[tconfrey]

"Growing support is easing collaboration via org files."

The same month this article was written (4/22) I tried to make the case for org as an interchange format for productivity tools [1] and pointed to most of the same tools supporting org. (Disclosure, my browser extension, BrainTool[2] is listed in the article). I still love the idea of a local-first, plain text model for sharing personal data across productivity apps, but three years on its not clear to me that the momentum has been maintained. Are folks still building new things on top of org?

[1] https://braintool.org/2022/04/29/Tools4Thought-should-use-Or...

[2] https://chromewebstore.google.com/detail/braintool-go-beyond...

[adityaathalye]

Post author here... My dream: a grammar or specification for org text. So that someone can make a self-contained parser library (I'm not smart enough). So that others can embed it in all their tools, and so that someone can make an LSP for org text.

IMHO, the integration of org-mode into Emacs is a double-edged sword. Great for me, because I get batteries-included writing support. /However/ terrible for distribution. The source code of the canonical implementation is the specification. Unfortunately, it is hardwired to the Emacs binary. That fact alone makes org text verboten to most of the known universe. Even if that weren't true, the absence of a standard makes it incredibly difficult for outsiders to maintain feature-parity of their own implementations with the canon, and with each other.

[internet_points]

> The source code of the canonical implementation is the specification.

https://orgmode.org/worg/org-syntax.html is the specification.

[alwayslikethis]

The problem is a lot of the semantics is dependent on one of a myriad of user's configs. As a simple example, TODO keywords, tag inheritance, etc.

[dietr1ch]

You could specify what's vanilla default in the org packages.

For that example, there's a file-local way to specify custom TODO keywords[^0]. Maybe that can get into the standard, but I'd be Ok with with a primitive, vanilla specification for org-mode first as striving to support everything would ensure no projects form around an emacs-free org-mode anytime soon.

---

[^0]: https://orgmode.org/manual/Per_002dfile-keywords.html

[tconfrey]

Completely agree, on both counts. Users often ask me how they can edit their BrainTool.org file outside the extension. There's no way I can in good conscience point the average knowledge worker toward installing an emacs distro!

FWIW I use orgajs (https://github.com/orgapp/orgajs). Its well-supported and gives me a pretty complete AST. Highly recommend if you're working in JS.

[arminiusreturns]

My hacky solution for this when working with others has been using git post-receive/commit hooks to execute the org document (babel, pandoc, emacs html export, etc), so anybody with any editor can, if they understand the format, make changes without having to have emacs.

I heavily abuse commit hooks in my hacky CI/CD pipelines though, so ymmv.

[internet_points]

> three years on

I have been using org-mode for nearly two decades. Org-mode itself has existed for 22 years (beating markdown by 1 year!). For something as important as an archival format, I wouldn't base my decisions on the vagaries of the latest trends.

[tconfrey]

org is timeless for sure! Personally I've been using emacs since the 80's. But as the developer of a personal knowledge management application I was hoping that an ecosystem of specialized, easy to use, tools based on org would emerge.

BrainTool is a browser extension for managing tabs and bookmarks. It saves your data in an org file. One feature is to allow an item to be marked as a TODO. For me personally I then integrate my BrainTool.org file into my overall org workflow and see that TODO in my org-agenda. But ideally a naive BrainTool user would be able to see that task in a tool like Todoist or whatever.

For a while there it seemed like there was some momentum!