Feature Request: Org mode

[codygman]

Vision: 100% Haskell clone of https://orgmode.org/ using Yi

Why?

1. Performance

I don't know about anyone else that uses org mode, but once my org mode documents get to be over 1MB I start getting lots of weird little hangs and general laginess. I've tried debugging org-mode before, and it seems all roads lead back to https://github.com/codygman/figure-out-emacs-long-lines-issue/blob/master/figuring-out-emacs-display-issues.org

2. Familiarity with Haskell

For me personally, but I'm sure there's a non-zero overlap of people who use org-mode and haskell.

3. Babel Performance

I've recently changed my entire workflow to use something akin to outlined tasks interspersed with babel source code blocks which are sometimes tangled to source files. For an example of part of this workflow, see http://www.howardism.org/Technical/Emacs/literate-devops.html.

However, performance of babel which writes results to buffers (with very long lines usually that slow down emacs (even moreso on the osx machine I use at work)) and/or to files. The lack of concurrency here slows everything down.

4. Tramp performance

Would only worry about this after #1071 of course, but I think having a much faster tramp mode would blur the line between an interactive ssh session and a literate devops document that is reproducible.

How? Roughly speaking, I think implementing it would look like:

• toggle expand of nodes
• TODO/DONE functionality
• haskell source code blocks
• shell source code blocks
• extend source code blocks to respect tramp :dir /ssh:server.remote
• allow any node to be annotated with schedule DATE or deadline DATE
• org agenda

I haven't researched everything, but I did find some org mode parsers available in Haskell:

• https://hackage.haskell.org/package/orgmode-parse (attoparsec based)
• https://hackage.haskell.org/package/orgmode (parsec based)

Given my reasoning for this is increased performance, I'd be more easily swayed towards the attoparsc based parser unless I see it's lacking some pretty big conveniences or is much less mature. I have no prior experience with either though.

I've had this in my head forever and have delayed posting until I had a better proposal, but I've finally decided to just post this up :)

Thoughts?

[Fuuzetsu]

org mode does a lot of things, 100% sounds far too ambitious. For example you'd have to implement nested parsers. We should outline actual features we want first.

[codygman]

@Fuuzetsu Don't necessarily disagree, I'll try to add more on the minimal feature set my workflow would need. Hopefully that coincides with others needs as well.

[ghost]

Besides the ambitious goal of org-agenda a lot of the same functionality can be achieved with markdown-mode in emacs, which would be great to have in yi