[[concept]] are my access points (purple). For example, right now I'm writing a note about John Grinder's model and added an access point for [[breathing]] which brought up my old note on reduction of fear. Along with it, it brought up other notes about oxytocin and dopamine.
Here's how my capture workflow looks like... I cite things from #note/capture into my #note/observation note as a block embed. At the same time, I create access points that connect me to old notes, in context of personal interpretation on what I've cited...
Example of my "Project" note. It's essentially like a MOC or workbench note which I use to track chronological change over time. When access points connect new ideas to this note, I update this project note
I probably can do better on metadata for future Dataview queries, but I like my stuff minimalistic. I only track date created and date updated. I use the "update time on edited" plugin.
• • •
Missing some Tweet in this thread? You can try to
force a refresh
A thought experiment. Maybe a good frictionless PKM is building to augment the primitive parts of our human brain.
We have, according to chronological order of development: 1. Hindbrain (rhombencephalon) 2. Midbrain (mesencephalon) 3. Forebrain (prosencephlon)
Also: 1. Neuro-endocrine systems 2. Neuro-vascular systems 3. Neurotransmitters 4. Limbic system
Hindbrain being the most ancient, containing the brainstem and cerebellum. The brainstem regulates our physiological processes automatically, such as breathing/heart-rate and so on. What I'm more interested in is the cerebellum
Starting a thread that will document my exploration into @Heptabase - mainly because it has a cost barrier. Let's see if it's worth the hype, because my expectations are raised when you need upfront costs.
@Heptabase Upfront, I really do not like the cost barrier - it forces you to chalk up a year's worth of "monthly subscriptions" and do not offer trial subscriptions. That leaves a bad taste in my mouth from the start. But of course, function comes first, so let's see if it's worth it.
@Heptabase I have to admit the onboarding tutorial/tour was really well thought out - it's pretty spectacular. It's probably one of the best I've been through so far - interactive, easy to follow, and has literally arrows pointing at where to click so I didn't feel lost at all.
Been analysing the process of recall. It's starting to seem that there is no distinct phase for input versus output. Whenever there is an input, the mind is already giving auto-outputs. It's a biofeedback system. Hence synthesis can only happen through feedback and appendage.
"Threading mode" came from analysing what our minds do when receiving six sense inputs; but it did not adequately dive into how our mind automatically feeds back and creates big works of synthesis by wrestling concepts on paper.
Seems that to really understand this, we need to dive into subconscious mechanics, beyond working memory.
Hope dataview gets up to its notion-like tables soon. Heard it was in the works... This would seriously take obsidian apart from other knowledge graph apps and make it serious competition with notion.
The best part about notion's databases is not that it is a simple spreadsheet or table, but that it literally is linked with existing pages with attributes/metadata. Relational databases introduce interactibility between databases. It's like no-code programming for the lay.
It's not just another feature. Imo every thought should be default be able to generate its own metadata - eg. Date/time. It should be able to have customisable metadata according to the input type - ie. an actionable would differ from a knowable.
When we want a sequential chain of thought, we nest them as equal level children under the same parent. The alternative would be a very ugly way:
Parent
- A
- B
- C
- ...
This way makes more sense but looks horrible.
The maths doesn't work out either. When use method 1, A/B/C are related via an OR (x) function. In method 2, A/B/C are related via an AND (%) function. We intuitively feel this burden that something is not right. And yet we don't like going into indenting hell.
Part of the reason why I enjoyed writing in prose form in obsidian was because I didn't have to figure out whether I had to indent or write a sibling bullet. Isn't it intuitive? No it isn't, especially when you consider how a bullet is queried in future...
It should be structured exactly how a thought works. One transient thought comes along, and is auto-tagged by datetime and custom metadata. Tags change its type - "actionable" vs "knowable". Like a tweet, you can chain another thought to this (thought train)
New thoughts can be assigned relationships to the previous thought - parent/sibling/child (up/sideways/down). Main idea is that you never break the flow of the current thought and stay in the present.