Analysis of Everything

The point of the paper was to look at a collection, and analyze its creation and use.

Analysis of Everything, a community of geeks.

Nathan Oostendorp is a mild-looking young man with unruly brown hair, slightly hunched shoulders and a shy smile that still manages to look slightly mischievous. Born in Ann Arbor, Michigan while his father was in medical school, but raised and educated in Holland, Michigan, Nate recently celebrated his 21st birthday. He's also the driving force behind Everything, a popular web site which is an experimental, creative, frustrating and exciting collection.

    Everything is a collection of definitions, each of which is supplied by the users of the site. It's kind of like an encyclopedia where the user is able to add an entry for anything that hasn't been defined by someone else yet. Each definition is described as a node, and each node is linked. In the documentation for Everything, there are several attempts to explain what Everything is supposed to be. "An information casserole," according to the Everything FAQ is to "catalog all human knowledge, and show the interconnections (links) between all the people, places, things and ideas (nodes)."

    Everything is a project of BSI, or BlockStackers Intergalactic. Oostendorp started this company in August of 1998 with three friends of his: Rob Malda, Jeff Bates and Kurt DeMaagd. Even though the members of BSI are obviously flip about the conventions of business, as one can see from the full company name and the official company motto ("Pants are Optional"), the sentiment of these four men is completely serious. This group is very passionate about what they do, which is program. According to Oostendorp, BSI was developed at the same time Everything was created to act as an umbrella for the projects being done by the group listed above. "The hope was to have some sort of idea machine, or a self sufficient think tank," said Oostendorp.

    The actual creation of Everything is reminiscent to Samuel Taylor Cooleridge writing Kubla Khan, except that luckily no neighbor stopped Oostendorp from completing his work. "You mean my motivations (for creating Everything)? I'm not sure, they are difficult to deconstruct. Everything seemed to give me some sort of 'geek purpose'. A mission, etc.," said Oostendorp. A major tenet of belief for all members of BSI is the superiority of open-source software design. Open source software is a term penned by Eric Raymond, in an attempt to avoid the baggage associated with the term "free", and means software whose source code is distributed to whomever wants it, without financial obligation or proprietary control. Netscape is a prime example of this phenomenon. Any person can get the code that generates Netscape, and use it to develop rival products, without owing one cent to Netscape itself. Open-source has a long history, according to Tim O'Reilly. The original infra-structure from which the Internet was developed was a series of files shared between researchers at different sites. The Internet itself is a collection that has grown from and been shaped around this original code developed by people freely contributing their code. These original hackers developed code they thought was useful, and then told friends and colleagues about their work, thinking these folks might also garner some value from the code.

   Kevin Kelly, Executive Editor of Wired Magazine, describes a game of Pong played by 5,000 people at a computer convention. Half of the group had red paddles and the other half had green. A computer could recognize the movements of the paddles, and through the cooperation of 2500 people, the game paddles were moved. The group was able to operate though the application of a model where many minds cooperated to complete a task. This is an analogy to the drive of open-source software; many hands trying to accomplish a goal.

    Raymond makes a further analogy to the Cathedral and the Bazaar. He asserts that there are two methods of developing a piece of software. One can treat it like a cathedral, where the edifice is singular, organized and pre-determined, or like a bazaar where the structure is a collection of several smaller units, disorganized and constantly changing. To carry the metaphor further, one can see the difference as well in completed structures. A completed cathedral is immovable, stately and somehow distant, whereas a bazaar is constantly shifting and always changing, much more like a living organism. "The open source world behaves in many respects like a free market or an ecology, a collection of selfish agents attempting to maximize utility which in the process produces a self-correcting spontaneous order more elaborate and efficient than any amount of central planning could have achieved," says Raymond.
    The members of BSI are very committed to the philosophy and beliefs prescribed by the open-source movement. The other major BSI project, Slashdot, is a news service that is heavily populated with news items related to Linux, Perl, BSD, GNU and any other open-source software. Salon Magazine describes Slashdot as "the ubergeek news portal." It also highlights negative press surrounding Microsoft, which is widely seen as the antithesis of the open-source movement, especially with the release of the Halloween documents. Everything itself is almost like a showcase of open-source software. The computer managing the site is run on Linux, an operating system developed by Linus Torvalds. The story of Linux runs very parallel to the history of open-source itself. After all, Torvalds development of Linux mirrored in many ways the creation of the Disk Operating System by Bill Gates, since both were new ways to look at operating systems. The two men could not have taken more different paths with their products though. Whereas Gates used his innovation to start a major computing company that would someday become the target of the United States Justice Department for monopolistic tendencies, Torvalds immediately gave his innovation away. The source code for Linux became public domain, and anyone was free to take Torvalds' kernel and adapt it to his or her own use. There are currently 7,000,000 users of Linux, which does not even approach the number of people who use MS Windows. Still, many technology-oriented users swear by the Linux operating system, citing fewer crashes and greater control, and it is widely recognized as the main contender to the MS stranglehold on the operating system market.

    Everything also uses Perl to program common gateway interfaces (CGI's) to update the frequent changes that define the site. Perl is a programming language developed initially by Larry Wall, but which was released through the open-source paradigm. Perl is based on natural language patterns, and as Walls himself says "Perl is modeled on human languages. And the reason human languages are complex is because they have to deal with reality." He goes on to explain "This is important, and a little hard to understand. Since English is a mess, it maps well onto the problem space, which is also a mess"e; Similarly, Perl was designed to be a mess, though in the nicest possible ways." Other sites that use Perl for the same purpose Everything does are Amazon.com and Yahoo.com.

Even the graphics on Everything are produced by GIMP, an open-source image editing program. Practicing what they preach, BSI offers the source code for both Slashdot and Everything. Oostendorp even makes a few recommendations for how it might be useful to the individual user. For instance, the ease of hyper-linking would make Everything a great program for organizing class notes or a company training manual.

    The heavy presence of open-source software is important to understanding Everything in two ways. Firstly, the audience of Everything tends to be dominated by people from the open-source community, many of whom have made a direct migration from Slashdot. A recent poll conducted by Slashdot shows that the majority of the users are white, male, young and typically involved in the computer field, either as network administrators, programmers or students mostly. Everything does not collect that sort of demographic information, but analyses of self-descriptions in the personal nodes reveal much the same breakdown.

Secondly, open-source is at the very philosophical core of what Everything tries to do. The definitions can be provided by anybody, and approved or rejected by everybody. For instance, even though Oostendorp created the site, and maintains it still, he lists one of his favorite things about Everything to be "The fact I can explore it and be just as fascinated as joe user. All I've really done is create an excellent machine. I find what some people do in the machine absolutely amazing." The underpinnings of the open-source movement are so pervasive that it rarely goes recognized by users of the collection. It would be like the patrons of art museums questioning whether or not they should include paintings in their collection. However, Oostendorp does say "More important than the actual structure is the ideal that I feel Everything tries to convey: anyone can contribute, anyone can add- and everyone who comes in shapes the database in their own way."

According to the Everything FAQ, the idea for Everything was born in Malda's living room while he and Oostendorp were discussing some databases like the All Music Guide , and they thought about having a database that would be abstract enough that almost anything could be included in it. The problem with that theoretical database was maintenance. Any administrator for such a site would have a full time job making connections between concepts, much less monitoring the write-ups themselves for content. Oostendorp consequently designed Everything to enable the user population itself to keep order within the collection. Oostendorp's main role is to provide documentation and restart the server when the system crashes, as he modestly claims.

    In order to add nodes to Everything, one must register as a user. The only things required to become a user are choosing an account name and supplying an email address. Non-members may still freely browse the nodes, and may also vote on the write-ups for each node. The voting system is designed to allow member control of the content of Everything. One may either vote for or against any particular write-up. Currently, it is up to the administrators to go through and monitor the total of votes, "nuking" write-ups or nodes that have received a number of negative votes. In a future version of the site, to be released in the summer of 1999, an automated process will count up the votes and act accordingly. Also, soft links are created anytime any person, whether a registered user or not, moves from one node to another.

    There are currently over 50,000 nodes on Everything, 4500 registered users, and almost 250,000 links between the nodes, averaging 10 links per node. There are two ways to link concepts. A member may either create hard links by bracketing a word or concept within their write-ups, or any user creates links by moving from one page to another, which Oostendorp has taken to calling a soft link. The hard links never disappear unless a write-up is destroyed, while soft links fade with disuse. Even better, the soft links made most often are listed at the top of each category, signifying a hierarchical structure that is determined by actual use. So, while browsing one may do a very direct search from the links in the write-up, or a much more tangential search from the lists of Person, Place, Thing or Idea placed below the write-ups. Soft links also get "stronger" the more frequently they are accessed. Links that appear on the top of each list are the most common, while links at the bottom may have only been made a couple of times.

    This is important considering that any user may leap randomly from node to node, which still creates a soft link between the nodes. Rather than being a system flaw, this allows for a definite jump in logic that may strengthen previously untested cognitive links. Oostendorp sees this as a valuable tool for "thinking outside the box." In a way, the linking model of Everything closely follows semantic network modeling in human cognition. Hard linking is like formalized education, whereas the soft linking is like learning from experience. Just like in Everything, the more often two concepts in human memory are linked, the stronger the connection becomes. The similarity to human cognitive modeling has made Oostendorp remark half facetiously "Some concerns have been raised about the danger of Everything, because of its neural network structure, gaining its own form of intelligence and trying to take over the Internet. But what's to say it hasn't already? Even so, we think the idea of living under the domination of a computer intelligence of our own creation is pretty cool."

    The actual content of Everything reflects the interests of the users very directly. The start page for Everything lists a collection of nodes under a "Popular" category, as determined by number of links leading to those nodes. The nodes under this category range from "nate", which is Oostendorp's home node, to a definition of ambient music. Selecting the "random" option from the navigation list provides me with a node related to "Slugworth", defined as a person, the nefarious candy maker in "Charlie and the Chocolate Factory", by Roald Dahl who steals all of Willy Wonka's good ideas. Another search reveals "heineken" which is defined as an import beer that greatly facilitates coding in Perl.

    If the demographics of Everything at all resemble those of Slashdot, then the majority of users are males in their early twenties. A user who goes by the sobriquet of "pi" says "Almost any cool pseudo-underground Internet thing has at least a rough explanation about it on Everything." It is important to remember that Everything is still a relatively new collection, less than a year in existence. The nodes that first appeared related most directly to the interests of the user group, who might describe themselves as geeks.

Everything itself offers a definition of geek. "Geek is a term that originated in the circus world; the original geeks were those people who were somehow amazingly disfigured or had some freakish natural ability that made them worth observing. Nowadays those 'natural abilities' have more to do with learned abilities in things like computers or physics, but in some ways the geek is still interesting enough in his strange attraction to what others considered too utterly boring to be worth examining."
    The geek roots of Everything cannot be ignored in estimating its importance to its target audience. Oostendorp defines his audience as "Smart geeks mostly. People in love with complex systems. Everything is like information candy… brief, and sometimes clever. Everything users tend to binge and leave and come back." An uninitiated observer might consider the content of Everything to be somewhat immature. The collection allows the user to explain why Dana Scully is "such a babe" and has a node for every single character, no matter how minor, who appears on the Simpsons. Of course, Everything has defined its members via a node. "Members of Everything" has two write-ups, like every other node. 'HeadKnob' defines Members of Everything as "A bunch of cynical, Linux-using geeks who have read too much Douglas Adams, played too much Magic: The Gathering and like to post definitions from The New Hacker's Dictionary ." In the second write-up, 'jeffd' defines the same group as "A bunch of people who are willing to think differently and contribute to the improvement of the things in which they find fault instead of merely criticizing it." The two different write-ups have received roughly the same number of votes.

    Under the cool nodes, which are determined by system administrators, are a range of topics, mostly which either tickle the fancy of the site administrators, or are inside jokes among that same group. The random nodes at the bottom of the screen shot are a representative selection of Everything nodes, displaying the tenor of the whole site. Again, this ranges from personal experiences with technology related topics, to experiences with geography. For instance, one of the write-ups for "Michigan" is "If you seek a pleasant peninsula, look around you. Then move to Florida. It's cold, wet, and in the north, miles and miles of it are exactly identical."

    It's important to mention that it is extremely hard to get meaningful data out of Everything. Even on well-written nodes there is likely to be an editorial edge to the definitions, or some sort of ironic twist to the write-ups. Sometimes links lead to dead ends, where no one has found time to define a node that was specified, and the nodes themselves are not really written to educate. In some ways, Everything is more like a time capsule of cultural terms. If an anthropologist two hundred years from now wanted to know what geeks in the late part of the 20th century thought was important, this would be her or his ticket. Not that each node is empty of informational value, but the emphasis is definitely on the "coolness" of the definition rather than the accuracy of it. It is like an arm wrestling contest for geeks. Who can come up with the wackiest definition that is still accurate?

    Several of the nodes that appear in the Random category are personal nodes, which provide self-descriptions of regular users. This is one of the most important functions of Everything. More than anything else, Everything is a community where a user can be assured that there are other people who get as obsessed about these topics as they do. This is an important occurrence for several of these users, many of whom might be considered outcasts in their "real" social groups. Despite the financial success of several notorious geeks, Bill Gates for instance, the interest in technology and other fringe topics still sets one apart in many of the social environments faced by the users of Everything. Several personal nodes mention the difficulty in finding a significant other that shares their interests. The node for "alien girlfriend" says "Man I bet at least an alien girlfriend would be interested in technology. They are probably all Kernel hackers too." Several nodes are dedicated to lamenting the ratio of men to women in computer science and engineering classes.

    An odd effect of the community formed through Everything is that it prevents, in large part, direct contact between its members. Not even Oostendorp is able to contact all the members of the site, and it is up to the individual whether or not they want to provide any personal information about themselves in their personal write-ups. There have been several times while browsing, and checking to identify the user who wrote a write-up I particularly enjoyed, that I ached to be able to contact that person. Some of the self descriptions in the personal nodes sketch profiles of people I would really like to meet, but have no way of doing so. This means that the community created by Slashdot is close-knit and familiar, but mostly through the definitions they have provided. It is amazing the sense of a person you can get by what they choose to write about, and how they write it. Without, for the most part, knowing each other's names or faces, the users of Everything have established a significant and meaningful community of fellows.

    The nodes on Everything create a shared meaning for this community of users. As part of writing the paper, I requested and received administrative rights on Everything. It was now within my power to "nuke" nodes, erasing them from the Everything neural net. It was amazing how reluctant I was to do this, a sentiment echoed by the other people who have the same access rights to the site. There is a feeling among the administrators of the site that the nodes are untouchable unless they are obviously wasteful or facetious. The context provided by sharing meaning in this way has a hint of the sacred underlying the sheer enjoyment of the collection. Everything provides this group of people with similar interests a venue with which to define themselves, at least through their thoughts and beliefs about several topics they hold dear.

    However, the very content that draws this group of "intelligent geeks" together may exclude other groups. One node entitled "Showing Everything to my girlfriend" recognizes that there is nothing in these nodes that would interest the significant other of the writer. This is a common artifact of the open-source movement. Open-source software tends to be those things that interest programmers, like operating systems, networking and programming languages. It definitely ignores things like word processors, spreadsheets and other basic office applications, either because they are not interesting to the varied programmers working on open source code, or because there is no drive to bring those programs beyond their current state. Meanwhile, companies like Microsoft and Corel work actively on those office products that get ignored by the open-source community. Another quirk of the open-source community that reflects itself in Everything is a severe insularity. The average consumer may have distinct differences with programmers as to what constitutes useful and elegant programs and documentation. This tendency may broaden the gap between these open-source advocates and the rest of the world. This could also happen on Everything if it never takes off past the original geek population who made the first contributions to the collection. It never says directly in the documentation that Everything is a site for geeks, by geeks, but increasingly technical or obscure write-ups will definitely push the average user away from the site. Since it is part of the lofty ideal of the developers to make Everything show the links between every person, place, thing and idea, one can assume they would like to avoid this cordoning off of their population. The new version of Everything will need to satisfy the needs of the audience who have to this point contributed so deeply of their times and identities. At the same time it will need to integrate a new phase of users who will link some of those "geek concepts" to a broader range of ideas and concepts.

    If Everything wants to stay within the circle it has already drawn, it should be able to easily continue the same sense of community it has already engendered in its members. However, the real challenge will be to integrate a broader group of members without alienating the core community. There are a few ways to do so.

    Everything could encourage several sister sites that target other specific populations. For instance, an Everything type intranet could be established for Sports Fans, so that the "Detroit Tigers" node would link to all players' names, or a node for "fly ball" would define that term. This model has been mirrored in business practice as a form of franchising that does not replicate the original institution perfectly, but instead clones the values or methods of that institution. You could have Everything sites that targeted as many groups are there are usenets, another method of self-definition by a community.

    Another idea is to broaden the scope of the current site by employing administrators to assign "cool" nodes, and even to write nodes that are broader than the defined geek purview. Everything has already done this to some degree, though not in an organized manner.

    The actual inherent barrier to the broadening of Everything is the inclination of the community currently driving it. When Oostendorp developed the site to act autonomously, he also surrendered the control that would allow him to single-handedly change the culture of Everything. It will ultimately be up to the current members of Everything to find value outside of the topics already addressed by the site. This would take a precipitating event of some sort, as one of the pride points of this group is their narrow focus, and how that focus separates them from others.

    Still, even confined to its current community of users, Everything is an exciting, dynamic organism. The collection itself is engaging, almost addictive in the way the World Wide Web was in 1995, and certainly fun. The description of the site as "information candy" seems very appropriate, because while it is difficult to pull any factual information from the nodes, it is still satisfying on many levels. Yet, besides its entertainment value, Everything provides an important service to its community. It's a place for the disenfranchised to come and realize they are not the only ones who think in this way, that there are others who share their interests in these fringe topics. While in some ways, the community of Everything is like a herd of cats, like any family, they provide support to each other, sometimes fight, but always know whom to trust.