Unpopular #TfT / #PKM opinion: spacial representation of knowledge (whether 2D or 3D) is a complete red herring and, on the whole, a waste of time.
Graph views in @RoamResearch and @obsdmd and @AthensResearch and @logseq do not actually *encode knowledge* in any meaningful way.
What they symbolize, rather, is *progress.*
What they symbolize, rather, is *progress.*
It's very satisfying looking over a big graph overview and being reminded of all the connections you've made. It's motivating and encouraging. I don't think I would have "clicked" with @RoamResearch in the early days without some help visualizing what links and pages are *doing.*
When people share their expansive, color-coded graph overviews, they're showing off the work they've done. The sense of progression and accomplishment is what makes those feel meaningful.
But when someone shows you a galaxy-looking graph overview with 15,000+ nodes, your reaction is NOT "Ah, yes, I see what you mean."
Graph overviews do not, by themselves, *teach* you anything. Most of the actual information in the graph is lost by representing it all spacially. And whatever small amount of information it does retain is not readily digestible to someone unfamiliar with the underlying entities.
The actual knowledge work happens in textual form. The written word is primary.
A graph overview spacially *reflects* the *true* format that stores knowledge. It does not itself give a true (a.k.a. "full" + "faithful") representation of knowledge qua knowledge.
A graph overview spacially *reflects* the *true* format that stores knowledge. It does not itself give a true (a.k.a. "full" + "faithful") representation of knowledge qua knowledge.
Generating a graph over a note network must, by definition, throw away +90% of the actual information in the notes. Spacial representations are inherently limited in what they can productively convey.
Important clarification that I'm realizing from reading replies: In this tweet I intended "representation of Knowledge" in a specific sense—namely, representation of *complex, interconnected knowledge networks.*
• • •
Missing some Tweet in this thread? You can try to
force a refresh
