I went down a rabbit hole with the built-in Remember mode in Emacs today.

Persistent Scratch with remember-note

The Emacs *scratch* buffer is handy. It's always available and easy to get to. However, by default it doesn't persist between runs of Emacs. So, when I quit Emacs, anything in the scratch buffer is lost. Sometimes I get burned by this, so I thought I'd fix it using the persistent-scratch package.

While looking for details about the package, I spotted a couple comments about using the built-in remember-notes function instead.

Return the notes buffer, creating it if needed, and maybe switch to it. This buffer is for notes that you want to preserve across Emacs sessions. The notes are saved in ‘remember-data-file’.

OK, I tried it and it works great.

I now have a persistent Org mode file as my scratchpad, and I can easily capture new notes as org headings to it using remember.

New frame defaults

Now that I had a nice, persistent scratchpad, I wanted it to be the default buffer for new frames and at startup. Easy enough:

;; Persistent notes (like persistent-scratch, but built-in)
(setq initial-buffer-choice 'remember-notes
   remember-data-file "~/Documents/Notes/remember-notes"
   remember-notes-initial-major-mode 'org-mode
   remember-notes-auto-save-visited-file-name t
   remember-in-new-frame t))

(defun my/switch-to-scratch-buffer (f)
  (with-selected-frame f
    (remember-notes t)))
 
(add-hook 'after-make-frame-functions #'my/switch-to-scratch-buffer)

I'm not sure all of that is necessary now that I've learned about remember-in-new-frame but it works as-is, so I'll leave it.

Remember mode via Raycast for system-wide capture

I'd created a keymap (C-c o n) for bringing up the remember-notes buffer while in Emacs, but thought it would be nice having a system-wide option. I've never understood how to configure =org-protocol= so I cheated and created a Raycast script:

#!/bin/bash

# Required parameters:
# @raycast.schemaVersion 1
# @raycast.title Remember
# @raycast.mode silent

# Optional parameters:
# @raycast.icon images/emacs.png

# Documentation:
# @raycast.description Open Emacs remember in new frame
# @raycast.author Jack Baty
# @raycast.authorURL https://baty.net

emacsclient -e "(remember-other-frame)"

Now I can trigger Raycast, type part of "Remember" (usually just "R" is enough) and it brings up a new Emacs frame containing the remember buffer. I type or paste whatever, and hit C-c C-c which adds the note to the remember file and kills the frame. Handy!

Creating emails via Remember mode

While reading the Remember manual, I spotted the option to save Remember notes as emails. WUT?!

(add-to-list 'remember-handler-functions 'remember-store-in-mailbox)

By default, this uses an mbox file, ~/Mail/remember, which I can use Mutt to view.

neomutt -f ~/Mail/remember

Crazy, right? Yeah, I'm not going to view my captured notes in Mutt, but it's cool that I could. But it got me thinking about my urge to put everything into Notmuch. Notmuch indexes Maildir-formatted emails, not mboxes. What if I converted the mbox to Maildir and used Emacs to read the notes where I'm already reading my emails? Plus super fast search and tagging, etc.?!

I've not gotten to that yet, but it's something to think about.

Now what?

This whole thing started because I wanted to make my scratch buffer persistent. Instead, I've ended up with a whole new capture setup for quick notes.

Why not use Org-capture or Howm-remember or a million possible other things for this? Fair question. The answer is that I didn't know anything about Remember mode when I woke up this morning, and now I do. It remains to be seen whether I should have used Org-capture for this, but it was fun learning how to do it this way.

It's possible we've, er, forgotten about Remember mode too soon.

lmq-10/denote-search

This package provides a search utility for Denote, the simple-to-use, focused-in-scope, and effective note-taking tool for Emacs.

The command denote-search is the main point of entry. It accepts a query, which should be a regular expression, and then searches the contents of all the notes stored in denote-directory for it. The results are put in a buffer which allows folding and further filtering; all standard commands offered by Xref are available as well.

This package has the same code principles as Denote: to be simple-to-use, focused-in-scope, and effective. We build upon Xref to be good Emacs citizens, and don't use any dependencies other than Denote and built-in libraries.

This works pretty well. It returns a nice, foldable list of results. One can also refine the results by filtering by file names, etc. I'll probably continue using Deadgrep, but denote-search is nice to have in the quiver.

When I bought the Nikon Z f recently, the way I intended to offset the cost was to sell the Leica SL2. The mistake I made was to buy the Nikon before selling the Leica.

Yesterday, I gathered up the SL2 with accessories and took some photos of it for the listing. Picking up the camera reminded me that, as nice as the Z f is, it's not in the same league as the SL2. Everything about the Leica feels nicer. Controls, covering, grip, doors, battery, all of it.

I've been lost in the Nikon's menu system for a week. I still don't know where anything is. There's more "stuff" available on the Nikon, but the Leica's menus are simple, clear, and easy to understand. Sometimes less is more, you know?

I started to reconsider the move to Nikon for my digital photography.

On the other hand, the Z f is significantly smaller and lighter. Accessories are less expensive. The RAW files look good, and are much smaller (due to being 24MP vs 47MP, sure, but still). It's better at everything in low light. And auto-focus with the Nikon is miles ahead of the Leica. Eye detection with continuous auto-focus is a thing to behold.

I'm keeping both cameras.

The Nikon is for everyday stuff. It's perfect for the dog and my grandson and out-and-about photography. The Leica will remain for more "serious" things like studio portraits, which I want to do more of.

That's settled then. 😉.

My use of Howm has continued to grow organically. In other words, I don't set out to use Howm, it just happens :).

One of Howm's unique features is its TODO implementation:

Format of schedule and todo:
[2002-10-21]@1  schedule -- (shown in schedule part. @3 = "3 days schedule to 10-23")
[2002-10-21]+7  todo -- float up slowly from the date in 7 days
[2002-10-21]!7  deadline -- float up fast from 7 days before the date
[2002-10-21]-1  reminder -- float at the date and sink slowly one unit per 1 day
[2002-10-21]~30 defer -- float at the date and repeat sinking and floating with 30 days period
[2002-10-21].   done -- sink forever
(Numbers after marks are default values.)

How to remember:
- Schedule at(@) the date
- Reminder sinks down(-).
- Todo floats up(+).
- Deadline needs attention!
- Defer waves(~) up and down.
- Done is the end(.).

These TODOs can be placed anywhere in any Howm file, but sometimes I just want to capture a task and have it added to an existing list of TODOs. I created a simple org-capture template for this:

  ("h" "Todo to Howm" plain
           (file+headline , "~/Documents/Notes/Denote/Howm/20250224T081548--todo.org" "Inbox")
           "- [%<%F>]+ %?")

Running the capture template puts a new, dated TODO into a dedicated Howm TODO file. After a few uses for testing, it looks like this:

* Inbox
- [2025-02-24]. [2025-02-24]:+ Create capture template for Howm dates
- [2025-02-24]. [2025-02-24]:+ Move TODO file to root of Howm tree
- [2025-02-24]+ Fix categories on Baty.net
- [2025-02-24]~ Buy replacement hinges for laundry room
- [2025-02-24]+ Something to do
- [2025-02-28]@ Meeting with tax guys

They show up when needed on the Howm menu, like so:

At some point I should add a prompt for the date to use as well as the type of TODO, but for now, it's a small, useful quality-of-life improvement.

We are born with hearts that want to leave a mark on the world. This ability is not lost, it is taken from us.

Mike Monteiro, How to draw an orange

Currently, my blog has two types of posts, Journal and Post. Journal entries are comprised of things I collect or think about during each day. I post one per day. Posts are your typical blog post. Posts are (usually) a bit longer and are about a single topic.

I'd like to eliminate the Journal posts, but have not found a way to do that. I could post every random thing that occurs to me as a full Post, but that feels too noisy and weird. I could move the Journal posts to the wiki. I still do that, sometimes. Or, I could stop writing them. Only the first two options are feasible, so which is it? I don't know.

For now, I'd like to move more of the Journal notes into their own Posts, and leave the journal for truly unstructured bits and bobs. It would be noise, mostly.

Still noodlin' on it.

I started free writing again a couple weeks ago, but this time I'm doing it digitally, using the AlphaSmart Neo 2.

I dove hard into "Morning Pages" after reading Julia Cameron's The Artist's Way, but it didn't last long. I mean, c'mon, there's no way I can consistently write 3 lined pages, longhand, every morning.

The intended benefits of doing hard-core morning pages are lost on me, but I can get behind the practice of doing a little free writing each day. For me, it's more like forced journaling.

My version of morning pages is more like "Morning Pages Lite". I set a timer for ten minutes, press the On button, and start typing. Ten minutes is plenty long enough for me to uplug the tubes in my head. I'm not an artist, so that's all I need.

I haven't written morning pages longhand in quite a while. As much as I love journaling with fountain pens and paper, writing quickly for long periods as an exercise turns into a chore, and I don't enjoy it.

For some reason, I didn't post about it when I bought my first Neo, so I don't remember when I got it. Around 2012, I think. I remember paying $25.

For digital journaling, using the Neo is fun and quirky and about as far as one can get from writing on a computer. It shows only four lines of text at a time. The keyboard isn't fancy, but it's fine. If I want to transfer my writing to a computer, I plug in a USB cable, open my text editor, and press "Send", like this.

The last set of batteries I put in the Neo lasted nearly 3 years. That's 3 years!

Free writing for a few minutes in the morning is something I find useful. Changing the way I do it once it a while keeps it interesting.

I "settled" on using notmuch for my email, but now I'm reconsidering that decision.

Notmuch is great, but using it forces me to have two email stores. One is IMAP (via Fastmail). The other is my local notmuch database. Mbsync keeps things kind of synced, but it's really only a few flags. Notmuch doesn't delete or move email on the server. This means I'm managing, for example, my Inbox, in two places. It's not hard, but is it necessary?

I need to read and manage email on my phone, and on both computers. After spending days wrangling issues with syncing between Macs, and knowing that there is a lot that I'm missing when viewing email on my phone, I'm beginning to feel that maybe doing my email this way isn't ideal.

Email in Emacs is important to me, and if notmuch is causing issues, there's always Mu4e. I used Mu4e for a long time before switching to notmuch. Mu4e is easier and feels "nicer" (subjective, I know). It also is a direct mapping of what's on the IMAP server. This means that my email is my email is my email, if you know what I mean. There's one version of it. I still have local copies of everything in ~/Mail, too.

So, my dream of putting my life in notmuch is on hold. I need things to be simpler right now, and using Mu4e makes things simpler for me.