Jim Moran, the museum director observed my setup by saying: “Not that it’s necessarily a bad thing, but you’ve created an interesting problem for yourself there…”. And my first setup, which couldn’t have consisted of more than 30 letters (and a healthy sprinkling of ampersands), took over two hours to set up.

I tried doing a lot of things with layering that didn’t really work, and for my 2nd setup I decided to simplify. 8 hours wasn’t nearly enough to get a good sense for what the Vandercook press, which I was set up on, could do.

I can’t really say what happened to the time. All of a sudden I looked up and it was time to go. But I left feeling a new respect for design and how space informs perception, and I left bringing that back to work on Monday. Back to wireframes.

Update: Night Beacon is the winner of the 2010 CHI Student Design Competition.

Night Beacon is a project developed during my Interaction Design course as part of a team of five. As part of the CHI student design competition, our team was challenged to think about the problem of walking. Early in the process, we decided to narrow the problem of walking into that of walking at night. By narrowing the problem, the potential solution could tackle the reconstruction of notions about trust and safety in city spaces. We completed a contextual inquiry, created personas, conducted field research with paper prototypes, and came up with lo-fi and hi-fi prototypes for an Android application that included user testing for feedback.

The paper we wrote drew on this design experience, and was named as a finalist to be presented in the April at the CHI 2010 ACM conference (along with our peer group, mibo).

Re-imagining the site(s) comes at a time of change for our organization, which records and archives lecture materials, making new kinds of knowledge available to the public. The availability of these resources invites predictions and visions for how it may be used.

In initiating this kind of site redesign, A List Apart’s “The Content Strategist as a Digital Curator” is a good place to start, and it addresses the relationship between digital space and physical space that has become more and more important for content specialists to understand.

Without diverging too much, let me say that my field jump into HCI started as my fascination with a troubling metaphor back in my new media studies classes in undergrad. At the time, digital media theory introduced the idea that digital space is like physical space, and must be navigated in cognitively similar ways. I noticed a conflict in treating the relationship of digital space to physical space as a metaphor and a reality, and I wanted to investigate further.

I didn’t know then that Information Architecture was a burgeoning field, and it bears mentioning that I once saw the relationship as a metaphor, and now, as a practitioner, I see digital space as a real, albeit different kind of space. Much like the terms “calculator” and “computer” once referred to the clerical professions of people and used those models for building the machines we use today, digitization has created a vast information landscape that exceeds the human resource of attention. The limitation of attention makes the problem of getting through information a problem of navigation (among many other kinds of problems).

So in re-imagining the archive, I decided to visit the newly expanded University of Michigan Museum of Art, this time while thinking about its goal to present and preserve a collection within the limitations of a physical space.

In doing so, I realized that UMMA recently solved a problem of limited space for a too-large collection by doubling the size of the exhibition space with the Maxine and Stuart Frankel Wing.

Physical space directs attention. When inside the UMMA, you see doorways leading to rooms, and in those rooms you can actually see the exhibition’s contents. While a user landing on a website can judge in a very short time about if they will stay or where they will go next (well under a second), and will not likely scroll down to do so, in a physical space a visitor can make that kind of decision just by looking through a doorway. UMMA augments this by providing information on the walls, guiding the way from exhibit to exhibit.

It’s a totally common strategy, but the limitation of information (only the nearby wings are shown, and information is reduced relative to the user’s position in the museum) within the physical space helps people get through it. And these decisions have been made in part by my SI colleagues, so you can bet they were well considered.

Physical space can hint at what’s next door. One interesting update to the museum were the “open storage” galleries. In these galleries, related objects were placed together on shelves with much less context than the exhibited objects. The related objects were sorted by collection on different shelves. And the structure of the shelves was suggestive of how physical space enables perception– the shelves were lit and separated by door dividers with opaque, vertical beams. However, when leaning in close to see the objects, you can still see into the adjoining shelf space, perceiving both separation and relationship at the same time. I took a picture:

When designing, a digital space must provide relationship cues in places that are available to the user to understand. Finding meaningful relationships between stored objects– in my case, lecture objects– and presenting them as neighbors may make an archive more browseable. What I’m talking about is grounding what we call recommender systems in a navigation scheme that takes its cognitive cues from physical space.

I’m a brand new designer, and it’s therefore just the time to wax philosophical in order to create extraordinary future design. How do we imagine our ability, as humans, to see and understand our vast landscape of information? If we best extend our ability to do so, what will we see? What will we learn? The question precedes understanding the difference between seeing the information landscape from a horse and buggy and seeing it from an airplane. And yes, that’s a metaphor.

Game Boy Classic

The past few weeks I’ve been getting an education in affordances. This is not to be confused with having Norman’s Design of Everyday Things assigned, followed by a brief discussion, conclusion and let’s move on. Rather, we’ve been getting into the nitty-gritty details of why affordances matter, and how they work.

Two weeks ago, we were asked to make an affordance matrix and bring it into class. We took pictures of everyday objects that afford pushing, squeezing and twisting and arranged the pictures into a matrix.

I decided that a good place to start was with the very classic and tactilely pleasing Game Boy. Yes, because I am old school.

It’s hard to find objects whose interactions rely solely on pushing, squeezing or twisting. More often, we interact with objects on a procedural basis. For example, think of a wheelbarrow. We squeeze the handles, then pull up, then continue to squeeze in order to push. We could think about the interactions procedurally, or we could think about the sum combination of interactions in terms of accomplishing the object’s desired function. This is what got me to the triad– the desire to try to introduce a new scale to our perception of affordances.

There are some problems with the triad– the concentric circles imply continuity between interaction styles, and the styles themselves are nominal. In other words, the interactions don’t flow into each other. However, if we think about the overall object in terms of task, the triad works better.

Yesterday I met with two classmates in order to brainstorm an affordance device. That is, we have to plan and build a cube-shaped device that affords pushing, squeezing and pulling and that disregards outcome. Pictures are forthcoming, pending completion of the assignment.

Update: here’s some pictures of our “affordance cube.” We chose to work with soft and tactile materials. Sponges fit the bill. Other materials we used include paint, pink duct tape, a child’s tile, and …glue!

My teammates and I decided to focus on the problem of “bad” links– links you want to purge from history. During this process, I began thinking about the metaphor of clothing storage. To some extent, I hypothesized, link storage is like storing clothes in a closet– some you use every day, some you never pull out, some you want to get rid of, or never wanted in the first place.

This falls in line with the research that says frequency of visitation falls into 3 categories (described in This Presentation by U-M SOCHI, Slide 9):

• Fast (short term, high volume, sites like shopping)
• Medium (regular hourly checking, Gmail, Facebook)
• Slow (rare but regular visits, personal interest sites)

I did some sketching to see how useful the clothing storage metaphor might be use for browser history interactions. Here’s what I came up with:

Clothing storage sketch

Ultimately, I’d like to work with my group a little more to see if the metaphor is useful. We want to find a meaningful way to get “bad” or unwanted links out of the browsing history.

A major problem of this design is a lack of parameters around “Fast,” “Medium” and “Slow” interactions.

After an initial obligatory self-searching, I noted that part of what makes the installation so intriguing is the flow of crawled text as it builds into a final representation. The installation displays partial stories before pushing the information derived into a category.

Here’s Nelson Mandela:

Nelson Mandela Persona

And here’s Nelson Mandela being built by Personas (to get a good look, you might want to watch the video full-screen):

[vodpod id=ExternalVideo.865893&w=425&h=350&fv=i%3D8079]

Okay, so we have one interesting public figure who became the President of post-Apartheid South Africa and who arguably was a positive force for change in the world. So what about the baddest of the bad?

Charles Manson:

Charles Manson Persona

It’s reassuring to see that “Illegal” is the largest category for Manson, but “social” is among the second largest category, and prominent categories like Sports and Management are downright confusing. Furthermore, it is strange to see the name Manson afforded categories like “art” and “professional” at all.

[vodpod id=ExternalVideo.865895&w=425&h=350&fv=i%3D8086]

Okay, we’ve done good man and bad man now. I decided to try out my current favorite comic (and person to follow on Twitter), Russell Brand. Here he is:

Russell Brand Persona

Here his largest categories make sense; online, movies and social. But Sports and Military make less sense.

[vodpod id=ExternalVideo.865897&w=425&h=350&fv=i%3D8089]

And finally, I didn’t have any girls. Jezebel did a nice piece a little while back about prominent women in the Media (What Do People’s Online Personas Say About Them?), which offers more visualizations if you don’t want to DIY. But I decided to try out Gwen Stefani, general rad chick and lead singer of No Doubt since as long as I can remember.

Here’s Gwen:

Gwen Stefani Persona

[vodpod id=ExternalVideo.865898&w=425&h=350&fv=i%3D8096]

Conclusions

After watching the growth of these fine (well, 3/4ths of them) folks, as well as observing the construction of my own persona, I recognized that the project might show just as much, if not more, about the body of data from which the personas are derived. For example, all of these personas, none of whom are athletes, include substantial sports sections. Maybe this has to do with the way that language is constructed to organize information about sports. For example, many sites have sports pages or sections, and there are many opportunities for vocabulary that falls within the umbrella of sports.

I’m actually not trying to go into deep analysis mode, so I guess I’ll leave it at that. It’s hard not to draw some conclusions. But if you found this post in the least bit interesting, you’ll want to read up on Personas. The explanation of the installation notes that the project was “not supposed to be utilitarian. Instead, it was meant to expose a process that often happens behind the scenes.”

Witty Comedies Featuring a Strong Female Lead (a la Baby Mama and Juno)

Ballet & Modern Dance

Dysfunctional-Family TV Dramas (a la Dexter, Six Feet Under and I imagine Arrested Development)

However, when I browse on demand movies, I often begin watching as a form of browsing. If I don’t like a movie, I stop it and look for something else, usually within the first 5 minutes or so. However, my recommender system thinks I watched the entire movie, and provides recommendations based on movies I flat-out didn’t like. Usually I end up disliking them for reasons that can’t be described with text– for example, I don’t like the lighting, the acting is bad or it isn’t what I thought it would be based on the description. Furthermore, upon login, Netflix tells me:

“You watched: [title of the movie I didn't actually watch]”

This is both inaccurate and annoying. Netflix should keep track of whether or not people finish on demand movies, perhaps by finding a sweet spot at which most people continue or navigate away.

But that’s just my two cents. Netflix can, of course, do what they want.

You can go to my portfolio page about the project to read more about our project, The CVGA Archive Visualization, which used data from the GiantBomb website as the backbone for an interface design that allowed for the exploration of video games in new and meaningful ways.