Hacker News Reader

Comment by jasode | Hacker News Reader

>, org-mode eliminates the risk of hitting any sort of barrier

Ultimately, it depends on the type of user but the "barrier" may be that it's plain text. So Org Mode's data format as a desirable selling point is also its ceiling of functionality.

Yes, I store 90% of my notes in plain text. But I also store some programming notes in MS Excel because I need a datagrid as UI with dynamic calculations. I store some project notes in MS Word because I need to embed graphics along with text. I suppose I could use Org Mode as a "single source of truth" by linking to .xlsx and .docx files but I reorganize my folder structures too often to make such file links stable.

I made a previous comment why Emacs Org Mode along with everything else I tried in the last few decades to organize life is a very hard problem of "boiling the ocean": https://news.ycombinator.com/item?id=23112843

View on HN

Actually org mode can do all of those things:

* Recalculate formulas a la excel: https://orgmode.org/worg/org-tutorials/org-spreadsheet-intro...

* Actually you can go way further than cloning basic excel functionality. For example, you can write inline code to generate or retrieve some data, feed that into a graphing program, and then inline the resulting chart back into the org file. https://orgmode.org/worg/org-contrib/babel/intro.html

* Display inline images: https://stackoverflow.com/questions/17621495/emacs-org-displ...

Just because it's possible in theory doesn't make it a good idea. In org-mode, anything that isn't plaintext is quite painful to set up and to maintain, at least until you've memorized two dozen arcane key chords. I would love for Emacs to get a Notion-like UI (still with plaintext underneath obviously) one day but until then, I'm okay with people using the best tool for the job, which isn't a text editor.

(I'm an enthusiastic org-mode user, by the way.)

In practice, I've used org tables (as a minor mode) even before I've started to use org mode regularly -- so much I've liked it (use case: manual entry of smal tables with occasional formulas)

I find inline pictures very helpful when Org Babel is used as analog of Jupyter notebooks (plantuml diagrams, pandas plots).

I honestly tried to use org-mode tables and formulas to record and process student grades and it was a huge failure. I am very comfortable with emacs, but this particular aspect of org-mode took 10x the time a simple spreadsheet software would have taken. And I don't see this time decreasing with repeated usage.

I am back to using org-mode for other stuff.

Org-mode documents can include graphics[0] and (formula-driven) spreadsheets[1].

[0]: https://orgmode.org/org.html#Images

[1]: https://orgmode.org/org.html#The-Spreadsheet

Yes but I use MS Excel features like pivot tables and MS Word graphics features like drawing arrows as overlays on top of screenshots to annotate them.

The other limitation of out-of-the-box Org Mode "plain text" architecture is that it limits collaboration of letting other users edit my file because fine-grain locking is not as straightforward as with a db file like SQLite. My "master todo list" is easier to manage if I can let my significant other can enter items for me instead of me doing all the manual data entry myself.

I do understand that since Emacs has Lisp, it can ultimately do anything if you really want to customize/extend/research it.

These are actually not that hard to do in org-mode, if you are willing to use babel and push the pivot out to something like pandas. :D

That said, no, I don't actually think it is a replacement for spreadsheets. Especially not if you are already fluent in them.

Come to think of it, someone should write pivot tables for Emacs. Could be an alternate mode for opening CSV files (which then could also be used when editing code snippets in Org Mode).

(EDIT: In the meantime, I guess you could shell out to https://www.gnu.org/software/datamash/manual/datamash.html.)

https://www.gnu.org/software/datamash/manual/datamash.html Fixed link

On the topic of commandline spreadsheets; I'm a huge fan of sc(1), a cute curses-based spreadsheet calculator.

https://en.wikipedia.org/wiki/Sc_(spreadsheet_calculator)

I'm not sure what you are describing. Note that doing this with python and org is already pretty easy. Or R. No, you don't fully open the csv in org. But, that usually seems a feature. The firehose that is large csv is typically best as a related file.

I'm thinking of the middle-sized use cases. A table with couple hundred to couple thousand rows. Small enough that you can load it into Emacs without a performance penalty, but too large to pivot by hand. Personal and freelance business budgeting come to mind.

However, at this point I'm OK with shelling out to external tool for data wrangling.

That is still easy with pandas/babel. Just set the table as an input for the snippet, and then you can output a pivoted table. (That make sense?)

Yes, it does, and is probably the best option for most people.

> by linking to .xlsx and .docx files but I reorganize my folder structures too often to make such file links stable.

Linking implies too many manual operations to keep things tidy when moving headings around. Try Org attachments[1].

- Basically, a heading has an attachment directory, whose name is by default generated by Org, and optionally inherited from some level in the outline hierarchy.

- Any attachment is represented by a heading with the :ATTACH: tag and an ID in the :ID: property of its drawer. Actual files can be e.g. copied, moved, or symlinked.

- Attachment directories, the :ATTACH: tag and :ID: property are conveniently added by one of the attach commands, but can also be customized (`org-attach' is the dispatcher, bound by default to <C-c C-a>).

- For a given heading, run a command (conveniently, from the dispatcher) to visit the attachment directory for the heading from the configured file browser or Dired: <C-c C-a f> and <C-c C-a F>, respectively.

If it makes sense to move attachments when moving headings around, or when otherwise adding files externally to the attachment directory, there is `org-attach-sync' (<C-c C-a z>):

  Synchronize the current task with its attachment directory, in case you added attachments yourself.

[1]: https://orgmode.org/manual/Attachments.html

Further explanation: When moving headings around, one can choose to also move the attachments along, for example, by relying on OLD-HEADING and NEW-HEADING's respective attachment directories. Changes to these directories can include external modifications (e.g. addition of files produced by a spreadsheet app). `org-attach-sync' makes sure these changes are propagated to the Org file (and/or its memory representation).

When referencing one of the attached files within the heading, one can use the `attachment' link type, which points to files resolving to the attachment directory[1]. The format is [[attachment:FILE-IN-ATTACHMENT-DIR][DESCRIPTION]] and can be easily inserted with completion by `org-insert-link'.

For more complicated attachment linking scenarios it might be better to configure the attachment directory to be absolute (`org-attach-id-dir').

Conclusion: The use of Org attachments and the attachment: link type stabilizes links to files.

[1]: https://orgmode.org/manual/Attachment-links.html

Reply on news.ycombinator.com