July 23 , 2006
Why believe Wikipedia?: Simply by appearing in the Britannica, an article has credibility. But that's not true for Wikipedia because you might hit an article a moment after a loon has altered it. Yet, Wikipedia has (and deserves) credibility, in part because of its willingness to acknowledge its fallibility.
The end of the story (Or: The tyranny of rectangles: Journalism can't get stories right because the world doesn't fit into rectangles.
Book report (Or: My obsession): The first draft of my book is done. Here's a brief report on Chapter 8.
Walking the Walk:
What I'm playing: Gun is disappointing. Indigo Prophecy progresses from cool to idiotic.
Bogus contest: Metadata for traditional authorities
Why so long between issues
As one of the articles below says, I've become mildly obsessed. Every day I wake up and think, "I really should be working on a new issue." But every day there's just a little more of my book I should be writing. I have been unable to tear myself away.
Now I have a week between drafts. So I'm filling my interstice by providing a little something for you to fill yours.
Lord bless the interstices!
A month later: I turned in a complete first draft a couple of weeks ago and yesterday heard from my editor. She likes it. In fact, she was enthusiastic. She has edits and questions, and some ideas about how to make it a bigger seller, but at least I did not get the dreaded reaction: "It's a great start. You've given us lots to work with."
So, in a week I'll get her commented version and will spend August sanding and polishing, as well as ripping out some rotten floorboards. Metaphorically speaking.
Why believe Wikipedia?
Simply appearing in the Encyclopedia Britannica confers authority on an article. Simply appearing in Wikipedia does not, because you might hit the 90 second stretch before some loon's rewriting of history or science is found and fixed. Yet, Wikipedia is in some ways as reliable as the Britannica, and in some ways it is more reliable. Where does it get its authority?
There are a few reasons we'll accept a Wikipedia article as credible.
First, we apply the same rules of thumb as we do when listening to someone for the first time: Does she sound like she knows what she's talking about? Does she seem fair? Does she seem to have some perspective? Does she blatantly contradict herself?
And, we are generally more likely to believe a major article than one on an obscure topic because it's more likely to have been worked on by many people. Plus, we may already know something about the topic. If the article on the JFK assassination says he was poisoned by Rasputin, we'll be disputin' that article.
The article gains credibility if we see it has a long edit history. It becomes yet more credible if the discussion pages are long and rich. (As someone pointed out to me a few months ago — who were you, dammit? — those pages are going to become remarkable artifacts as future historians try to understand our attitudes and beliefs. Imagine we had discussion pages for the 1950's Wikipedia page on segregation.)
There's one more sign of credibility of a Wikipedia page: If it contains a warning about the reliability of the page, we'll trust it more. This is only superficially contradictory. Wikipedia has a page that lists the available notices. Here are some of the warnings available in the Disputes category.
The fact that Wikipedia encourages us to use these notices give us confidence that Wikipedia is putting our interests over its own.
So, why is it that you don't see such frank notices in traditional sources such as newspapers and encyclopedias? Is it because their articles don't ever suffer from any of these human weaknesses? Oh, sure, newspapers issue corrections after the fact, and "This is non-neutral opinion" is implicit on the Op-Ed page. But why isn't there any finer grain framing of the reliability and nature of what's presented to us in their pages? Can we come to any conclusion except that traditional authorities are more interested in maintaining authority than in helping us reach the truth?
Which in the long run will be devastating to their credibility.
danah boyd has a terrific post on her problems getting the entry about her at Wikipedia corrected, pointing out the extent to which Wikipedia relies on the media. As other Wikipedians have pointed out, the person danah was dealing with there does not speak for all Wikipedians. In fact, no one speaks for all Wikipedians. But do check the discussion page at Wikipedia.
The Scaredy Cat Encyclopedia
The Encyclopedia Britannica has refused my request to interview an editor for 15 minutes about the process by which it chooses authors. I explained that this is for a book. But, the head of the Britannica's communications group decided - based on what? - that they don't want to support people who are "cheerleading for the downfall of businesses that they deem to be part of an old regime."
All part of the command-and-control mentality at some of our great institutions of knowledge.
Go team! Sis boom bah!
The end of the story (Or: The tyranny of rectangles)
If you've ever been part of a story covered by a newspaper, it's a near certainty that you didn't think the story got it exactly right.
1Even if there were no outright mistakes, you read it thinking that the emphasis was wrong, that it didn't quite capture all sides, that there was more to it than that, that a turn of phrase was prejudicial. You would have written it slightly differently. At least.
This is not because reporters aren't good at their job. By and large they are, and it is hard job requiring skill, experience and persistence. It also generally doesn't pay that well. The problem is not with the reporters. Lord bless them and multiply them. The problem is with the notion of "the story." And the real problem with "the story" isn't the story, it's the "the." There are better and worse ways of recounting what went on, but there is no one right way. But, the physics and economics of paper just about require that the story be told in one and only one way. Paper fixes ideas the way flypaper fixes flies.
Newspapers are beginning to recognize that the Web gives them a way to get past The Story. Sort of. Opening up a comment thread next to an article, or pointing to blog posts that discuss it are starts, but that still presents The Story as the object at the center of peripheral discussions. The page in which The Story is embedded can — and well may — become the primary artifact, and it will contain multiple stories, from multiple people, about multiple topics.
Imagine what a page about, say, the Iraqi elections or a new stem cell discovery could contain. Stories have to repel alternative versions. Stories are defensive. But topics are warm and welcoming. Topics accrete stories the way corral reefs (reeves?) accrete diverse life forms.
So, maybe this is a way newspapers can hold onto their value. Stories become commoditized. The New York Times could be the place that, yes, writes stories but that, more importantly, assembles topic pages that contain the stories, pictures, archival information, maps, video, and links to what the rest of the world is saying. In fact, that's pretty much exactly what the NY Times is doing with its topic pages. But, The Times seems to be looking at topic pages as an additional service, one that they hope will let them rise in Google rankings so that when you search for, say, "stem cells," their page will show up ahead of Wikipedia's. The Times is being slow to create these pages. And they're still charging for access to their archival material. The Stravinsky topic page lists an article from 1987 about a judge blocking a Stravinsky biography, but the link takes you to a $3.95 tollbooth. This will prevent the pages from being broadly useful and ranked high by Google. It's also annoying.
Imagine if the Times conceived of itself as a type of agile think tank, with the topic pages as their main artifact. It would have to open up its archive - a boon for an informed democracy - and could make money on ads. The Times' value wouldn't be its ability to generate stories. The Times would be about understanding what's going on. And for that we need lots more than stories. If the Times doesn't do it, then someone else will, aggregating stories from multiple newspapers, and assimilating them to powerful, rich topic pages.
Stories just want to be contextualized. Context is king. Or queen. Or at least The Dauphin.
But, topics suffer from the same limitation as stories: They have to be fit within rectangles. We all understand that there is no one story about the current Mideast conflict. But if we go up a level to build a page about the topic that can embrace multiple stories, how do we decide what the topic is? Is it the Mideast conflict? Rise of Islam? History of Israel? War on terrorism? Western colonialism and its aftermath? Through hyperlinks we can bust out of the boxes, but we nevertheless have to build the boxes in the first place, just as physical books need to be put on a particular shelf.
We are on the verge of a serious change, though. Tags allow readers to categorize items the way the readers think about them, so that one person's "Israeli aggression" is another person's "Israel's war on terror." By using tags, readers can collectively decide on the rectangles, their contents, and their links. Chains of links can eventually lead across rectangles to shared materials, although the great danger is that we will box ourselves in...but that's a topic for another day.
As it stands, human editors build better boxes than machines do. The Times' topic pages that its editors have worked on are noticeably better than ones auto-assembled on the basis of metadata. It may always be thus, although I expect that the best rectangles will come from individual amateurs and collections of readers engaged in perpetual discussion.
Why have rectangles at all? Attention has become synonymous with focus. But attention seems in fact to be a dialectic between focus and openness to distraction. And that makes the paper rectangles by which we've been expressing ideas for millennia unnatural when compared to the rectangles-with-links we're building for one another on line.
1For example, this article in the Harvard Crimson got my position on the reliability of Wikipedia backwards. Argh. And the entry on me in Wikipedia links to it as a reliable account of what I said. Double argh.
Jay Rosen, chair of the journalism department at NYU, today announced on his weblog a new site, NewAssignment.net, intended to bring together professionals and the wide world of citizen journalists. It's an interesting experiment that may help validate that what Jeff Jarvis calls "networked journalism" can produce high-quality results. But, because it's being done in public on a Web site, I suspect that the site where each story is developed is going to be more interesting, useful and revelatory than the story itself.
Book report (Or: My obsession)
I don't have Attention Deficit Disorder. I'm not autistic or have Aspbergers (by the way, here's a sure-fire laugh-getter for your eleven-year old: Do they have Aspbergers on Uranus?). I do have 50+ Fat Guy Type 2 Diabetes, but other than that I don't have any of the trendy diseases, with one exception: A few years ago I realized that I have a touch of obsessiveness. Generally it's nothing too disruptive. But for the past twelve months, it's surfaced as an inability to think about anything except the book I'm writing. So it could be worse, unless you live with me and don't agree that the 9th grade performance of The Music Man is really all about the tyranny of taxonomy. But answer me this: Why do all 76 trombones have to be clustered in the front? Is it natural or even useful for every one of the 110 cornets to be near the trombones? Why carry the tyranny of rows and columns all the way into musical performances? That's just sick.
Anyway, one odd manifestation of my obsession is that I never get to a point where I'm ready to talk about the book. There's always just one more day's writing to do. And, by the way, that is exactly why JOHOs have been so far apart.
So, now I've turned the whole rough draft in and my editor likes it. As I mentioned in a box at the head of this issue, I have a summer's worth of revising to do, but right now I'm waiting to get the marked-up manuscript from my editor. Ah, the sweet interstice of summer!
I'd give you an overview of the chapters I've written since last we talked, but I'm avoiding re-reading them. Suffice it to say that Chapter 5 is about how we're going to identify the pieces and parts that constitute The World Wide Mess. (I actually don't remember exactly what 5 is about, but I assure you that it will be absolutely fascinating and has already been optioned by George Clooney.) Chapter 6 is on the socializing of knowledge - social filters, Wikipedia, etc. - and especially on the effect that's having on the news media. Chapter 7 is on the importance of the implicit and the difficulty computers have in dealing with it. Chapter 8 is about the virtue of messiness. Chapter 9 draws conclusions about the nature and role of knowledge, and also includes a delightful recipe for sugar-free tiramisu. You can't go wrong with tiramisu!
I don't want to talk about 9. So, here's what chapter 8 talks about.
As discussed in previous issues of Joho, the book pretends there are three orders of order. In the first, we organize the objects themselves. In the second, we separate the metadata from the data and organize the metadata (e.g., a card catalog). In the third order, the data and metadata are both digital, so we can come with new ways of organizing them free of the constraints of the physical. Chapter 8 begins by saying that messes in the first two orders are inefficient and make life worse, but in the third order, a messy pile with lots of implicit and potential relationships within it actually reverses entropy. The pile itself can stay messy as different people organize the metadata as they want. For example, if our family photos are in a messy pile, we can't find anything easily. If my wife wants to organize them by year and I want to organize them by person, one of us has to lose. But, we can each organize our digital pile of digital photos the way we want without actually rearranging the digital photos at all. The more metadata attached to the photos and the more relationships discerned among them, the more potential for fruitful ways of organizing it.
The chapter then looks at the history of business org charts and why they have so much white space. Nowadays, companies are realizing the value of mapping the messy social networks that the org chart hides. But messiness isn't simply a part of our social order. The work of Eleanor Rosch showed in the 1970s that we order our concepts in sort-of kind-of relationships around clear exemplars (prototypes), thus punching Aristotelian essentialism right in its well-defined nose.
Then I look at the Semantic Web as an attempt to clean up the messy Web. I trace it back to Tim Berners-Lee's previous project, Enquire, a way of modeling systems by using sets of links that had meanings such as "x is part of y" and "x depends on y." The Web succeeded in part because the links are simpler than that. The Semantic Web would like to restore the rich meaning of links. Sounds good, but as a result, some versions of the Semantic Web hark back to the attempts to build universal, comprehensive taxonomies that I made cruel fun of in previous chapters. Other versions take a messier and more incremental approach. Certainly there are areas where the messy and the clean variations of the Semantic Web make sense, but there's also lots of sense in loose, unrestricted tagging and clustering, which leads to the loose order that Rosch talks about.
Conclusion: We're learning how to know in sort-of kind-of ways, in addition to our traditional ways.
Here's a snippet from Chapter 8:
...Rosch was in New Guinea to study how one of the local tribes, the Dani, categorized color. Color categorization was an interesting field for anthropologists and linguists because the 7.5 million colors humans can perceive form a continuum with seemingly no natural divisions. Yet, pioneering work by Berlin and Kay in the late 1960s showed that across 98 different languages, there seem to be only eleven basic color categories. Russian has no single word for blue, the French have no single word for brown, and the Dani, remarkably, have only two basic colors, but every culture's basic colors seem to come only from within that group of eleven.
Rosch showed the Dani color swatches and thirty seconds later asked them to pick the color from an array of samples. She found that they identified the basic colors more accurately than non-basic colors. This was precisely the same as with Americans and subjects from twenty-three different language backgrounds. It seems that although we disagree about how many basic colors there are and even what they are, when we lump and split we identify some swatches as a prototypical example of colors and others as sort-of, kind-of, to-some-degree examples.
This flies in the face of the Aristotelian idea. For Aristotle, a thing is a member of a category if it satisfies the the category's definition. Thus, anything in a category is an equally good example of it. After all, it shares the essence of the category. But, when it comes to color, it seems that we don't work that way. Tomato-red is a great example of red, the sort of red you could point to if someone didn't know what "red meant, but a setting sun may be reddish, an orangey-red, or a light red – it's red, but not a good example of red.
Rosch quickly realized that this might be true not only of how we deal with colors but how we generally think about our world...
Middle World Resources
Walking the Walk
Raytheon is supplementing its intranet's internally-developed taxonomy with a folksonomy. Librarians pay attention to the tags users create, creating synonyms where appropriate. It's a good hybrid case.
For the Hyperlinked Organization
I like Diigo in theory. It's a social bookmarking site like Delicious, but it also lets you highlight text and leave sticky notes that other Diigo users can find. (This is pretty much exactly what Third Voice did about six years ago.) All this is swell. I just don't know if I'll be using it regularly, in part because there are so many places I can put social bookmarks that I'll soon need a meta-social-bookmarker to keep track of them.
I finished Gun, enjoying the fact that it dares to be set in something other than space-grunge corridors of an interstellar vehicle beset by grues. It's good to get a breath of fresh air now and then. But, the narrative is pretty dopey and the game play is too easy. Like Grand Theft Auto, there are plenty of side missions, but they're also too easy. Plus, every time I start it, it puts me back in an out-of-the-way corner of the landscape from which I have to ride, annoying me unreasonably.
I'm about a third of the way through The Indigo Prophecy, which is more interesting as genre than as an instance of the genre. It blends an adventure game with an interactive movie. Plus there are something like action sequences that require you to press buttons in sequence. The story so far is a too-easy mish-mash of haunted possession and a serial killer, but it's fairly interesting, if you can avoid the Angelina-Jolie-ish hot female detective and her embarrassingly stereotyped black-hip partner who is accompanied by a soul soundtrack. That aside, it's not bad. Really.
Later: I finished The Indigo Prophecy. Actually it gets bad. Very bad. The last third introduces ridiculous and uninteresting elements, leading to an anti-climactic and quite stupid conclusion. Avoid it.
Bogus contest: Metadata for traditional authorities
You know how Wikipedia encourages editors to stick in warnings of various sorts? What, did you skip the first article in this issue? You better go back and read it because it will be on the test.
So, your challenge is to come up with appropriate warning stickers for traditional knowledge authorities. For example:
New York Times
This column is predictable. This column is redundant with the previous n columns by this author.
Insufficient Chagrin. This article refers to Iraq without acknowledging that our poor editing practices led to a year of inaccurate reporting that helped get us into the war.
Right in theory. This map of downtown Boston is correct according to the final plans for the Big Dig, and thus is wrong.
Corporate quarterly earnings press release
Questionable self-knowledge . The self-description at the end of this release has been carefully constructed not to reflect the company's actual business but to enable us to claim that we are the "global leader" in something.
Answer by Bush administration press person
How about Those Red Sox? This answer is intended to change the topic to something more pleasant.
So, go forth and make risible. But please don't make fun of the Britannica. It's too important for that.
JOHO is a free, independent newsletter written and produced by David Weinberger. If you write him with corrections or criticisms, it will probably turn out to have been your fault.
To unsubscribe, send an email to email@example.com with "unsubscribe" in the subject line. If you have more than one email address, you must send the unsubscribe request from the email address you want unsubscribed. In case of difficulty, let me know: firstname.lastname@example.org
There's more information about subscribing, changing your address, etc., at www.hyperorg.com/forms/adminhome.html. In case of confusion, you can always send mail to me at email@example.com. There is no need for harshness or recriminations. Sometimes things just don't work out between people. .
Dr. Weinberger is represented by a fiercely aggressive legal team who responds to any provocation with massive litigatory procedures. This notice constitutes fair warning.
Any email sent to JOHO may be published in JOHO and snarkily commented on unless the email explicitly states that it's not for publication.
The Journal of the Hyperlinked Organization is a publication of Evident Marketing, Inc. "The Hyperlinked Organization" is trademarked by Open Text Corp. For information about trademarks owned by Evident Marketing, Inc., please see our Preemptive Trademarks™™ page at http://www.hyperorg.com/misc/trademarks.html
This work is licensed under a Creative Commons License.