that's actually a big use case of mine. i use my notes as a crm for the people and conversations i've had.

yml
- people hierarchy: people.{name}
- meet hierarchy: meet.journal.{date}.hint

i have a section in every meeting note for people that are involved

yml
# notes
- ...

# people
- [[people.jojanaho]]

this way, i can navigate to the people note and have backlinks to all conversations we've had

where airtable comes in - i like their aggregations and formula. before dendron, i had an airtable setup in the same way and would do an aggregation to see how long its been since i've last had a conversation with someone. this served as a good queue to remind me to catchup with folks i haven't talked to in a while

the question is if we want to re-implement airtable functionality inside dendron or have a good way of exporting/syncing data from dendron to airtable using a pod

i'm currently biased towards the pod idea because, at the end of the day, i think airtable will be the best tool to use for airtable like functionality. instead of trying to compete with airtable, i want it to be easy to import/export/sync your data from dendron to airtable so that you can reference it inside dendron and have it open to the relevant airtable entry where you have the full power of airtable at your disposable

instead of fully fledged airtable, I was thinking something slimmer and compatible with text; essentially ability to create inline anchors in the text (including csv, markdown tables and yaml) and streamlining existing linking syntax to reach these ids (e.g. [[my.note#anchor]])

that would allow foreign-key like functionality everywhere

got it. that makes sense and is high on my priority list as well. i was just thinking md files but being able to deep link to any plaintext file is better

this is actually why dendron's note reference syntax is rather verbose:

yml
(( 
    ref:[[name]]#anchor1:#anchor2
))

i want it to support more exotic references and operators moving forward

yml
(( 
    ref:[[foo.csv]],
    startLine: 2,
    endLine: 5
))

some other operators:

yml
startRegex: /^foo/  # use for anything
startKey: fooKey # use for yaml and frontmatter
startHTMLId: ... # use for <span> tags inside markdown to reference arbitrary sentences

Yeah, I was thinking something along these lines

markdown
## Ingredients:

| id   | name     |
| ---- | -------- |
| §cuc | cucumber |
| §tom | tomato   |
| §pot | potato   |

## Dishes

| id     | name     | ingredients    |
| ------ | -------- | -------------- |
| §soup1 | Odd Soup | $cuc $tom $pot |

basically getting intellisense + ctrl-click-open + diagnostics on id mismatches

and then extend something similar to yaml etc

(link and id syntax sucks, just slapped something there :))

and why not just use airtable? same reason as why not just use roam. Still want to have this running locally, compatible with git and fast text editing capabilities

I bet this could also use schemas somehow, e.g. saying that ingredients column in dishes is a reference to ingredients table; in that case no special link syntax would even be needed

haha, you know you can get me to do anything by bringing up schemas 😛

and yes, i like this a lot. schemas, in addition to describing the hierarchy, can also describe aspects about the note and be used as hooks for more advanced functionalities like foreign references

makes me wonder if there really isn't demand for locally running airtable, since I just cannot find such a product anywhere.

would have opened my wallet already

many times in the past

yeah, excel used to be a thing but now its office 360 and like everything else, is in the cloud

R, pandas, python notebooks are what i've done in the past

yep, I guess things like Access are the closest match, but they are way too clumsy

going back to foreign keys, i think it makes sense. ideally, i would like to be able to specify a foreign key as both a

link

and as a

note reference

(radical idea: in the future, they'll be the same thing. a link is just a special case of a note reference that doesn't show the content)

if you have ideas on this, would love to read any proposal/rfc/issue that you come up with

yeah, coming up with a unified, simple and short syntax is a challenge

...and non-conflicting one, not really sure which even would be conflicting cases (e.g. using hashtags in yaml probably doesn't fly due to #-commenting)

btw, what are your current plans regarding LSP; these kind of things probably only make sense after that is up and running

my current plan on lsp is to spend this weekend doing a deep dive into the specs/docs and writing my findings here: https://github.com/dendronhq/dendron-template/blob/master/vault/project.n.2020.lsp.md#L1:L1

i should be in a better place to figure out timelines and implementation details afterwards

sounds good. Fun fact: before foam appeared, I was writing something similar using LSP. Got to the point of intellisense and navigation working