Trailmeme and the Web of Intent


The quickened pace of content production and distribution have created a stream of information that we have trouble focusing our attention on, much less our intention. Establishing (indeed, re-establishing) a true Web of Intent will require different interaction with the stream, one that insists on active participation and creative work, rather than passive consumption.

A Web of Intent is a participatory web: a web where we’re active consumers of content. Webs of Intent are an extension of the remix culture that has emerged.

Intention is actually taking action, expending some energy or effort to do something. Intention is a lot more expensive, cognitively speaking, than merely attending. The power of collective intention is literally what changes the world, but we don’t have the tools to direct it yet.

We need intent-centric products and services that contextualize the stream, and propel publishing in a more meaningful and actionable form.

One such product that addresses that need exactly is Trailmeme, from Xerox (s XRX), which is currently in private beta.

Trailmeme is still in very early development and the UI is rough, but I’ve been following its progress with interest since I first spoke to the project manager, Venkatesh Rao, over two years ago.

Trailmeme is a conceptual necessity to the organization and creation of Web content in the endless flux of the stream. I feel it is evolving a new and valuable breed of content curation and creation.

I’m not allowed to say too much about Trailmeme, as it’s still under wraps, but I do want to give a few high-level observations that get at why the project is so interesting.

Trailmeme allows users to save and tag articles, but not just to consume and store them. Users “blaze” their own trails by creating a meaningful, intent-driven, interlinked network between pieces of web content. The result are annotations and contextualizations that are both personally relevant and, when shared and made public, a new way of re-engaging the material for readers and followers of the trail itself.

Users link digital objects (articles, websites, datasets, etc) in multiple relational paths, and provide their own commentary to the associations.

For example, the Trail on the “Scamville” expose launched last year by TechCrunch begins with a single node, or trail marker, and branches out in any number of creator- or collaboratively-defined paths to follow the story across various websites and articles.

For readers, these artifacts can be followed as an in-frame series of articles, or from a visual map view that displays all the possible paths.

The visualization element is one that will become increasingly indispensable to the web as streams are organized (ManyEyes and Tableau are particularly interesting).

We have a long way to go before the Web of Intent takes hold, but the ideas behind Trailmeme and others are fundamental to addressing some of our most pressing problems.

Hopefully, more systems like this will force an intentional engagement, one where passivity and reception proves not only unrewarding, but antithetical to the system, much in the same way that video games have elevated the interactivity of film and television.

The first step to ensuring intention is to promote systems that are inherently based on engagement and action, not passive consumption (e.g., more clients upon clients).

If we are to evolve a more manageable and effective web, we must remember that capturing attention is not the only ambition we must pursue; we must focus and foster intention as well.

Nova Spivack is a technology futurist, serial entrepreneur, and one of the leading voices on the next-generation of search, social media, and the Web. He is a co-founder of LiveMatrix and an active angel investor.



Nova Spivack spun that the next big thing was Twine and that was a big fail so he has credibility problems.

Trailmeme should say, “Thanks but no thanks” to his input.

Jack Smith

Thanks, Nova. Very interesting look at intent. The idea of visualizing relationships between things is an important one and it’s a better way to quickly get at the information we’re looking for.
“Establishing (indeed, re-establishing) a true Web of Intent will require different interaction with the stream, one that insists on active participation and creative work, rather than passive consumption.” In the case of social media, there are a few ways that intent can be determined. I agree with henchan’s comment that this will not scale. In social media, the content of a written status update or forum post generally contains strong signals of intent. But we shouldn’t assume that we can change the behavior of individuals and get everyone to participate unless there’s tremendous value for them in doing so. Far more people in social media online are lurkers than are participants. In the case of lurkers, a combination of passive and existing active signals can be used to divine intent without trying to change their behavior. The fact that a non participant arrives at a forum thread via a search engine, clicks on a link, or visits related conversations are all signals of intent albeit signals of varying strength. Combining signals from existing behaviors in the right way can achieve a Web of Intent with no heavy lifting.

-jack smith


An interesting article. Thanks.
“The first step to ensuring intention is to promote systems that are inherently based on engagement and action, not passive consumption ”
I generally disagree with this point. Though it may be applicable in a niche, it will not scale. The job of systems designers should be to automatically convert passive consumption into information of value to its users.
When the distinction between consumption and publishing is sufficiently diminished the perception of ‘hard work’ – one of the barriers to system adoption – is lowered.
Video games are a poor analogy. They give the appearance of intent, but this is achievable only because their content, UI, game objectives, narrative etc. are all constrained through top-down design. As a matter of practicality, Web-scale systems need to depend on the system’s emergent properties rather than on creative control.
I don’t know about Trailmeme, so I could only guess at how it works. But consider ant trails: a classical example of emergence. Each ant is usually both walking (consuming) and blazing (publishing) a trail simultaneously. Very rarely it will depart from all known trails and at that point it is only blazing. There is valuable information for the colony from each ant walking the trail.

Comments are closed.