The term data exhaust was one of several additions to my vocabulary throughout FOO Camp 2008 (the annual “Friends Of O’Reilly” campout and unconference at the O'Reilly campus in Sebastopol, CA, to which I was pleasantly surprised to have been invited [again!]). The rather clever phrase came up in the context of a discussion about metadata (i.e., data that describes other data, e.g., titles on photos posted on Flickr) and refers to the behavioral residue we leave behind through our online activities … and I’m riffing on it in the title of this post - my expression of impressions from the weekend - as I can’t think of any [other] short, pithy description that captures the overall experience … and this post will [thus] be a kind of venting – perhaps even a catalytic conversion - of some of the thoughts and feelings that came up for me during the weekend.
The metadata session was rather far-ranging … almost too much so, at times. There were some interesting observations made by a number of participants – who I will not name, in keeping with Foo Camp’s policy of encouraging open discussion by discouraging subsequent revelation and/or attribution of potentially sensitive statements by participants. Rather than delve into details, I’ll make a few meta-observations (pun intended).
One meta-observation is that the discussion reminded me of some of the “neats vs. scruffies” debates in AI. The “neats” were (are?) very intent on logical formalisms that could [theoretically] accommodate a broad range of phenomena; the “scruffies” were/are hackers that would try anything – and any combination of things – to get a system to behave intelligently. I was squarely in the “scruffy” camp (and continue to be rather scruffy in my approach to other application domains), and was surprised that there seemed to be so many meta-data “neats” [speaking up] in the session, as my impression of Foo Camp is that it is largely composed of people in the hacker culture. Several people were proposing various frameworks to try to encompass a broad range of potential uses – and abuses – of metadata … and I found myself thinking that we don’t know any more about metadata than we do about intelligence, and so I believe scruffy approaches are more likely to succeed, at least in the near term. Another, related, meta-observeration is that I suspect metadata and intelligence - artificial or otherwise - are closely related on several dimensions
Yet another meta-observation is that it seems like many of the people engaged in the discussion are “privacy fundamentalists” (or acting primarily on their behalf) – eager to ensure that no evil can possibly ensue from the aggregation and use of metadata ... or, at least, not without our partially informed consent. I have at least two fundamental problems with this perspective. One is that I don’t believe most people will be willing to take the time to educate themselves and adjust whatever metadata collection settings are offered through any kind of user interface (not that I think an intuitive user interface for metadata would be easy to engineer). For example, how many people ever adjust the cookie settings for their browser? And using a browser is more intuitive - or, at least, more familiar - than thinking about and dealing more directly with metadata.
Even more bothersome for me is that for all concerns about making the web safe for demography (sociography?), there appear to be relatively few examples of the abuse of metadata. I asked for examples of such abuse during the session; other participants mentioned the Chinese blogger whose identity was revealed to authorities by Yahoo, and a company called USAinfo that purportedly sells/sold names and addresses of gullible senior citizens (though I cannot track down any info about that). I’m sure there are other incidents of abuse, but I imagine they are on par with the wildly overestimated – and overreported – instances of sexual predation on MySpace. But, as I’ve noted before, I tend to be in the “privacy unconcerned” category.
The next session I attended was about online “griefing”, focusing primarily on a former prominent blogger – or, more precisely, a prominent former blogger – who was the victim of a vindictive, sexually oriented and life-threatening vendetta that started in comments on one of her blog posts and soon moved into an apparently accountability-free site where anonymous blog authors were authorized and encouraged to attack her.
For all my talk about “privacy unconcernedness” and my railing against the overblowing of risks on MySpace, sexual violence against women is a hot button for me. The quickest way I can self-induce sharp anger is to simply think about a date rape incident a woman close to me once told me about (my heart is pounding as I write this now). I was really angry when I first read about the escalating series of events that engulfed this former blogger, even though (fortunately) no physical harm befell her.
Anyhow, it was good (for me) to hear this being discussed openly, even though there were some [men] in the session who were not sympathetic. One of the men who has also received death threats on his blog noted how Hollywood celebrities have the money to pay to insulate themselves from those who might physically harm them; online celebrities often do not have the same financial means to do so, and so are more vulnerable.
Another man likened the Internet to a giant schoolyard, complete with bullies. I started wondering about issues revolving around meanness in the online and offline worlds, and the related issue of scale. Is the Internet becoming a meaner place? Is the real world becoming meaner? I found myself thinking about Crossfire (which is off the air), The O’Reilly Factor (which is not), Rush Limbaugh, and other shows and celebrities in the mass media who tend to promote griefing and hate (though perhaps television is not a good proxy for “real life” [either]). Someone commented that negative posts get far more traffic than positive posts, and it’s pretty clear that negativity gets more attention in the traditional news media and in political campaigns ... and (not surprisingly) in our brains.
Focusing primarily on the online space, I wonder what relative levels of meanness are found in different social networks, e.g., MySpace,. Facebook, LinedIn, YouTube, Flickr, LiveJournal, etc. Do walled gardens – vs. the relatively more open blogosphere - help reduce the meanness levels? In a later Foo Camp session, someone noted that comments by YouTube users who have actually produced their own videos are far more supportive than the comments by those who merely comment on others’ videos. But in this case, we have blog authors (vs. commenters) who are writing blog posts that are incredibly vehement … although these are authors of blogs whose sole purpose appears to have been to attack another blogger … so maybe it is true that, in general, those who are producing their own content are less mean than those who are only criticizing others’ productions.
Another example of extreme meanness took place in March when hackers embedded flashing animations in an epilepsy forum that were designed to trigger a seizure in someone suffering from photosensitive epilepsy. Although no one died, several people suffered seizures visiting the site. What would possibly motivate anyone to do such a thing? I was wondering about potential equivalents in the real world; nothing came to mind then, but a later Lightning Round Talk at Foo Camp included a “game” called “Toast” where a pickup truck pulls up to an unsuspecting pedestrian around 3am, people jump up from the truck bed, yell “Toast!!” and start throwing toast at the person. I didn’t think this sounded particularly fun – or funny – and I'm not even sure if the speaker was just making up the game, but I was wondering what would happen if the victim was someone who suffered from epilepsy … and if it’s less of a stretch for people who do think this kind of thing is fun[ny] to imagine hacking an epilepsy site would also be fun[ny] … and in reflecting now on the laughter in the audience over the “game” of “Toast!”, I wonder how many Foo campers might have found the story of the epilepsy site hack funny if it had been presented in the Lightning Round by the speaker presenting “Toast!” … but maybe I’m taking this all too seriously.
[Update: I just found Joi Ito's fabulous photos from Foo Camp, and am embedding a photo he took of me that offers some visual evidence that I do sometimes take things rather seriously.]
I also found myself wondering if there is a difference in meanness levels on the Internet vs. real life, what accounts for that difference? Is it the Bystander Effect, wherein the larger the set of potential witnesses to a crime, the less likely any single one of them is to prevent or report it (the prototypical case being the 1964 murder of Kitty Genovese in NYC). Are the vast numbers of lurkers on the web creating orders of magnitude more bystanders than exists in real life? Are any measures taken in real life to reduce meanness levels, e.g., in high schools (and other schoolyards) applicable to the Internet? Someone in the earlier metadata session had noted the importance of preserving pseudonymity for bloggers and other people producing content that may be offensive to the repressive regimes they live under. Accountability measures that may reduce the threat to individuals in one culture or segment of society may increase the threat to individuals in others.
After lunch, I attended a session on ubiquitous computing. The session leader started off making some interesting distinctions between "pragmatic" and "academic", and between "USA" and "Eurasia" (the latter of which I think incorrectly combines Europe and Asia, as I see developments in the regions as at least as different as, say, between the USA and Europe). One of the more interesting observations was that once we understand it, it ceases to be ubiquitous computing, which parallels an oft-articulated view of artificial intelligence. Another was that ubiquitous computing makes physical space mutable or programmable - which got me thinking about interesting new terms, substituting "world" for "web", such as the programmable world or perhaps the read/write world. Personally, I think the increasing interconnections between the web and people, places and things in the real world will be the hallmark of Web 3.0.
One of the interesting questions was whether we could identify 4-5 patterns that are emerging in this programmable world. Two of the more interesting ones were "the invisible becomes visible" (contrary to the invisibility promoted by Mark Weiser in his original vision for ubicomp in 1991, but very much in keeping with a provocative paper presented at UbiComp 2006 by Yvonne Rogers) and "the transient becomes permanent" (though this seems to be more of a pattern on the web - Web 2.0 or Web 3.0 - in general). There were also a few references to Genevieve Bell & Paul Dourish's paper on Yesterday’s Tomorrows: Notes on Ubiquitous Computing’s Dominant Vision, and to some projects / companies I want to read up on soon, e.g., Olinda, Path Intelligence and Pachube.
The next session I attended was on decentralizing social networks, which was immediately contrasted with distributed social networks (DiSo), the latter framework being more amenable to central control and authority (and thus, it was argued, failure or abuse). I agree with an early observation that a big problem is that most people don’t understand who can see what (e.g., people posting personal comments on publicly posted photos not realizing the comments are also publicly viewable). However, much of the discussion was on very particular use cases that I personally did not find very compelling (e.g., sharing a photo with a few friends, who share it with a few of their friends, who then can annotate who is in the picture, which may extend beyond your comfort zone of who you intended to share it with). I think as soon as one adds granularity beyond the level of "share with friends" an "share with everyone", things get too complicated for most users (i.e., non Foo Campers) to understand. One of the more interesting suggestions was that there is a place-based metaphor that dominates the way most users think about the web (“my photos are on Flickr”), and this metaphor is less accurate as data and metadata can flow more freely to and through different sites and services.
Later in the afternoon, it was time for something completely different: The Lost Sport of Olympia (aka "the human labyrinth") an avant game led by Jane McGonigal. Players fill out a profile that helps them identify their "ancient strengths" (and thus the roles they might best play on their team). Most of the players form a human wall around the pattern of a labyrinth, and a designated - and blindfolded - runner has to find his or her way out of the labyrinth based solely on the directional humming employed by other players forming the wall. There is, of course, much more to the game, which has its own mythology and mystique, including its banishment from the Olympics (hence the name). I think it's easier - and more fun - to just watch the game than to explain it. I've posted some of my photos from the game on Flickr, and I'll embed a video taken by Loic Le Meur at the game site below.
There were a number of interesting Lightning Talks (and performances) in the evening, but we were explicitly asked not to blog about one of them, and given the intentionally edgy nature of some of those talks, I'll simply note that one of them was an update on one of the most fascinating topics (to me) in Web 2.0: object-centered sociality, the need for a system to present objects - photos, videos, etc. - around which to socialize and to support ways in which the objects can be used by participants, vs., say, simply putting people in a room - online or offline - and inviting them to talk to each other (without any objects to motivate socializing).
One of the most interesting discussions on Sunday was about the notion of "friending" - adding a link to another person in a social networking service - and how that differs based on the objects that are the center of the social networking service. For example, on the Flickr photo-based social network, where people primarily post photos that they have taken themselves, adding someone as a "friend" means you can follow their photostream, and thus keep up with what's going on in their lives (though it was also noted that Flickr friends can also simply be people whose photography you admire). . However, on a music-based social network, such as Last.fm (or MyStrands), people's playlists are simply metadata for objects (songs) primarily created by others ... and although music playlists reveal interesting dimensions of people's tastes, the music you are playing does not generally afford the same depth of revelation as the photos you are taking ... although if you are a musician, the music you are creating probably reveals more about you than any photo you could take.
I did not propose any sessions at this year's Foo Camp. I led a session on passion at Foo Camp 2007 that was immensely illuminating (for me) - and quite surprising in the way it evolved over the course of the conversation(s). This year, I was tempted to lead a session on managing parallel universes: dealing with the alternate streams of metadata - especially comments - created by aggregators such as FriendFeed (or our new strands.com (currently in private beta)). I'm not sure why I decided not to - probably a mixture of feeling I'm too inexperienced in using such aggregators to lead such a discussion (especially among a group of prospective participants like Foo campers), and waiting too long to approach the "big board", which was pretty full by the time I got there to see what sessions were being proposed (and feeling a sense that there simply isn't room for everyone to propose a session). Essentially, I just didn't muster the gumption to propose the session. Perhaps the regret I now feel about falling into that gumption trap will help me muster the gumption to post a future blog entry about it. But I think I've written [more than] enough for now ...
[Update: Aaron Quigley both pointed out a broken link (to the Bell & Dourish paper - also fixed above) and shared a story related to the photosensitive epilepsy seizure attack:
I was wondering when someone might make rapidly flashing screen into a computer virus. The idea occured to me when I lived in Japan in 1997 and a children's cartoon show "Pocket Monsters" or Pokémon had a explosion effect on screen which caused hundreds of children around the country to have seizures [ read more of photosensitive seizures ]. The epilepsy site attack, I'm amazed to say, even more nasty than the idea of a computer virus with this built in.;
Thanks Aaron!]
Recent Comments