Thoughts on the Antinet's manual search
Nicolas Gatien: @ArchLeucoryx Any thoughts? Based on your twitter posts, I believe you are working with Logseq and Obsidian?
First--thank you for taking the time to write the thread. Scott discussed this in one of his videos but finding specific info is a bit harder ;) I think plenty of people are worried about "blinding" themselves to their notes without full-text search.
I use Logseq and Obsidian casually, among other applications like Dendron on VSCode, FSNotes, and even Twitter. Currently, I'm on a "null theory" of PKM, and I'm investigating the ways we represent our knowledge--literature, journaling, commonplacing, note-taking, oratory, communication, etc. I don't think there are any rules to knowledge! All representations are mutable and interchangeable; your notes are not your knowledge; the map is not the territory. But that doesn't mean we can't be open to better maps that give us new perspectives on the old.
(Note: I don't have an Antinet myself, but I think I have a good picture of it. I tried making an alt-Zettelkasten in TiddlyWiki that used sequences of sequences rather than trees of trees, and timestamping instead of alphanumeric IDs.)
The lack of direct search is definitely not a bug. The Antinet still uses a hand-curated index, so you're not lost every time you open your drawer. The manual search is really just something you do, right? It's practical usage. At the very least, the process of re-reading is an aid to memory; recall in of itself makes memories stronger.
I'm reminded of this other application, ThreadHelper...it was a sidebar extension to Twitter designed to be a serendipity engine. Every time you wrote in your compose box, the sidebar would bring up related tweets from your account and your bookmarks (you had to import your Twitter data locally, I think). And one of the developers, exGenesis, explains that "[your] body of work should be a living system but water is drying up nearly as soon as it's posted". I can't help but think that the process of re-reading your notes is like "watering" them again, digital or no. You're coming back to your past self and following up with their thinking. This way, your thinking literally compounds over time.
Meanwhile, the Zettelkasten falters a bit at explicit recall, at least later on. An Antinet small enough that you can keep it's contents in your head and reliably put in and take out the exact card is an archive. And explicit recall does make sense! Sometimes you need the full fidelity of either your own or someone else's writing. But it's kind of interesting that we haven't really investigated the shift from explicit recall to associative recall in the slipbox, as the slipbox grows.
Further, the key to associativity seems to be surprise and discovery. Eva Thomas writes in her article that Luhmann described the "surprises we encounter while interacting with the slip-box [...] as a process of finding something you were not looking for, but something that you can still use". (Unfortunately I don't know what quote that is in his actual paper, Communication with Slipboxes.) Is a mature Antinet a discovery engine just by reading through the notes and walking down the associative trails to find what you need? That seems too good to be true. But it also makes a lot of sense. What happens when the user of the Zettelkasten begins to forget what they have written? Well, the more the user forgets, the more surprised they are when re-reading their writing. Forgetting is integral to serendipity; a creative partner. At this stage, you put a note in there to "lose control of it, not to gain control over it".
To be surprised, we must be unaware.
To be surprised of known information, we must make it unknown. Defamiliarize; misplace; forget. It is only when we know we have lost something that we wish to get it back, to reclaim it more tightly.