One of the many things I love about Semble and Margin is that they offer something more than a way to save links on the web. They provide novel ways to connect ideas and build knowledge together.
Someone discovers an article and adds it to a collection. Someone else finds it, connects it to another source, and adds a note explaining why the relationship matters. Another person challenges that interpretation, offers a different perspective, or makes a connection nobody had considered.
This dialogue continues in a mesh of annotations, collections, threads, and on and on, all across the Atmosphere and wider web.
A link is just the starting point. Annotations become invitations to a discussion. Connections becomes arguments. Collections become part of a shared and ever-evolving web of connected ideas, evidence, interpretations and relationships (both human and informational).
Forest for the trees
Semble and Margin excite me because they make room for the social life of information. What we find, what we think about it, and how our thinking changes through encounters and exchanges with other people.
These tools give us ways to leave traces of curiosity, to make our reasoning visible, and to weave together the intellectual work of others into something more.
But as these collections and conversations grow, stepping back to see the whole picture becomes harder.
We can explore individual cards and follow connections, read annotations and participate in discussions. But how do we see a larger context taking shape?
Which ideas cluster together? Where are the strongest relationships? Which sources remain isolated? Where do different people's interests overlap? What unexpected bridges might appear between collections, arguments and conversations that have developed independently?
Creating a purview
Meet Assemblage.
Assemblage turns Semble collections and Margin annotations into an visual knowledge graph. Links and annotations become nodes, relationships become visible connections, and the collection becomes a landscape to explore. You can move cards around, trace relationships and draw new connections directly on the canvas.
The idea is to offer a different way of seeing what we're already building together.
For this first version, I'm working with the links in my own repository as a deliberately small (and manageable) starting point.
This first iteration is an opportunity to experiment with how collections and their relationships might be explored spatially, rather than navigating from page to page, from one record to another.
I'm curious about what becomes possible when we can see these relationships laid out in front of us.
Can a visual map reveal patterns that are difficult to notice in a list? Can it help us find gaps in an argument, connect seemingly unrelated ideas, or discover a new direction for research? Could it make the relationships between people's interests and ideas more apparent, helping us understand not just a collection, but the conversations and shared enquiries growing around it?
And what happens when the map isn't just something we look at, but something we can directly manipulate and extend, drawing new connections, and adding more layers to the context and knowledge we're building?
Over the horizon
There's a bigger ambition here.
I'd love to map every link and relationship across Semble and Margin, and eventually work with others to visualise the knowledge graphs emerging from their collections.
Imagine being able to explore the connections within your own research, discover overlaps with someone else's collection, and trace an idea as it travels between sources, annotations, people and conversations.
Not a definitive map of what everyone knows, and certainly not a replacement for the conversations themselves. More like a way to make some of the richness already there easier to see and to create new opportunities for people to connect.
The most interesting thing about a knowledge graph may not be the nodes or the edges. It's what happens between them. The ideas that travel. The interpretations that change. The connections that only become possible when different people bring their perspectives together.
How much of the world do we need to represent before the representation becomes useful? Perhaps the answer isn't to make the map bigger and bigger, but to find new ways of seeing what's already there.
This first version is a small step towards that possibility. There's plenty left to figure out, but I'm excited to see what becomes visible when we can step back and look at the whole map.
—
Try Assemblage
Contribute
If you want to follow the progress with Assemblage, follow me on Bluesky or Mu, or subscribe to this blog.