The New Gatekeepers, Part 7: Solutions

Media | Access/Network
This is an addendum to the New Gatekeepers series (formally, part 7). In the series, particularly part 4, I described that the need for gatekeepers is a result of discursive postings; in order to minimize the influence of gatekeepers, we need aggregatable declarations.


“Aggregatable declarations” encompasses such activities as filling out a ballot and voting, whether for political office or for American Idol; participating in a survey; rating a movie on the Internet Movie Database; writing a review for a products review site like Epinions or Amazon or Zagat; sign a petition; participate in a political rally; buy stock. In all these cases, people take a declarative action which can then be summed up to form some aggregate picture of how many total people are making that statement. It is sort of arbitrary that the goal for such aggregates may be 51% or 2/3 or 4 stars or 5,000 signatures or $70/share, but those are concrete enough goals for advocates to strive for (or against).

Aggregatable declarations are crucial for markets and democracies. They not only help decision-makers compile information (this is somtimes called “data-mining”). More importantly, they can allow participants to find people with whom they share opinions.

What’s unfortunate is that so much of the communications essential to both democracy and markets escapes aggregation — letters to the editor; letters to elected officials; Internet forums; blog posts (elected officials do internally track “for” or “against” certain issues, and publications often do as well, but it is not openly done). And it’s a shame because they could easily be aggregated in an open manner.

Beyond the examples listed above, there have been few practical deployments for rating the world of ideas, particularly those ideas which are present on the Internet. There is the Google PageRank, and there is the Technorati rank, which is a straight count of links to a web page. Both are essentially popularity-ranking systems. By linking to a page, you are implicitly endorsing its popularity, often without any cognizance of that. This is a chief flaw in the system (see Seth Finkelstein’s Jew Watch, Google, and Search Engine Optimization), and is part of the reason that Google is believed to be at work on a replacement called TrustRank (Seth, recently, perplexed at Jeff Jarvis’s mistrust of TrustRank: “Such algorithms are the missing piece of building a journalism data-mining business… Moreover, something useful would be the best thing ever to happen to ‘citizen journalism’!”) In the meantime, both Google and Technorati embraced a way to allow hyperlinks to be specified without implying an endorsement, with the nofollow attribute. Neither has commented on the rate of adoption of this approach since the beginning of this year.

There’s a more direct way of solving the endorsement problem: separate endorsements from links. Endorsements are a type of aggregatable declaration, which this point in the essay is begging for a shorthand term. Suppose we call it “peg” to match the dictionary meanings of making a mark, of classifying something. Thus an Internet user can choose to spend their time blogging or pegging.

Now pegging would not seem to be socially rewarding in a way that blogging is, but thirty million “American Idol” viewers would disagree. In theory, it is much easier for the average person to register a simple declaration than to churn out a short blurb to make a blog post. The only thing that’s missing is the conception of pegging as a cohesive activity, and specific examples of how it would solve real-world problems in information management. As I mentioned in part 6, I use journalism simply because it is a very strong model for this.

Problem: How can we get an appraisal of a given article– so that the publisher, and the readers, can be aware of how it has been received by its audience?

Current Approaches: You have to hope that people leave comments; or that the editor publishes all letters in response. Beyond a certain number of comments, an aggregated summary is absolutely necessary. There are also third-party tools whch service now use a binary ranking (Google, StumbleUpon), and others use a 1-5 rating system (Furl). These approaches are not well integrated into any site or index; Google does not openly display the number of votes a page gets.

Proposed Solution: ViewPoints is a palette of responses for deliberations. Instead of asking the reader to rate content on a single scale, which is just mindless, the ViewPoints provide a list of choices which actively engage the reader. Do you agree? disagree? Feel the conclusion was unjustified? Find the post off-topic? There are seventeen in all; see below.

Consider Also: The ViewPoints is not just a way to react to a primary source. It can be used to react to comments as well. In theory, a primary source would attract a number of critics, and then the rest of the readers, instead of passively lurking, would register their ViewPoints with the various responses accordingly. It would help to shape an online conversation.

Implementation: I conceived of it four years ago, and implemented as a custom module for this site. I will be making this into a standard Drupal module.

Problem: How can readers best “triangulate” writers/publishers, that is, to judge a current article in the context of what else they’ve written?

Current Approaches: The reader has to comb through a lot of articles. Who has the time to do this?

Proposed Solution: the HearsayNetwork. Person X says or thinks that person Y says or thinks something about Z. That’s what hearsay is, by definition. We can invite Y to state what they say about Z; and they can enter different opinions over time.

Consider Also: These can leverage the ViewPoints, to allow people to rate other people’s opinions.

Implementation: I’ve designed it; I wish I had the time to build such a system.

Problem: How does a publisher solicit tips from readers in an open manner?

Current Approaches: Readers may suggest tips in the comments, but there is no consistent expectation of whether the publisher will choose to look at that them or devote a new article on them. Publishers solicit email, but that remains private and closed.

Proposed Solution: the Tipster Network. Create a structured form for submitting links and short ideas to the publisher. These can be reviewed openly, and pegged by other readers.

Consider Also: Blogs were originally designed to organize links to other things on the Internet. Now suppose that links are supplied, evaluated and filtered by the users. What value is left for the publisher to do in the department of posting links? The closer the relationship between the readers and the publisher, the more that the suggested links would mirror what the publisher was going to post.

Implementation: I suggested that Joshua Schachter’s service be augmented to do this; quite simply one can add a bookmark with accompanying comment, and specify as one of the tags the target publication.

What’s ironic is that I shared the Hearsay Network; Dan Gillmor made a note of it. The idea of the tipster network came out of a direct response to Jay Rosen who gave a list of guidelines for submitting links to his PressThink blog. That Rosen and Gillmor didn’t pursue these ideas further, I can only speculate: they perhaps get many suggestions (but we have no idea, without the accountability of Tipster). But what’s odd is that the readers of Gillmor and Rosen didn’t.

The wry reader may wonder whether this hearby proves or disproves the whole theory of gatekeepers. Certainly I realize that I need to spend time reaching the right audience– fellow software engineers.

Will such a system of aggregating declarations ultimately serve as a suitable replacement for gatekeepers? As a faster way of doing editors do? Perhaps; there’s one way to find out, and that’s to get these software tools developed.

  • Read the Posting Guidelines
  • ViewPoints
  • Login/register to post
  • Response summary: 1 comments, 0 Viewpoints
    aggregrate ranking, hmmmm Allison_Nevitt Mar 18 ’06 5:45AM
    . for content Jon Garfunkel Aug 06 ’06 1:07PM