Context Management

Note: a while back, Christian Crumlish & Erin Malone asked me to write a sidebar for a book they were working on … an ambitious tome of design patterns for social software. The book, (Designing Social Interfaces) was published last year, and it’s excellent. I’m proud to be part of it. Christian encouraged contributors to publish their portions online … I’m finally getting around to doing so.

In addition to what I’ve posted below, I’ll point out that there have been several infamous screw-ups with context management since I wrote this … including Google Buzz and Facebook’s Groups, Places and other services.

Also to add: I don’t think we need a new discipline for context management. To my mind, it’s just good information architecture.


There was a time when we could be fairly certain where we were at any given time. Just looking at one’s surroundings would let us know if we were in a public park or a quiet library, a dance hall or a funeral parlor. And our actions and conversations could easily adapt to these contexts: in a library, we’d know not to yell “heads up” and toss a football, and we’d know to avoid doing the hustle during someone’s eulogy.

But as more and more of our lives are lived via the web, and the contexts we inhabit are increasingly made of digits rather than atoms, our long-held assumptions about reality are dissolving under our typing-and-texting fingertips.

A pre-web example of this problem is something most people have experienced: accidentally emailing with “reply all” rather than “reply.”  Most email applications make it brutally easy to click Reply All by accident. In the physical world in which we evolved, the difference between a private conversation and a public one required more physical effort and provided more sensory clues. But in an email application, there’s almost no difference:  the buttons are usually identical and only a few pixels apart.

You’d think we would have learned something from our embarrassments with email, but newer applications aren’t much of an improvement. Twitter, for example, allows basically the same mistake if you use “@” instead of “d.” Not only that, but you have to put a space after the “d.”

Twitter users, by the time of this writing, are used to seeing at least a few of these errors made by their friends every week, usually followed by another tweet explaining that was a “mis-tweet” or cursing the d vs @ convention.

At least with those applications, it’s basically a binary choice for a single piece of data: one message goes either to one or multiple recipients: the contexts are straightforward, and relatively transparent. But on many popular social nework platforms, the problem becomes exponentially more complicated.

Because of its history, Facebook is an especially good example. Facebook started as a social web application with a built-in context: undergraduates at Harvard. Soon it expanded to other colleges and universities, but its contextual architecture continued to be based on school affiliation. The power of designing for a shared real-world context allowed Facebook’s structure to assume a lot about its users: they would have a lot in common, including their ages, their college culture, and circles of friends.

Facebook’s context provided a safe haven for college students to express themselves with their peers in all their immature, formative glory; for the first time a generation of late-teens unwittingly documented their transition to adulthood in a published format. But it was OK, because anybody on Facebook with them was “there” only because they were already “there” at their college, at that time.

But then, in 2006 when Facebook opened its virtual doors to anyone 13 or over with an email address, everything changed.  Graduates who were now starting their careers found their middle-aged coworkers asking to be friends on Facebook. I recall some of my younger office friends reeling at the thought that their cube-mates and managers might see their photos or read their embarrassing teenage rants “out of context.”

The Facebook example serves a discussion of context well because it’s probably the largest virtual place to have ever so suddenly unhinged itself from its physical place. Its inhabitants, who could previously afford an assumed mental model of “this web place corresponds to the physical place where I spent my college years,” found themselves in a radically different place. A contextual shift that would have required massive physical effort in the physical world was accomplished with a few lines of code and the flip of a switch.

Not that there wasn’t warning. The folks who run Facebook had announced the change was coming. So why weren’t more people ready? In part because such a reality shift doesn’t have much precedent; few people were used to thinking about the implications of such a change. But also because the platform didn’t provide any tools for managing the context conversion.

This lack of tools for managing multiple contexts is behind some of the biggest complaints about Facebook and social network platforms (such as MySpace and LinkedIn). For Facebook, long-time residents realized they would like to still keep up their immature and embarrassing memories from college to share just with their college friends, just like before — they wanted to preserve that context in its own space. But Facebook provided no capabilities for segmenting the experience. It was all or nothing, for every “friend” you added. And then, when Facebook launched its News feed — showing all your activities to your friends, and those of your friends to you — users rebelled in part because they hadn’t been given adequate tools for managing the contexts where their information might appear. This is to say nothing of the disastrous launch of Facebook’s “Beacon” service, where all users were opted in by default to share information about their purchases on other affiliated sites.

On MySpace, the early bugbear was the threat of predator activity and the lack of privacy. Again, the platform was built with the assumption that users were fine with collapsing their contexts into one space, where everything was viewable by every “friend” added. And on LinkedIn, users have often complained the platform doesn’t allow them to keep legitimate peer connections separate from others such as recruiters.

Not all platforms have made these mistakes. The Flickr photo site has long distinguished between Family and Friends, Private and Public. LiveJournal, a pioneering social platform, has provided robust permissions controls to its users for years, allowing creation of many different user-and-group combinations.

However, there’s still an important missing feature, one which should be considered for all social platforms even as they add new context-creation abilities. It’s either impossible or difficult for users to review their profiles and posts from others’ point of view.

Giving users the ability to create new contexts is a great step, but they also need the ability to easily simulate each user-category’s experience of their space. If a user creates a “co-workers” group and tries to carefully expose only their professional information, there’s no straightforward way to view their own space using that filter. With the Reply All problem described earlier, we at least get a chance to proof-read our message before hitting the button. But most social platforms don’t even give us that ability.

This function — perhaps call it “View as Different User Type” — is just one example of a whole class of design patterns we still need for managing the mind-bending complexity we’ve created for ourselves on the web. There are certainly others waiting to be explored. For example, what if we had more than just one way to say “no thank you” to an invitation or request, depending on type of person requesting? Or a way to send a friendly explanatory note with your refusal, thereby adding context to an otherwise cold interaction? Or what about the option to simply turn off whole portions of site functionality for some groups and not others? Maybe I’d love to get zombie-throwing-game invitations from my relatives, but not from people I haven’t seen since middle school?

In the rush to allow everyone to do everything online, designers often forget that some of the limitations of physical life are actually helpful, comforting, and even necessary. We’re a social species, but we’re also a nesting species, given to having our little nook in the tribal cave. Maybe we should take a step back and think of these patterns not unlike their originator, Mr Alexander, did — how have people lived and interacted successfully over many generations? What can we learn from the best of those structures, even in the structureless clouds of cyberspace? Ideally, the result would be the best of both worlds: architectures that fit our ingrained assumptions about the world, while giving us the magical ability to link across divides that were impossible to cross before.

You Are (Mostly) Here: Digital Space & the Context Problem

Here’s the presentation I did for A Summit 2009 in Memphis, TN. It’s an update of what I did for IDEA 2008; it’s not hugely different, but I think it pulls the ideas together a little better. The PDF is downloadable from SlideShare. The notes are legible only at full-screen or on the PDF.

Oh, my binary heart

David Weinberger’s most recent JOHO post shows us some thinking he’s doing about the history (and nature) of “information” as a concept.

The whole thing is great reading, so go and read it.

Some of it explores a point that I touched on in my presentation for IDEA earlier this month: that computers are very literal machines that take the organic, nuanced ambiguities of our lived experience and (by necessity) chop it up into binary “is or is not” data.

Bits have this symbolic quality because, while the universe is made of differences, those differences are not abstract. They are differences in a taste, or a smell, or an extent, or a color, or some property that only registers on a billion dollar piece of equipment. The world’s differences are exactly not abstract: Green, not red. Five kilograms, not ten. There are no differences that are only differences.

The example I gave at IDEA was how on Facebook, you have about six choices to describe the current romantic relationship you’re in: something that normally is described to others through contextual cues (a ring on your finger, the tone of voice and phrasing you use when mentioning the significant other in conversation, how you treat other people of your sig-other’s gender, etc). These cues give us incredibly rich textures for understanding the contours of another person’s romantic life; but Facebook (again, out of necessity) has to limit your choices to a handful of terms in a drop-down menu — terms that the system renders as mutually exclusive, by the fact that you can only select one.

More and more of the substance of our lives is being housed, communicated & experienced (by ourselves and others) in the Network. And the Network is made of computers that render everything into binary choices. Granted, we’re making things more fine-grained in many systems, and giving people a chance to add more context, but that can only go so far.

Weinberger uses photography as an example:

We turn a visual scene into bits in our camera because we care about the visual differences at that moment, for some human motive. We bit-ify the scene by attending to one set of differences — visible differences — because of some personal motivation. The bits that we capture depend entirely on what level of precision we care about, which we can adjust on a camera by setting the resolution. To do the bit-ifying abstraction, we need analog equipment that stores the bits in a particular and very real medium. Bits are a construction, an abstraction, a tool, in a way that, say, atoms are not. They exist because they stand for something that is not made of bits.

All this speaks to the implications of Simulation, something I’m obsessing about lately as it relates especially to Context. (And which I won’t go into here… not another tangent!)

Dave’s example reminds me of something I remember Neil Young complaining about years ago (in Guitar Player magazine) in terms of what we lose when we put music into a digital medium. He likened it to looking out a screen door at the richly contoured world outside — but each tiny square in screen turn what is seen through its confines into an estimated average “pixel” of visible information. In all that averaging, something vital is inevitably lost. (I couldn’t find the magazine interview, but I did find him saying something similar in the New York Times in 1997: “When you do an analog recording, and you take it to digital, you lose everything. You turn a universe of sounds into an average. The music becomes more abrupt and more agitating, and all of the subtleties are gone.”)

Of course, since that interview (probably 15 years ago) digital music has become much more advanced — reconstructing incredibly dense, high-resolution information about an analog original. Is that the answer, for the same thing that’s happening to our analog lives as they’re gradually soaked up by the great digital Network sponge? Higher and higher resolution until it’s almost real? Maybe. But in every case where we’re supposed to decide on an input to that system (such as which label describes our relationship), we’re being asked to turn something ineffable into language — not only our own, expressively ambiguous language, but the predefined language of a binary system.

Given that many of our lives are increasingly experienced and mediated via the digital layer, the question arises: to what degree will it change the way we think about identity, humanity, even love?

Context Collapse

First of all, I didn’t realize that Michael Wesch had a blog. Now that I’ve found it, I have a lot of back-reading to do.

But here’s a recent post on the subject of Context, as it relates to web-cams and YouTube-like expression. Digital Ethnography — Context Collapse

The problem is not lack of context. It is context collapse: an infinite number of contexts collapsing upon one another into that single moment of recording. The images, actions, and words captured by the lens at any moment can be transported to anywhere on the planet and preserved (the performer must assume) for all time. The little glass lens becomes the gateway to a blackhole sucking all of time and space – virtually all possible contexts – in upon itself.

By the way, I’m working on a talk on context for IDEA Conference. Are you registered yet?

Public vs Published

When I first heard about the Kozinski story (some mature content in the story), it was on NPR’s All Things Considered. The interviewer spoke with the LA Times reporter, who went on about how the judge had “published” offensive material on a “public website.”

I won’t go into detail on the story itself. But I urge anyone to take the LA Times article with a grain or two of salt. Evidently, the thing got started when someone who had an ax to grind with the judge sent links and info to the media, and said media went on to make it all look as horrible as possible. However, the more we learn about the details in the case, the more it sounds like the LA Times is twisting the truth a great deal. **

To me, though, the content issue isn’t as interesting (or challenging) as the “public website” idea.

Basically, this was a web server with an IP and URL on the Internet that was intended for family to share files on, and whatever else (possibly email server too? I don’t know). It’s the sort of thing that many thousands of people run — I lease one of my own that hosts this blog. But the difference is that Kozinski (or, evidently, his grown son) set it up to be private for just their use. Or at least he thought he had — he didn’t count on a disgruntled individual looking beyond the “index” page (that clearly signaled it as a private site) and discovering other directories where images and what-not were listed.

Lawrence Lessig has a great post here: The Kozinski mess (Lessig Blog). He makes the case that this wasn’t a ‘public’ site at all, since it wasn’t intended to be public. You could only see this content if you typed various additional directories onto the base URL. Lessig likens it to having a faulty lock on your front door, and someone snooping in your private stuff and then telling about it. (Saying it was an improperly installed lock would be more accurate, IMHO.)

The comments on the page go on and on — much debate about the content and the context, private and public and what those things mean in this situation.

One point I don’t see being made (possibly because I didn’t read it all) is that there’s now a difference between “public” and “published.”

It used to be that anything extremely public — that is, able to be seen by more than just a handful of people — could only be there if it was published that way on purpose. It was impossible for more than just the people in physical proximity to hear you, see you or look at your stuff unless you put a lot of time and money into making it that way: publishing a book, setting up a radio or TV station and broadcasting, or (on the low end) using something like a CB radio to purposely send out a public signal (and even then, laws limited the power and reach of such a device).

But the Internet has obliterated that assumption. Now, we can do all kinds of things that are intended for a private context that unwittingly end up more public than we intended. By now almost everyone online has sent an email to more people than they meant to, or accidentally sent a private note to everyone on Twitter. Or perhaps you’ve published a blog article that you only thought a few regular readers would see, but find out that others have read it who were offended because they didn’t get the context?

We need to distinguish between “public” and “published.” We may even need to distinguish between various shades of “published” — the same way we legally distinguish between shades of personal injury — by determining intent.

There’s an informative thread over at Groklaw as well.

**About the supposedly pornographic content, I’ll only say that it sounds like there was no “pornography” as typically understood on the judge’s server, but only content that had accumulated from the many “bad-taste jokes” that get passed around the net all the time. That is, nothing more offensive than you’d see on an episode of Jackass or South Park. Whether or not that sort of thing is your cup of tea, and whether or not you think it is harmfully degrading to any segment of society, is certainly your right. Some of the items described are things that I roll my eyes at as silly, vulgar humor, and then forget about. But describing a video (which is currently on YouTube) where an amorously confused donkey tries mount a guy who was (inadvisedly) trying to relieve himself in a field as “bestiality” is pretty absurd. Monty Python it ain’t; but Caligula it ain’t either.