"Always bet on text" is a good philosophy that I *almost* buy, except for one thing:

I'm yet to find any reasonable plaintext representation of graphs. Text can do trees just fine, but trees are inherently limiting. If there's no efficient notation for DAGs and graphs with cycles - i.e. structures that can't trivially be flattened without duplicating nodes - then text itself is heavily limited as a tool for thought.

Anyone knows of any interesting work in this space?

@temporal would an incidence matrix count? Could be stored in plaintext


@kuba It works for capturing all the core information, but it's very space-inefficient as a notation - in most practical cases, you'd have extremely large and sparse matrices. This also means it doesn't yield itself for easy comprehension.

Attached is a screenshot illustrating the problem on trees.

For trees, we have a family of compact notations that communicate the structure clearly. I wonder if something similar is even possible for graphs in general.

· · Web · 1 · 0 · 0

@temporal i just stumbled upon this, didn't look far into that but might be useful for you:

@kuba Thanks. I've seen this before; I often use its "competitor", PlantUML. Related to this is Graphviz and its Dot format.

They do a good job at presenting graphs piece by piece, but they do this explicitly by listing every edge.

Attached is another screenshot that shows this notation applied to an example tree. It expresses the same information in roughly the same space, but doesn't help comprehension the way indentation-based notations do.

So, I wonder if we can do better for graphs in general.

Sign in to participate in the conversation
Mastodon for Tech Folks

This Mastodon instance is for people interested in technology. Discussions aren't limited to technology, because tech folks shouldn't be limited to technology either!