Awakened grains of sand

Many people assume that web thinking comes naturally to anyone born after 1990. Yes and no. Yes, the experience of 21st-century digital natives is qualitatively unlike what all prior generations shared, as far back as the dawn of genus homo over a million years ago. But no, the virtual dimension hasn’t fundamentally changed us — at least not yet.

Consider the physics of things in the real world. As hominids we have a million years of experience with the properties and behavior of physical things. And as individuals we have lifetimes of such experience. Based on those experiences, here are some things we know:

  • If I give you a thing then it’s yours until you give it back. We can’t both use it.

  • If I want to make a collection of things, I get a bucket and put things into it. A thing is either in the bucket or it isn’t. The same thing can’t be in two buckets at the same time.

  • If I invite you to work on a project using my collection of things, you have to visit my house.

In the virtual dimension none of these laws apply.

  • I can give you a link to a thing, and we can both use it. If either of us improves it, we both share the improvement.

  • If I want to make a collection of things, I can invent a tag that identifies things in a collection. We can invent many tags for many collections. A thing can be in more than one collection.

  • If I invite you to work on a project using my collection of things, you can join me in a virtual place.

As we begin to colonize the virtual dimension, we begin to learn about the physics of virtual things. But we still haven’t fully internalized the laws and properties that govern them. All of us, digital natives included, still tend to treat virtual things as if they were physical things.

Making and sharing collections of things is a common scenario that illustrates the tension between these two different sets of laws. When the things are physical objects — say, coins — there’s no choice about how to collect them.

Scenario 1: Somebody passes around a bucket, people toss coins into the bucket, whoever holds the bucket has the coins.

But when the things are virtual objects — say, chunks of information, existing as web resources, identified by URLs — there is a choice. Somebody can still pass around a virtual bucket, people can still toss virtual coins into it, and whoever holds the virtual bucket will have access to the collection of coins. That’s exactly what happens when we use an email conversation as the virtual bucket. Or:

Scenario 2: We can invent a name for a virtual bucket, and toss web resources into it by tagging them with that name. Now anyone who searches for the tag will have access to the collection of resources.

This new scenario raises important questions. How do we choose a name for the collection? How do we choose a service that binds the name to things in the collection? How do we restrict access to the collection? On what terms are these name-binding and access-control services provided to us?

I can’t answer all these questions at once, so let’s start with the most fundamental one: How can we name things?

In A tale of two dams I made up a tag, WestStDamKeene, that could be used by anyone to co-ordinate web resources related to an ongoing conversation in my city about the fate of the Ashuelot River dam on West Street. The tag is 14 characters long. With 14 characters you can make quite a lot of names. If each character can be a letter or a digit, then there are 36 choices for each character — 26 letters plus 10 digits. There are 3614 strings of 14 such characters. When faced with big numbers like that I turn to my favorite scientific calculator, WolframAlpha. It evaluates the number like so:

In a recent talk I failed (spectacularly) to convey the point I’m about to make, so I’ll try it again and more carefully here. We can make about as many 14-character tags as there are grains of sand on Earth. True, a lot of those won’t be nice mnemonic names like WestStDamKeene, instead they’ll look like good strong unguessable passwords. But there are still unimaginably many mnemonic names to be found in this vast namespace. Each of those can serve as a virtual bucket that we can use to make and share collections of arbitrarily many web resources.

The implications take a while to sink in. Grains of sand are inert physical objects. They just lie around; we can’t do much with them. But names can be activated. I can create a 14-character name today — actually I just did: WestStDamKeene — that won’t be found if you search for it today on Google or Bing. But soon you will be able to find at least one hit for the term. At first the essay I’m now typing will be the only hit from among the 30 billion indexed by Google and 11 billion indexed by Bing. But if others use the same term in documents they post to the web, then those documents will join this one to form a WestStDamKeene cluster.

This isn’t even tagging in the conventional sense. Set aside, for now, the clusters that can form around that tag on Delicious, or on WordPress, or on Flickr, or on Twitter, or even at the intersection of all those tag-oriented services. Consider only what can happen when you make up a never-before-seen term, utter it in a document placed on the web, and invite others to utter it in documents that they place on the web. When we enact this scenario we are, in effect, creating an ad hoc web service that we can use to make and share collections of things related to (in this case) a particular dam project in a particular town. Of course tag-oriented services enhance our ability to make and share such collections. But even without them we can still do it. In the realm of virtual things, names are as plentiful as grains of sand. But they aren’t inert. We can wake them up and use them to co-ordinate our activities.

21 thoughts on “Awakened grains of sand

  1. Jon, the reason I’ve made an effort to follow all of your published work since Byte Magazine days is exemplified in this post. Deep thinking, leading to novel and pragmatic insights, beautifully explained. Well done!

    Hope the new Delicious folks are paying attention.

    (I’m amused that the category for this post is “Uncategorized.”)

    1. Patrick, I agree, this is a very insightful post, and exemplifies why I continue to follow Jon as well.

      I’m looking forward to the answers to the other questions, i.e. “How do we choose a service that binds the name to things in the collection? How do we restrict access to the collection? On what terms are these name-binding and access-control services provided to us? “

  2. Thanks Patrick! Yeah, that is funny. I had actually written the tag (as a WP tag, not a WP category) so this post would show up in the various queries I linked to. But then I forgot to commit the version with the tag.

  3. Strictly speaking the old metaphors are correct. A document tagged in two ways does not actually appear in two tag folders at once. It is actually several unique lines in a relational database table that are never duplicated. Instead, instances are create by the software or web server and rendered once each by a browser or desktop GUI. Changes from these new copies are written to the single disk location of the file. The illusion is that a copy is the original, and thanks to the speed of computers it is convincing, but what is happening no different than 300 philosophers across Europe and North Africa copying, classifying and commenting on Aristotle in the 13th century, just more common now…

    1. I think there are several qualitative differences between what those 13th-century philosophers could do and what we can do. For example, they could not:

      – designate a master copy of an Aristotle work

      – identify it with a globally unique name

      – collaborate, in a shared information space, on the exegesis of that Aristotle work

      – do that collaboration indirectly by way of links

  4. Maybe I’m overshooting, but I’d bring it all the way back to the primary source here John; Tim Berners-Lee’s “URI – Axioms of Web Architecture”:

    “The Web is a universal information space. It is a space in the sense that things in it have an address. The “addresses”, “names”, or as we call them here identifiers, are the subject of this article. They are called Universal Resource Identifiers (URIs).”

    Anything which cannot be named and collected, anything which cannot be linked, is not of the web. The supplementary tagging services you name constitute a more human-focused sub-address space of the web.

    1. You’re right of course. I’m focusing on the tag-based address space here precisely because it is human-focused. People often can’t control the URIs that represent their web activity, so the opportunities that flow from intentional naming aren’t apparent. With tags it’s clear: You can make up any name, you can invest it with any meaning you intend, you can agree to share that meaning with others and thereby collaborate in that context.

      It’s worth noting that while tags are indeed “just a subset” of web namespace, the indexed web — on the order of 10^11 URIs — is negligible by comparison to the 10^21 permutations of a 14-character tag.

  5. At, says:

    The whole article is based on three premises about human cognition. All three are false. What is clear to me is that many people who engage in “web thinking” as the author calls it have an absurdly caricaturish notion of human thought and interaction before the advent of the internet.

    And here’s why they’re false:

    1. Human beings have had shared spaces and things such as temples, communal dwellings, canals, bridges, roads etc. that many individuals could and did contribute to and improve, and in many cases the whole community could use simultaneously. The internet is not a radical departure in this regard – it’s yet another example of the kind of social construct that’s as old as our species.

    Temples, canals, and bridges are examples of infrastructure that’s created and used collectively. Their shared creation and use requires co-location.

    A web-based software project is another example of infrastructure that’s created and used collectively. In this case the shared creation and use doesn’t require co-location.

    We’re still sorting out the dynamics of virtual teams vis-a-vis co-located teams.

    To say that the rules governing virtual teamwork may be qualitatively different from those governing co-located teamwork in no way challenges the notion that humans always have and always will work in teams. It only suggests that with the advent of the web, differences in how we work in teams may be more than differences of degree. They might be — I think they are — differences in kind.

    2. So a cardinal can’t simultaneously be a member of the category “things that are red” and the category “bird” and the category “animal” – at least that’s what the article claims.

    Here I’m talking specifically about a common group task: making a list of things. Real-world physics dictate that the list is written by one person, to whom others suggest items for inclusion. Web physics enable a loosely-coupled and decentralized approach that is qualitatively new and, for the most part, not yet exploited.

  6. What you are describing is content addressing vs location addressing. But this is not a new thing from a URI perspective. The ‘http’ for URLs is so pervasive that other protocols are rarely considered. The design though totally expected protocols like ‘tag’ to be used. There’s a world of difference from an automation perspective between WestStDamKeene and WestStDamKeene. More significant though is RDFa which makes something like WestStDamKeene machine readable.

    1. Hmm, yes. Last bit again with a little bracket fuzzing.

      There’s a world of difference from an automation perspective between [a href=’tag:WestStDamKeene’]WestStDamKeene[/a] and [span]WestStDamKeene[/span]. More significant though is RDFa which makes something like [span class=’tag:def’]WestStDamKeene[/span] machine readable.

  7. There’s a world of difference from an automation perspective between ay-href-WestStDamKeene and WestStDamKeene.

    Agreed. I’ve always championed the ay-href world and always will. I’m also increasingly a champion of the non-ay-href world, which I see as complementary. If you and I and some others agree to use WestStDamKeene for a mutual purpose, then we can do quite a lot with little or no deep software support, just an informal contract and the dynamics of the web.


  8. This comment is at best tangentially related. One of the problems with these really large numbers is visualizing them in any meaningful way. I really liked the Radiation Dose Chart in this post. I think it does a great job helping compare numbers that may have wildly different scales.

  9. In the self-centered network, my friends are important people; from my departure, after a few connections can cover a large crowd. In social networks, everyone Center.

    TED U.S. media organizations responsible for innovation and Chris Anderson proposed a new concept of “mass to accelerate innovation.” According to this concept, the innovator is not to hurry to hide his good ideas, but to boldly announced, so that others involved in improving, and he continues to progress, eventually everyone will become a winner!

    American business strategy expert Michael Porter in 2011 the first issue of “Harvard Business Review” on the proposed capitalist crisis, “creating shared value” should be a new target for business success. “Creating Shared Value” is the Chinese people have the most to understand, but also Chinese enterprises are now sorely lacking!

    Network Research Center of Northwestern University’s Professor Albert-Laszlo Barabasi said, the outbreak of social network communication compliance laws, namely the occurrence of extremely rapid dissemination of information, unpredictability, and culminate in a short time, but the transmission cycle is very short. The level of the spread of the first five minutes has been determined.

    Professor of medical sociology at Harvard Medical School study indicated that Nicholes A. Christakis, a friend of human behavior by high-impact, but this effect only to the third, fourth little. To obesity, for example, once friends of fat people obesity risk increased by 45%, second 25%, 10% three, four times 2%. Beware of your friends behind!

    Stanford University doctoral Xiaolin Shi at the University of Michigan PhD thesis issued that in all major social networks are important there is a node composed by the small network, and has the following characteristics: First, there is an important node between high correlation, to form a small network; two, relatively small network can accurately predict the characteristics of a large network; three important nodes more durable, more rapid growth of new connections.

    University of Michigan professor Lada A. Adamic and other people’s research shows that social games: the game by friends to invite participants to join a higher viscosity and easier to invite his friends to join. The key to success was not invited to spread a large net, but continued for a few friends an invitation.

    University of Miami Neil F. Johnson study found that community and the entrant is a relationship between two-way choice: the relative lack of community will choose their own talents, entrants will be selected so that people do not yet have their own societies to join, it is easy to lasting relationships.

    Men and women
    Harvard Business School MBA Bill Heil wrote in 2009, on Twitter, although men and women the same amount of powder, but a fan of men more than women 15%; and general social networking is different from men are more active. Posts of the average for all users 1, the top 10% active users made 90% of the posts.

    Harvard Miko · aj Jan Piskorski study showed that: all kinds of business and social networking will be the true wishes of the people as a cover. For example, in commercial networks, who will have regular work to expand business opportunities as an excuse to quit looking for the opportunity.

    The famous Professor MEJ Newman Santa Fe Institute in 2001, proposed a “self-centered network” concept: self-centered network, my friends are important people; from my departure, after a few connections on the can cover a large crowd. In social networks, everyone Center

Leave a Reply