> [!tldr] Why is there not a spreadsheet application that works like Obsidian?
[[Obsidian]] sits on a folder in your computer and treats the content of the plain text files as the source of truth. You could very easily imagine a [[Spreadsheet]] application that sits on top of a folder allow you to do spreadsheet-like things directly on the underlying source files, be they [[CSV]], [[JSON]], or [[XML]].
My original idea here was mostly for CSVs, so I'll focus on that for a minute.
> [!warning] Challenges:
> - CSVs have no method to describe [[Data Types]]
> - The concept of "formulas" would need to be managed as part of the overlay layer, not on the files themselves
> - ...unless you could bake formulas in as [[Frontmatter]]?
...Come to think of it, [[Why does CSV + Frontmatter Not Exist?]].
> [!example] Why it doesn't exist...
> - spreadsheets serve this need
> - [[Jupyter Notebook]]s, also
> - no strong push for this in the market
> - baking in features means adding constraints, which defeats the purpose
So far as I know, no real product exists in this space. Building one could be fun, but also would probably take a real programmer using something like [[Rust]] to do it right.
How much of a product space is there if you:
- don't change what makes [[CSV]] special
- add in field types
- Pivot between formats
- add in derived fields
- simple: `A` = `B` + `C`
- less simple: `A` = `B` + ave(`c`)
- even less simple `A` = if(whatever)
- least simple: `A` = ave(`B`) where `C` == `D`
- cache columnar averages & sums & whatnot
... man between this and [[Aaron's Hypothetical Modeling Tool]] I've got two infinitely deep rabbit hole projects I could jump down.
****
# More
## Source
- self