Author: cgerrish

Unemployed philosopher

The Twittering Machine: A Network of Accelerants & Silences

Published November 3, 2009 by cgerrish

In 1922, the Artist Paul Klee completed a work entitled Die Zwitscher-Maschine, which has been translated into English as The Twittering Machine. The work is part of the Museum of Modern Art’s collection. The catalog description delves beneath the surface interpretation of a machine for making pleasant bird sounds:

Upon closer inspection, however, an uneasy sensation of looming menace begins to manifest itself. Composed of a wiry, nervous line, these creatures bear a resemblance to birds only in their beaks and feathered silhouettes; they appear closer to deformations of nature. The hand crank conjures up the idea that this “machine” is a music box, where the birds function as bait to lure victims to the pit over which the machine hovers. We can imagine the fiendish cacophony made by the shrieking birds, their legs drawn thin and taut as they strain against the machine to which they are fused.

I’d like to take a moment to delve beneath the surface of the machinery that makes up the micro-messaging ecosystem. The micro-messaging service Twitter approaches a kind of frictionless channel for messaging. Everything about it seems to be built for growth and speed. After the substantial speed bump of signing up and orienting yourself in the system, both listening for, and creating messages are very simple. Select the other people (and pseudo-people) you’d like to follow, type fewer than 140 characters of hypertext into a text field and press update.

It’s been noted many times that the follow social structure – the directed social graph – allows for the rapid expansion of social linking. A participant is able to follow many others without the requirement of reciprocity. I follow you, you don’t have to follow me. And the retweet has emerged as a method of transmitting messages across multiple overlapping social graphs. Hash tags have become a conventional method of indicating a message tied to a particular event. All of these tools are dedicated to the acceleration of the growth of the Twitter network. Twitter’s trending topics and search functionality allow users to find high-velocity memes as they begin to achieve broad circulation through the social mesh.

Examples of the benefits of the speed of the Twitter network abound – news about the earthquake in China, the plane landing on the Hudson River, reports on wild fires in Southern California, etc. The news media, feeling somewhat usurped here, has also brought the negative aspects of this lack of friction in the system. Usually this kind of story is meant to point to the filtering role of their own editorial facility. The New York Times put a Twitter contretemps involving Stephen Fry in the first section of their newspaper.

…that little tweet set off a frenzy of vitriolic attacks and counterattacks on Twitter, drawing an untold number of people into an increasingly charged debate and thrusting brumplum — in reality a man from Birmingham, England, named Richard — unhappily into the public’s angry glare. It was an example once again of the extraordinary power of Twitter to distribute information and to sway the opinions of vast groups of people in tiny amounts of time.

While the frictionless micro-messaging environment can accelerate important, or amusing, messages (news, stories, jokes, musings), it can also assemble and ignite an angry mob at lightning speed. And if we look at the tool set we’re offered, everything is oriented toward accelerating people, memes and the growth of the Twitter network. In this frame, value is equated with high velocity and numbers of retweets, likes, mentions, @replies, followers and now, list memberships. We’re looking for what, or who, is going viral.

But if we examine the often banal items that achieve virality, we have to question the value = virality equation. Often the high numbers are a result of lowest-common denominator dynamics, it’s not the cream rising to the top. There are certain kinds of viral items that we’d prefer to be vaccinated against. We’d like a micro-messaging Center for Disease Control warning us about time-wasting viral memes spreading through the Network. Where are the tools to retard the spread of a meme? Where’s the don’t like, or the visible gesture of hiding something from your messaging stream. It’s as though we’ve put a poll in the field that only asks for a positive response and excludes all negative reaction. Feedback loops require both positive and negative inputs.

Feedback is a mechanism, process or signal that is looped back to control a system within itself. Such a loop is called a feedback loop. Intuitively many systems have an obvious input and output; feeding back part of the output so as to increase the input is positive feedback; feeding back part of the output in such a way as to partially oppose the input is negative feedback.

Negative feedback helps to maintain stability in a system in spite of external changes. It is related to homeostasis. For example, in a population of foxes (predators) and rabbits (prey), an increase in the number of foxes will cause a reduction in the number of rabbits; the smaller rabbit population will sustain fewer foxes, and the fox population will fall back. In an electronic amplifier feeding back a negative copy of the output to the input will tend to cancel distortion, making the output a more accurate replica of the input signal.

Positive feedback amplifies possibilities of divergences (evolution, change of goals); it is the condition to change, evolution, growth; it gives the system the ability to access new points of equilibrium.

One method of judging the relative merit of the memes circulating through the stream is to pay close attention to the silences, the negative gestures. While the wisdom of crowds may have crowned an item the real-time meme of the moment, there’s a simple way to filter for value. Using a small portfolio (lists, if you will) of people who’s opinion you respect – look to see if they remain silent on a topic. In the task of discovering value, this kind of silence is golden.

In the early stages of growth, accelerants are essential. As a Network matures, unless it develops a deeper tool set, it simply becomes a twittering machine. The value it appears to create is illusory, and so the commons begins its descent to knee-jerk reactions to the high numbers generated by the lowest common denominator. Perhaps the inevitable result is that overlapping publics contract and the social space becomes a much more private circulatory system. The crucible of dispute, debate and dialogue that produces real value may only emerge away from the public network of accelerants where the social contract allows such differences (both positive and negative feedback). One of the better statements of this kind of contract was for the Nettime mailing list:

Nettime was vector for experimenting with net critique that would confront it with the possibility of inventing new forms of discourse and dialogue in a new medium. Consensus is not the goal. There’s no governing fantasy according to which the differences within this ‘group’ will on some ever-deferred day be resolved. the differences are Nettime; they might be dialectical, implying each other, or they might be differential, making absolutely no reference whatsoever to each others’ terms. Net critique, if understood as a shared practice in and against a never pre-defined techno-local environment, contains many modes of possible participation.

In navigating the electrical storms and fires of the micro-messaging Network, we’d do well to have some tools in our kit in addition to that can of gasoline.

4 Comments

Shadows & Light: Privacy in the Panopticon

Published October 30, 2009 by cgerrish

Before the turn of the millennium, Scott McNealy declared privacy dead:

The chief executive officer of Sun Microsystems said Monday that consumer privacy issues are a “red herring.”

“You have zero privacy anyway,” Scott McNealy told a group of reporters and analysts Monday night at an event to launch his company’s new Jini technology.

“Get over it.”

McNealy’s comments came only hours after competitor Intel (INTC) reversed course under pressure and disabled identification features in its forthcoming Pentium III chip.

At one time, privacy was a function of a general laziness and the unlinked quality of information. While there may have been lots of publicly available information about a person, it was rather difficult to track down and assemble. We’ve developed a whole mythology around the kind of person who can root out the details about a person and put the pieces together into a picture that makes some kind of sense.

There was a kind of power in the invisibility we once had. Oddly, it was a kind of anonymity that was derived from the density of the urban environment. The city was a place you could go to get lost, to start over, to create a new identity. That’s why it took a detective to find the traces and clues that filled out the picture of a person. Today, that kind of invisibility has mostly vanished. If I want to know something about Sergey Brin, I can use any number of services that will scour the Network looking of publicly available information, and then I can pay for information that’s more obscure or privately held. Shoe leather is no longer a requirement.

Just as there’s a kind of ‘security through obscurity,’ there was a kind of privacy through obscurity. The methods by which information about a person used to be stored were enshrouded in shadow, even darkness. One piece of information wasn’t linked to the next. The trail was obscured, you had to stumble through the darkness to get from one piece of information to the next. Now information is linked into a web– it’s created, searched, and collated. In the UK, surveillance cameras are used to create a visual real-time mesh of video that can track you through your day. You are being recorded, it’s just a question of whether anyone is currently looking at the data or not.

…under a law enacted in 2000 to regulate surveillance powers, it is legal for localities to follow residents secretly. Local governments regularly use these surveillance powers — which they “self-authorize,�? without oversight from judges or law enforcement officers — to investigate malfeasance like illegally dumping industrial waste, loan-sharking and falsely claiming welfare benefits.

The private moment, that little space between this and that, the in-between time when no one is looking— this invisible space is growing smaller and smaller, the more connected we become. Privacy through obscurity is no longer a dependable strategy. The things that were hidden in plain sight, are now easily found.

There is some data that remains private. Our medical records and financial records are two examples of personal data that is actively encrypted and kept private. Generally a court order is required to pry open these vaults of information. In some sense, that’s the new definition of privacy. It’s data that can be accessed by the individual, the data custodian, and, by court order, the government. In addition, should this data inadvertently leak out from the data custodian, the individual has a well-established legal recourse against the custodian.

In order for the private to remain truly secret, it would need to be unconnected. As the connections between us are made visible by our electronically networked environment, we begin to see that we have always inhabited networks of one kind or another. To be unconnected to all networks is to no longer be among the living. The private is something that we are prohibited from sharing based on a social or legal contract. Viewed as a system, the private requires more energy to maintain its contracts regarding the non-sharing of information. Linking private personal data among private systems of record while still honoring the non-sharing contract takes even more energy. The network itself doesn’t distinguish between private and public information packets.

Gregory (Scotland Yard detective): “Is there any other point to which you would wish to draw my attention?”
Holmes: “To the curious incident of the dog in the night-time.”
Gregory: “The dog did nothing in the night-time.”
Holmes: “That was the curious incident.”

And just as we can deduce the nature of hidden facts based on the dog who didn’t bark in the night, the private can often be deduced by correlating public gestures/connections to and through the locus of personal identity. But privacy isn’t dead, it’s just as it always was– an agreement among a group of people to enact useful boundaries on the sharing of information.

Comments closed

The Page and the Item: The Dynamics of Context and Collection

Published October 28, 2009 by cgerrish

songs-of-innocence-title-page-william-blake

In thinking about building things on the web, I stopped to consider the raw materials from which we build. HTML markup retrieved from a web server and rendered by a browser, that’s where it all started. But even at that early moment, there was the implied structure of the document. The markup that existed was there to render visible the form of the academic paper. Headings, paragraphs, quotations, tabular data display– these are the formal elements of the document.

And very early on, the metaphor of the page gained purchase. We nodded our heads and spoke of the ‘web page.’ The static web page and the static book page have similar kinds of boundaries. The web page could theoretically be infinitely long, but the usability experts indicated that users didn’t scroll much beyond the length of a book’s page. And just like that, an infinity was tamed. The edge of the world was discovered.

As the content on a web page became dynamic, infinity migrated to the combinations and permutations a database could produce. As long as the data continued to grow and change, the items presented in a particular page could be of an infinite variety. The boundaries to the north, south, east and west remained consistent with the book’s page, but the objects emerging from the depths of the backend could be practically without end.

The document and the page have been structurally ingrained into the architecture of content management systems such that the smallest building block becomes a page linked to a hierarchical document tree. The elements that can be placed into a page are those for which the system has templates. And while most systems allow the manual writing and insertion of raw HTML, it’s a practice that is discouraged because it ruins the uniformity of the CMS’s output. The content management system is an industrial machine for creating hierarchically organized sets of pages.

The other major organizational structure on the web is time-sequenced content. To some extent, news media takes this approach to organization, new material is published each day to replace the material from the previous day. What’s lacking in the model is the continuity of sequence. Yesterday’s news is fish wrap, rather than the next step in a sequence. Blog posts and Tweets (micromessages) have the form of a sequenced set of texts by an author or group of authors. In this sense they are more like the output of a columnist or the writer of serial fiction. Blog posts can also be assigned categories and tags so that they can be sequenced across other conceptual frames. Tweets don’t have the extra infrastructure to house categories and tags, so the practice of adding a hashtag has been bolted on. More elegant solutions like the original track feature have failed to resurface.

Rather than referring to time-sequenced pages, here we more commonly talk about items in a feed. We’re interested in the source of the feed in order to gauge its authority, along with it’s velocity and trajectory. And unlike a hierarchical organization of pages, the items in a time-sequenced feed need have no semantic relationship to each other. The items are such that they can be organized in an arbitrary large variety of collections either within a particular feed or among a diverse set of feeds.

tweet_link

The page and the item converge at the URL (Uniform Resource Locator). Because of our page-centric view of the web, here we’d like to say this is the point where the item becomes a page. And yet, the web becomes much more interesting if we resist this temptation. The item has no native context, the page wants to own its context. The item allows the user to create a collection, a playlist, a feed that suits her own needs, wants and desires. The page needs to reinforce the hierarchy of which it is a part. The key to the dynamic context of the item is that it both has a URL and can contain a URL, and it doesn’t have a single right context.

Information architecture has largely concerned itself with pages and hierarchies, and the economics of the web have centered around the page-view model. As the item begins to emerge as a basic building block, it will be very interesting to see what kind of economics and architectural patterns arise. The containers, the playlists, where we assemble items will command an interesting new role in the assignment of context. And in this landscape, the item and the context are always already social, two-way and dynamic.

Comments closed

Real-Time Collaboration, Serious Play and the Enterprise

Published October 26, 2009 by cgerrish

With the advent of Windows 7 and the upgrades to the MS Office franchise, the talk is that there’ll be a big round of corporate upgrades. Many corporations are still running Windows XP, Internet Explorer 6.x and Office 2003 (or lower). Vista didn’t tempt them, but the good press for Windows 7 is supposed to do the trick. After all, they have to upgrade at some point, right?

If corporate America takes the plunge, one has to wonder if this will be the last upgrade cycle of this kind. The distribution and installation of software on to desktop and laptop computers is a messy business. Businesses require a very compelling reason to upgrade given the current model.

Google has put forward the model of the browser as operating system by working backwards from the Chrome browser to the Chrome OS. The integration of the Office Suite into the hardware starts in the cloud and moves to the local machine. When Microsoft tried a similar move in the other direction, the government stepped in.

Both Google and Microsoft have developed cloud-based Office Suite offerings moving from opposite directions. Looking down the road a bit, we can see that the next upgrade cycle will be “software + services” for Microsoft, and “services + software” for Google. The obvious motivation will be cloud-based software’s cost savings over the current model of distribution, installation, compatibility, upgrade and service of software installed on a local system. The sheer cost and pain of a firm-wide software upgrade is so frightening that most corporations defer it as long as possible. It’s entirely possible that some firms will skip the last installation and jump directly to the cloud.

Collaboration within the enterprise takes place via email, attached documents and shared network drives. The productivity software footprint defines the boundaries of the modes of collaboration. The big real-time innovation was the introduction of mobile push email via the Blackberry. This innovation reduced latency in the work process by detaching email from the desktop and allowing it to accompany a person wherever she might go. The introduction of Sharepoint and network-stored group editable documents is slowly seeping into the work process. But most corporate workers don’t know how to collaborate outside of the existing models of Microsoft’s Office products. Generally, this just an acceleration of the switch from production of hard copies to soft copies (typewriters to word processors). When confronted with Sharepoint, they view it as a new front-end to shared network drives, a different kind of filing cabinet.

Meanwhile in the so-called consumer space, Facebook, Twitter and a host of real-time social media services have radically reduced the latency of group communication and collaboration. In addition to text– photos, audio and video have begun to play an important role in this collaboration stream. For the most part the corporate computing environment has been left behind. This is due to two factors, the desire to maintain a certain kind of command and control of information construction and distribution within the walls of the corporation; and the desire of IT departments to avoid risk by maintaining a legacy architecture. The real-time productivity of the Blackberry has been working its way down from the top of organizations; but the tool set remains the word processor, powerpoint and excel. The only accelerant in the mix is faster mobile email of soft copies of documents.

Ray Ozzie discusses the “3 screens and a cloud” model as the pattern for the development of human-computer interactions across both the consumer and enterprise computing spaces. The missing element from this model is the input device, screens are no longer simply an interface for reading. Bits are moving in both directions, and email is being de-centered as the primary message carrier.

As we look at innovations like Yammer and Google Wave, the question becomes how will the corporate worker learn how to collaborate in real time? Accelerating network-stored documents and their transmittal via email moves the current model to near maximum efficiency. Further productivity gains will need to expand and change the model. Generally these kinds of innovations enter through the back door, or through a skunk works project, within small autonomous teams. But at some point, the bottom up innovation needs top down acceptance and support.

Luke Hohman of Enthiosys works with the concept of serious games in the management and development of software products. The collaboration processes he describes in his presentation to BayCHI may be the foundation for real-time collaboration throughout the enterprise.

Luke Hohmann on Serious Games: Luke Hohmann of Enthiosys on Serious Games

The lessons that we can take from Twitter and Facebook are that the leap to real-time collaboration is not one that requires a 4-year college degree and specialized training. It’s not an elite mode of interaction that needs to work its way down from the executive leadership team. It’s an increasingly ordinary mode of interaction that simply needs to be unleashed within the enterprise. But for that to happen, the enterprise will need to learn how to incorporate self-organizing activity. (Oh, and let employees use the video camera and microphone built in to their hardware) This will be a difficult move because the very foundation of the corporation itself is the creation and optimization of managed hierarchical organizational structures. It’s only when the activity of serious play can be reconciled with return on investment that the enterprise will come to terms with real-time collaboration.

One Comment