(hidden) n212LBQqE

words:	0 + 922 = 922
last update:

Linked Mentions (1 in 1 page)

January 4th, 2022 (1)
- [[(hidden) n212LBQqE]]
  - https://roamresearch.com/#[[(hidden) G97ohL3l8]]/app/developer-documentation/page/tIaOPdXCj
  - roamAlphaAPI resources
    - https://roamresearch.com/#[[(hidden) G97ohL3l8]]/app/developer-documentation/page/K_unTEjJx
      - (!) https://www.zsolt.blog/2021/01/Roam-Data-Structure-Query.html
        
        insanely good.
        
        {{[[TODO]]}} need to go through all material linked there
        
        further - "learn [[(hidden) 8iobLNJvy]]'s [[(hidden) xH-Ph8S9m]] today":
        
        http://www.learndatalogtoday.org/
        
        also "datomic datalog" in 7min
        
        https://www.youtube.com/watch?v=bAilFQdaiHk&feature=emb_title
        
        further
        
        https://roamresearch.com/#[[(hidden) G97ohL3l8]]/app/Zsolt-Blog/page/WUn5PuTDV
        
        further
        
        https://twitter.com/Roamfu/status/1363681695057121281
        
        further
        
        https://twitter.com/VladyslavSitalo/status/1362261990547197954
        
        https://twitter.com/VladyslavSitalo/status/1415296623693942791
        
        https://github.com/mathieudutour/gatsby-digital-garden/
        
        #observation people have been doing [[export-html-pages]]-like stuff for a while already!
        
        [[(hidden) zawAAHB3b]]
        
        https://github.com/mathieudutour/gatsby-digital-garden/issues/31#issuecomment-752807330
        
        https://github.com/roam-garden/roam-export/tree/757a777aac27777021f3ebab3c073893d7884951
        
        https://github.com/nikvdp/roam2agora
        
        {{[[TODO]]}} checkout [[(hidden) jZNl6HGob]]
        
        https://anagora.org/index
        
        [[(hidden) X3vwrXyD7]]
        
        https://github.com/dvargas92495/generate-roam-site-action
        
        #observation oh snap, they're doing everything __directly in [[roam]]__, and then using [[lambda functions]] ([[(hidden) g2-_urXV9]]/[[(hidden) _Ekrn0YvL]] or whatever)
        
        further https://github.com/dvargas92495/generate-roam-site/blob/main/src/index.ts
        
        further https://github.com/dvargas92495/roamjs-components
        
        [[(hidden) zJw7j5jxr]] holy shit, so much thoughts coming from this
        
        (hidden) b7GCT2svs:: "The Killers - Mr. Brightside (Lyrics)" ("it was just a kiss")
        
        https://youtu.be/UFKf08us2AI
        
        [[roam-traverse-graph]] [[export-html-pages]]
        
        thinking about how you can e.g. do in [[roam]]:
        
```javascript - further: - link - link```

vs ```javascript - further: link - further: link```

and how you could do that with linked references, if you were doing it:

```javascript - [[further]]: - link - link```
        
        vs ```javascript - [[further]]: link - [[further]]: link```
        
        {{[[TODO]]}} would [[roam attributes]] work better for this?
        
        i.e., you'd do the mentioning twice probably, like in the second part of either examples
        
        thought is that necessary? i think it's irrelevant actually
        
        initial thought - no, because the `(hidden) G97ohL3l8::` is in the end, and you'd have to manually write out the page's title instead of being able to search
        
        but wait, i never really use attributes, and that's not right! you can do `(hidden) G97ohL3l8::` and roam throws you in the beginning, and you can search just fine!
        
        okay so then that's no longer a problem
        
        next up - is this useful
        
        re:: ((EfE5J1OtG))
        
        why i thought it'd be useful is that you can only have one attribute per block. thus it'd force you to have more specific (verbose?) page titles instead of combining multiple ones:
        
        e.g. instead of [[roam-traverse-graph]] [[export-html-pages]], i could have a page titled "export-html-pages-plugin-of-roam-traverse-graph" or smthn, i don't yet know what'd be the optimal naming convention, but we could try to figure it out.
        
        i've even seen some people on [[roam]]'s [[(hidden) VZrbM9KJQ]] talk about some conventions
        
        but the main question here is -- is this overall a good thing?
        
        {{[[TODO]]}} #observation am i using attributes wrong?
        
        {{[[TODO]]}} more complex attribute handling in [[roam-traverse-graph]]
        
```javascript {{[[TODO]]}} ~~observation::~~ #observation am i using attributes wrong?```

or maybe like this? XD

observation:: {{[[TODO]]}} am i using attributes wrong?

{{[[TODO]]}} would need some extra integration help w/ the [[TODO]] toggle from [[roam]]

{{[[TODO]]}} figure out what [[roam attributes]] are all about / what their purpose is

[again](((vssQKBfFO))), the learning material of datalog could help

{{[[TODO]]}} thought also interesting what the [[roam]] people intended with them

cc:: [[(hidden) 1h9jVJvM-]]

{{[[TODO]]}} #[[(hidden) jOMmlCP4S]] if [[roam]] could provide a hooks/events api that you could listen from,

then the stuff that people are doing directly on roam, and we do in our runner locally / on github actions with [[roam-traverse-graph]] -- we could do all that in a remote environment as well, __but with much faster time intervals__ - they wouldn't even be intervals anymore, rather - reactions to events [1] happening live, as the user acts.

[1] in [[(hidden) xH-Ph8S9m]], it's [[(hidden) Vr3N0zG5Q]] isn't it?

(hidden) AV7J4zqkD:: ((YG-4tevRA))

this would likely require [[(hidden) BislYexyN]]s - detecting if changes have side effects that impact not only the edited block, but the whole page, or even the whole graph / parts of it

(hidden) AV7J4zqkD:: linking to a page - both pages need updating now

and if it's a meta page / has powers, then extra works to potentially any pages in the graph.

i.e. annoying stuff!

idk, maybe if we manage to keep the stuff fast, like it is now [1], then going thru the whole graph shouldn't be as bad?

e.g. we could have a queue of incoming events, and we'd only do 1 at a time, and if a new one comes in too fast, we'd wait for the currently one to complete, and in the meantime if more come in, we'd discard all the new ones except the newest, and then could go again.

[1] it takes ~5 seconds? to go from a `notes-private/graph.json` into a `notes-public/*.html`

and probably ~20 seconds when i introduce the stuff from the [[(hidden) nicjeov_l]] branch which is currently WIP.

((yxHMW9jlb))

ofc there's a lot of [[(hidden) BislYexyN]]s to be done, it's currently fairly naive. and, they seem more doable / easier than the side effect detection:

((RpqCi8rWX))

both assume 1. my size of the graph (1K pages, 8K words)

((FI6sCM5cI))

((xUKym_BXn))

((gf1mkreMz))

((spuKqLoxE))

((m1Fk_dVUP))

{{[[TODO]]}} datalog query builder w/ live preview, directly in [[roam]]?

maybe first gotta go thru the learning material above, lol

since this is getting so wide, i wonder how the generated html pages via [[export-html-pages]] will look like - i'd actually prefer if there was no wrapping and you could just scroll horizontally. will see soon once the export goes live

observation:: putting everything inside daily notes, vs referencing today's date in a specific page / block

if the page is public, but your daily notes aren't, then the stuff will also be public

but wait this is irrelevant because if using linked a linked reference which makes the stuff public, then it's the same effect, it juts ends up in the linked references of a page instead of directly inside it. [[duh]]!

{{[[TODO]]}} wait but i'm still confused myself now on how does the public vs public recursive logic work in [[roam-traverse-graph]] lol - is making a page "public" makes it public recursive as well?

e.g. even inline queries!!

:q [:find(count ?t):where[_ :node/title ?t]]

pages

:q [:find(count ?t):where[_ :block/string ?t]]

blocks