Hacker News new | ask | show | jobs
by angleofrepose 2621 days ago
What sense does it make to have org mode outside of emacs? I think the entire power of the system would be that it explicitly lives within emacs. From the repo:

> Org is probably the best and most complete plain text organizational system known to mankind. It has countless applications like authoring, publishing, task and time tracking, journal, blog, agenda, wiki etc...

Yes, absolutely.

> Unfortunately Org was originally developed for Emacs and therefore available only inside Emacs. It is a huge limiting factor for Org's development and popularization. Because of that it is not as popular outside of Emacs community as it should be.

"Unfortunately"? It is a dynamic tool, I don't understand what it would mean for it available outside of emacs except for if that were also a dynamic tool. Are you going to create a standalone dynamic org mode application? It seems better to make a "stripped down" emacs that only has org mode and literate programming features so as to not have as steep a learning curve as the entirety of emacs all at once. What other reason is there for this work except for avoiding the learning curve of emacs? Emacs has got to be one of the most configurable pieces of software outside of Operating Systems(maybe?) in existence.

Org mode is powerful because it is a plaintext system that gives hooks for emacs abilities like transformation (faces, folding, tagging/searching, agenda features) and execution of code with literate programming, and passing values around between code blocks and so on.

Why would this want to live outside of emacs? I do think the recently (and again two years ago) extremely popular post on org mode's markup language absolutely misses this point[1]. This post is also linked to from the repo.

I'd love to learn about what I'm missing here, and would love to get some answers to my questions or responses to my assumptions! I'm very interested in your roadmap and design decisions, I think your external links (except for that article) are fantastic. I will follow the project for sure.

I just think there is some mental disconnect with how org mode is represented and discussed on hn, and those discussions almost all seem to miss the majority of org mode abilities that absolutely depend on living in the style of environment that emacs provides. From reading the top discussions here, one might think org mode is in competition for the same space as markdown while that is not even remotely true [2].

[1]: https://news.ycombinator.com/item?id=19622019

[2]: https://hn.algolia.com/?query=org%20mode&sort=byPopularity&p...

PS: I am trying to do a more thorough review of the discussion about org mode on hn, and I will be taking actual notes and comment links down to better explain my ideas here rather that just general takeaways. Thanks for bearing with me, or letting me know if you disagree.

4 comments

> Why would this want to live outside of emacs?

It allows us to throw different use-case-optimized GUIs at a sync'ed folder of org-mode files. A clear example are the mobile apps, but there are others.

I've been using org-mode for a year, trying to replace a 5 year-old Evernote knowledge base/todo list, which itself inherited from a 12 year-old InfoSelect knowledge base.

For task management, org-mode on spacemacs complemented with beorg on mobile is proving to be the best system I've ever used.

As for the knowledge base features, however, I'm still far from being able to reproduce the speed and precision of recall that I had on InfoSelect, or even Evernote with a solid tag hierarchy. I also use Dynalist to have an outliner on mobile.

On mobile, the KB situations is even worse, since beorg is optimized for the task management use-case and does not provide proper search.

I would also love to be able to browse connections in my knowledge base, TheBrain-style, but that will always be clunky to unusable on a pure-text UI.

It would be awesome if I could get rid of Evernote and Dynalist and just point different apps at the same org file repository.

This is one reason having alternate parsers for the org markup is important.

> It allows us to throw different use-case-optimized GUIs at a sync'ed folder of org-mode files. A clear example are the mobile apps, but there are others.

This is a good point, I agree.

> As for the knowledge base features, however, I'm still far from being able to reproduce the speed and precision of recall that I had on InfoSelect, or even Evernote with a solid tag hierarchy. I also use Dynalist to have an outliner on mobile.

What speed and precision are you missing? It seems that org mode should have the capacity to establish any content connection scheme you could dream up.

> KB situations

What does KB mean?

> I would also love to be able to browse connections in my knowledge base, TheBrain-style, but that will always be clunky to unusable on a pure-text UI.

I don't think this is true. Depending on the kinds of connections you want to browse in your knowledge base there are plenty of ways to transform the ui to show them. That is what I see org mode to be at it's essence, a software tool to allow for transformations of text on screens. I think though, that your point is more about browsing. We should have text-based ui browsers for connections. All a gui like the brain is doing is putting animations and a mouse on top, we can definitely specify things via the keyboard in a number of ways to replace the mouse, and who wants animations to begin with? If the main pull at the end of the day is that GUIs are pretty and text UIs aren't then let me point you here [1].

[1]: https://lepisma.github.io/2017/10/28/ricing-org-mode/index.h...

> What does KB mean?

Knowledge Base

> What speed and precision are you missing? It seems that org mode should have the capacity to establish any content connection scheme you could dream up.

That's what I thought too, but I haven't been able to accomplish it yet.

In InfoSelect I had a tree on the left that I could filter instantly and incrementally while I typed in the full-text search box (imagine org-sparse-tree + helm-org-rifle on steroids). In Evernote I have a double hierarchy of notebooks and tags (vaguely based on The Secret Weapon [1]) that allows me to create a new project tag where I pull together new action notes, meeting notes, and old notes from my 17-yo knowledge base. I'll have anything between 5 and 20 ongoing projects that I can toggle quickly on the left hand navigation bar. The notes are shown in a table where I can click on the header to sort by title, created date, or updated date, so I can find what I need within a few seconds.

I can also do a mixed full text/tag/date search adding filtering criteria as I go. In org, I have to decide beforehand what kind of search I want, and if it's a mixed search the syntax becomes very cumbersome.

With org, I can do none of the above on mobile.

In practice, this means that I have more trouble finding things in my 14 month-old org setup than I do in my 17 year-old knowledge base.

> What does KB mean? "Knowledge Base", sorry for dropping the acronym without definition!

> All a gui like the brain is doing is putting animations and a mouse on top, we can definitely specify things via the keyboard in a number of ways to replace the mouse, and who wants animations to begin with?

I have zero problems with using the keyboard. I've been using vim and editors with vim keybindings since the 90s and my favorite laptop of all time doesn't even have a touchpad. There are cases however where specialized GUI affordances can be helpful. I would say that navigating a huge graph of nodes using mind-map-style partial views is probably one of them. I know there's a brave soul trying to pull this off in Org [2] and I sincerely wish them success.

> If the main pull at the end of the day is that GUIs are pretty and text UIs aren't then let me point you here

Pretty is also important but the default Spacemacs theme is pretty enough for me. Mixing fixed and proportional fonts would be nice, I've had that on my to-do list for a while.

[1] https://thesecretweapon.org/ [2] https://github.com/Kungsgeten/org-brain

> I just think there is some mental disconnect with how org mode is represented and discussed on hn, and those discussions almost all seem to miss the majority of org mode abilities that absolutely depend on living in the style of environment that emacs provides.

Honestly, part of it may be just baiting people :). The plaintext markup parts of Org Mode format are really just a Markdown equivalent; the true value comes from the features provided by Org Mode the application itself, and most of these are what they are precisely because of Emacs environment.

These days, with new Emacs users the story usually is "Come for Org Mode and/or Magit, stay for life"; personally, I love to encourage people to take that first step :).

Ha, you're probably right. Frustrating, but makes sense. Thanks for taking that on.
I have just recently found Org (last week) as I stumbled on spacemacs and then fell down that rabbit-hole, which had a branch off into this rabbit hole.

Now I'm having experiencing some Baader-Meinhof phenomenon as I've seen two org-mode articles on the front-page here since and never noticed any before.

Anyways, this seems sort of obvious to me as a total org (and emacs) beginner... the sheer flexibility and power from this being built on top of the emacs system is the real killer feature, though lots of other organization apps would do well to copy some the base features. I really don't understand why repeated tasks is so hard for other task management apps to implement well and is yet easily done by whoever put them into Org:

https://orgmode.org/manual/Repeated-tasks.html ...

   ** TODO Call Father
   DEADLINE: <2008-02-10 Sun ++1w>
   Marking this DONE shifts the date by at least one week, but also
   by as many weeks as it takes to get this date into the future.
   However, it stays on a Sunday, even if you called and marked it
   done on Saturday.

   ** TODO Empty kitchen trash
   DEADLINE: <2008-02-08 Fri 20:00 ++1d>
   Marking this DONE shifts the date by at least one day, and also
   by as many days as it takes to get the timestamp into the future.
   Since there is a time in the timestamp, the next deadline in the
   future will be on today's date if you complete the task before
   20:00.

   ** TODO Check the batteries in the smoke detectors
   DEADLINE: <2005-11-01 Tue .+1m>
   Marking this DONE will shift the date to one month after today.
Plain text on top of a programmable interface is the way to go. I like this comment from yesterdays 26.2 release [1].

[1]: https://news.ycombinator.com/item?id=19650131

I think just having the table creating aspects would be very useful as a way to quickly make plaintext tables
For sure, I think in the 2008 Google Talk by Carsten Dominik (org mode creator) his host makes a comment from the back of the room that he has a use case(among others) of creating html tables in org mode because its so easy. Org mode has so much functionality hooked in to plaintext structures it is ridiculous. Tab and S-Tab alone do so much work in an org mode workflow, it is things like this that make it hard for for me to understand what good an external parser for org mode would do. Org mode is incredibly feature rich for authoring.