There is a lot of note-taking software out there, and the trouble with most of it is that it's too darn complicated. Which is a shame, because note-taking is something a lot of people do with computers.
There are exceptions like Google Keep and thinktype but it seems to me that most of the big names are useless and confusing to most people - including me - because they require so much investment in learning how to use them. You need something simple and casual to take notes (like, say, a pen and paper) not something complicated.
So, take a breath and look at the landscape. There are, amongst other approaches:
- Commonplace books Leonardo da Vinci kept all of his notes in one big book. If he liked something he put it down. This is known as a commonplace book, and it is about how detailed your note-taking system should be unless you plan on thinking more elaborately than Leonardo da Vinci.
- Zettelkasten or a card file: small items of information stored on paper slips or cards that may be linked to each other through subject headings or other metadata such as numbers and tags.
- Commonplace journals or incremental notes are like a diary, but for notes: start a new page every day and fill it with what you're doing, not doing, or reading, or whatever.
- Spreadsheets. Some people know a bit about spreadsheets and fill them will all kinds of stuff, like the list of clothes they own and their washing schedule. People use the tool they know, rather than the tool that would do the job best - it's like they are pounding in screws with a hammer or opening a paint tin with a the point of a knife, and it upsets me. We will talk no more about trying to model a gloriously messy world with rectangular grids of values.
- Databases. Well, yes, but then everything's a database of sorts, isn't it? If by 'database' you mean 'relational database', then we're back to point 4, which we don't talk about.
The aim here is to make something simple that explicitly covers commonplace books and incremental notes, and enable a little of the functionality of zettelkasten using hashtags. After playing around with some designs for a while, I thought it might work to split commonplace books and journals into two similar apps, so that they can be run side-by-side, so text can be cut and pasted between them.
cj is a simple app that allows you to write and retrieve notes in a commonplace journal. You can have more than one journal; the default one is called Default, but you can create others.
cj generates a new note for you everyday (but you can still edit old notes, or create notes in the future). There is no explicit 'create note' feature; everyday has it's own note.
I toyed with the idea that notes from days before today cannot be edited. Think of it like this: last October, your favorite color was red, so you made a note of it. Now, your favorite color is blue. So, should you go back and edit the note from October, removing your choice from history, or just make a new note? I think the user can just resolve not to edit old notes, rather than have the app decide that for them.
The idea came from The Sephist's article and from using rednotebook for a while.
cb was a companion to cj, another simple app that allowed you to write and retrieve commonplace book notes. However, I abandoned this because it was worse than editing markdown files directly using something like MarkText or Typora.
The general idea is have an instance of cj open all the time, where you put notes and bits of text as they come up during the day. (You can have more than one instance of cj open, one for each journal, if you use multiple journals.) Then, have one or more instances of a markdown editor open, and copy-and-paste text from cj to the markdown editor as that information endures or needs categorization.
Prefix any to-do type journal entries with [ ] and turn them into [X] when they are completed. Find incomplete entries by searching for [ ].
Thereafter, because all the notes are just text files in directory trees, they can be manipulated, exported, reformatted by worthier and more appropriate tools.
cj was first written in Go, with the user interface done using the Fyne library (can't remember where the calendar widget came from).
The search code was copied and adapted from Andrew Healey's grup, but the need for simplicity and flexibility saw that (very fast code) retired and replaced by using grep. I've tested cj with grep version 3.7 on Linux, and it seems fine. You'd better have grep available on your PATH, otherwise searching won't work.
There's no indexing or anything fancy going on under the hood - what we have here is a basic text editor, grep and a small user interface.
Then cj was reimplemented in Tcl + Tk, which has a much better text editor widget. Inspired by the use of grep to do the searching, this version uses ncal to create the calendar widget.
Maybe in the future I'll do versions in Dart + Flutter, or Lua + wxWidgets or lupLua.
All the notes are stored as text files in a directory tree. The root is .cj.
The commonplace journals are stored in directories, one for each journal. The default journal is called Default. Inside each journal directory are directories for each year, and inside each of those, directories for each month. Each month directory contains text files for each day of the month. For example, if you made a note on January 5th 2023 in the default book, it would be stored in a file called .cj/Default/2023/01/05.txt.
You can shadow the entire .cj directory tree in cloud storage, archive them in a git repository (which you can upload to a private github repository), or backup all the notes using, rsync or zip, for example, zip -r <filename> .cj. I use a little bash script to name the backup files after the date they were made, for example:
today=`date +%Y-%m-%d`
filename="cj$today.zip"
cd ~
zip -r $filename .cjYou can use a script to create a single text file from all the data files, like so:
cd ~/.cj/Default
find . -type f -name "*.txt" | sort | while read file; do
echo ""
echo $file
cat $file
doneSend the output of this script into a text file, like so:
./export.sh > exportedThe output of this contains the reletive filename of each day's file, which doesn't look great. So, I use a little gawk script to clean the file up, and make it a little more markdown-flavored:
#!/usr/bin/gawk -f
# note breaks are prefixed with the filename, which looks like "./2023/07/22.txt"
# we have to parse this into three numbers (y m d) using patsplit()
# then convert those three numbers into a systime using mktime()
# so we can format the date however we like using strftime()
# https://www.gnu.org/software/gawk/manual/html_node/Time-Functions.html
BEGIN {
dateLinePattern = "^./[0-9]+/[0-9]+/[0-9]+.txt$"
FS = "\n"
prevYear = 1962
prevMonth = 4
}
{
if ($0 ~ dateLinePattern) {
patsplit($0, d, "[0-9]+")
sysTime = mktime(d[1] " " d[2] " " d[3] " 0 0 0")
if (d[1] != prevYear) {
print "\n# " strftime("%Y", sysTime)
prevYear = d[1]
}
if (d[2] != prevMonth) {
print "\n## " strftime("%B", sysTime)
prevMonth = d[2]
}
print "\n### " strftime("%A %d %B %Y", sysTime) " \n"
} else {
print
}
}Send the output of this script into a markdown file, like so:
gawk -f reformat.awk exported > exported.mdNo tricksy or closed file formats here, no sir.
-data <name of data directory> Defaults to -data=.cj.
-journal <name of journal to open initially> Defaults to -journal=Default.
-fontSize <size of font> Defaults to -fontSize=15
-width <width of window in pixels> Defaults to -width=1024
-height <height of window in pixels> Defaults to -height=640
- Better text editor (including spellchecking, found word highlighting, follow hyperlink, more visible caret, keyboard shortcuts for move word/delete line/goto start/goto end, Unicode support Unicode Character “𝕏” (U+1D54F))
- Support for moving text from
cjto a markdown editor, to facilitate short term to long term note workflow; maybe right-click popup 'copy selected text to commonplace book' ... - More support for hashtags (eg tap on a hashtag to find notes containing it, insert hashtag from dict)
- Support for creating backups, or git, or cloud (Fyne has some cloud support)
- Many little quality-of-life tweaks like colors, keyboard shortcuts, and setting the font face and size.
In the early 1990s, I wrote something called Idealist. That grew out of the idea of merging a database manager and a text editor, and morphed eventually into a package of components that were used to build applications in the museum, archive, and library sectors. 250,000-odd lines of C and Tcl, but I, like many others, just used it to take notes.
Everytime I use a computer, I end up taking notes. Playing a game, developing a new app, doing finances, reading about the worldwide political horrorshows, building bicycles, planning a vegetable garden, following a tv series, reading a book, ... everything seems to generate notes.
So, I've spent decades trying different ways of doing that, trying different apps and methodologies, storing files all over the place and eventually losing them, or not being able to transfer them into the shiny new fashionable app. Idealist is too old, only runs on Windows, and I don't have the source, so I can't use that. After years of nagging at myself, I surrendered to the itch and am developing my own new app(s), to fit my needs, and trying really hard to do it in the simplest way possible. Because plain and simple are good.