I like this. It's a lot like Emacs `diary` system. Though, you can group multiple events under one day heading. I will say that I like how `grep`-able this is. As long as you know the rules, it's pretty easy to parse. You could probably whip up an alarm system in bash and cron/systemd and let it alert you like all those big-box programs do.
Exactly. Some people are looking at the format and thinking "well, this doesn't do much" but they're missing the point: it's a plain text format that plays very well with other common tools such that 'doing more' is trivial.
Sorry to be a Luddite, but isn’t this completely useless?
- one line per event, so good luck finding multi-day events like “Grandma is visiting”;
- rigid metadata (dates, week numbers, weekdays) stored right next to the editable data (events), so copy-pasting errors are inevitable;
- the most important feature of the real calendar software (reminders) is thrown out;
- grepping is really not how most people interact with a calendar.
If you’re ready to ditch reminders, attachments, locations, use the paper diary planner. At least it won’t let you screw the dates with botched copy-paste.
Update: also, sorting by date must be done manually, my god.
I believe software is a form of literacy, not engineering. And I believe everyone should be literate it brings great benefits socially and individually
And if this works for him/her it’s a calendar on a piece of paper stuck to the fridge. Great, it works for them and some people might find it good for them too
> one line per event, so good luck finding multi-day events like “Grandma is visiting”
The format allows for variable granularity and ranges. If Grandma were visiting for a week, it would be fine:
2021-02-20 w07
Right now, the range (start-end) can only be hours, but changing that could fulfil your requirement, e.g.
2012-02-20 w07 Mon-Fri Grandma is visiting
> grepping is really not how most people interact with a calendar.
I don't think the creator ever suggested for one minute that this is a calendar for "most people"! Most people don't use Linux, macOS, or a command line.
> sorting by date must be done manually, my god.
| sort
is not much of a hardship. At least it's possible, unlike a typical GUI app that doesn't support sort.
You're quite right, oneeyedpigeon. The main audience of calendar.txt are people who always have terminal open, and who are using grep, sort and similar all the time anyway.
I use +plustags for multiday and recurring events. So for each line +grandma is visiting, I would add the tag +grandma.
I take similar approach with my courses, +tt for pentesting (from the word in Finnish). I found that for me, creating and validating recurring and multiday events was easier for me. Of course, your mileage may vary.
Smikhanov, you found copy pasting challenging. For me, copy pasting from some dedicated calendar software was a challenge. Copy-pasting with calendar .txt makes it easier for me to keep date, week number, weekday and the event together. And your comment on paper planners was on the spot, I wanted to catch some of their benefits, transparency and reliability with calendar.txt.
It's conflating a format for recording calendars with a syntax the user needs to write.
That syntax looks quite brittle and not very intuitive. One day = one line would get unwieldy fast. It doesn't localise well and there's no obvious way to implement recurring events beyond daily actions. (From what I can see at least?)
You could build a client on top of that but then, you lose the benefit.
It's a neat idea and it's good to see someone share a simple concept that works for them. But I don't see it working for most people.
> You could build a client on top of that but then, you lose the benefit.
disagree. having the data in a simple format makes it syncable, trackable, and guarantees that even if you don't have access to the software you're used to, you can always edit it by hand as long as you can access the file.
you can build some software on top of it; or better yet, build sync plugins for existing software, and benefit from nicer UX without giving up the benefits.
> Unix philosophy. It's one day, one line. You can grep (only show lines).
Surprisingly, grep is able to output several lines, so even if you were to use a (say) more sensible format of one line per event, grep could still output all lines, and display them.
As a file format, this is just fine, but I feel like there's some conflation between a file format and software meant to interact with that format.
I use plain text accounting for my business, and it's lovely to be able to enter bookkeeping data with a plain text editor and keep it all under version control, but the next step is using a program such as hledger or beancount to work with that data.
I'd love to work with plain text calendar software, but it would still need to do things such as provide a mechanism to work on my desktop and mobile device. It would need to handle recurring events automatically, and it would need to allow me to invite others to events and ideally to track when they've accepted or rejected my invites.
I'd need to be able to update, cancel or propose to move events in such a way that others would be updated.
I wish I knew of a good command line tool to interact with CalDav or similar servers, or that I could maintain my calendar in a file format and then handle the synchronization automatically.
Maybe calendar.txt could play a role in that, but on its own, it's not quite enough for me.
For reminders and other automated actions,
pydry suggested "orgzly revived with org mode"; and kruffalon suggested 'remind'. I have not tried those, but they might be better options for notifications and reminding than calendar.txt.
I use calendar.txt on a cell phone with Markor text editor. For my desktop, I just use micro editor on terminal, the same I use for most text editing tasks.
Bookkeeping using plain text files sounds both interesting and challenging.
> Bookkeeping using plain text files sounds both interesting and challenging.
It has quite a bit of a following; I personally would never switch back to some more "complete" application. I believe all the most popular implementations are listed here: https://plaintextaccounting.org/
Not marttt, but I'm a big fan of beancount after using it for the past few years for my company's books. Recycling an old comment of mine that might be helpful: https://news.ycombinator.com/item?id=37937942 (more upthread from that as well)
FWIW, I'm a single-person company with a tiny turnover and only a handful of expense categories. I've used ledger-cli for many years, but at one point, I started to migrate my data to pta [1]. The reason was that in my country (Europe), all reporting seems to implicitly rely on a numbered chart of accounts (something like "101 Cash", "201 Accounts payable" etc), and the journal format of ledger-cli is not built around this idea by default.
pta, however, does this numbering [2], so for filing taxes in my country, its journal format seems more straightforward in case auditors should have any interest in it. I also quite like its terseness (1 transaction = often 1 line, as compared to 3 in a ledger-cli journal -- gives a better overview and is somewhat better to parse further with sed or awk when needed [3]). Finally, it is a single Perl script, thus very portable and lightweight.
All that said, pta doesn't seem to have much of a following, as compared to ledger-cli, which, obviously, is a much more mature project.
There was also an interesting discussion regarding pta vs ledger, involving the author, OpenBSD dev Ingo Schwarze [4] -- well worth a read regarding the general plain text accounting philosophy also. Interesting stuff.
SimpleTask Cloudless has been my favourite todo.txt app for Android. Unfortunately, SimpleTask is in low maintenance mode [1], and does not work well in my setup.
According to mpcjanssen: "The suspending of the applications by Google #1102 and the continuously fixing breakage by Android API changes have reduced my Simpletask motivation to a very low level." [1].
I'm not sure it's even in maintenance mode at this point. I wish I had the skills to maintain it as I quite like SimpleTask and may, even if unsupported, go back to it. Nothing compares on Android.
I’d be interested to read a blog post about your setup. I’m about to start doing my own bookkeeping for my business, and was planning on doing it all in neovim.
pimsync synchronises CalDav to a directory where each event is an icalendar file. You can use khal or any other tool to operate in these files.
This tools is a successor to vdirsyncer, which filled a similar role but with several limitations.
I’ve working in Pimsync for many months now. The basics all work fine, but still working in some extended feature. Feedback is most welcome at this stage.
I used to use remind and wyrd[0] along with remind-caldav[1] for calendaring. I believe I either used remind-caldav or a custom python script set in my mailcap file (I don't remember which it was now) that would open up calendar attachments on emails and save them in remind and remind was sync'd with google calendar.
The details of how I had it all wired up are now lost to memory, but it seemed to accomplish a similar goal to calendar.txt while supporting standard calendar file formats and sync with Google Calendar.
I created a super-simple CalDav-esque web endpoint in my personal website that receives HTTP requests from a command line tool, stores the calendar events, and then returns them as a dynamic ICS file. Then, I use Apple’s Calendar app to subscribe to a “hidden” URL and iCloud automatically distributes the subscription to all my devices.
It sounds super dumb but it works so well.
The main feature is that I can categorize the events (personal, family, work, friends, etc.) and share individual URL’s with other people. Admitedly, I didn’t try solving this problem with Google Calendar, Apple Calendar, or any other calendar service, mostly because I wanted to own my data and also learn a bit more about CalDav itself.
This is close to what I want. I manage all my calendars in radical caldav server and want a sharable calendar that just shows the empty spots between all calendars with minimal detail fo easy scheduling.
It would be nice if Radicale could serve a FREEBUSY URL which takes contents from a your calendars. Some applications support this format and can help with scheduling meetings. I never got to use it in practice though.
I'd be way more comfortable with a VPN or some kind of access-gating for a personal calendar compared to just exposing it on the web and relying on obscurity..
Web calendars don't offer authentication. You have to build it into the URL anyway. If a service I use -- let's say my bank's chequing account, wants to offer a calendar I can subscribe to, I'll be given a URL that looks like https://somebank.com/api/cal?token=abc12345. Anyone who knows that URL can see the calendar as well. No different than my own web app where the URL is https://mysite.com/dev/cal_abc12345.ics.
For a personal calendar, I see no reason to make it any more secure than an obscure URL.
In some ways it worked well, but there were a few drawbacks, and eventually I switched to native calendar programs on desktop and mobile.
Drawbacks I personally felt:
* In the text file, recurring tasks didn't show up when I looked into the future, such as to schedule an appointment.
* Calendar invites over email weren't integrated, so I had to enter and update those manually anyway. (Though one advantage over the native calendar programs I use now is that todo.txt doesn't force the appointment headline someone else wrote into my calendar view, and refuse to let me edit it in my local copy.)
* I had to keep editing dates on tasks manually, every day, for my "current day view" of top of the file to work with priorities.
* No visual calendar views with the tools I was using.
* No device sync with the tools I was using (though possible).
* There are only so many ways in which I was willing to be a weirdo at once, and this one didn't make the cut.
I don't discourage anyone from trying todo.txt or calendar.txt. Just a heads-up of some things you might want to find solutions/workarounds for.
ill tell you a funny, in January I gave my old boss (who passionately hates technology) my text file and told him he only forgot to pay two days in the (hectic) year. (I make one txt per year) It listed what work was done and with who.
His response was WOAH, I've always wanted this! He then send the whole thing to the printer and put it on the wall in his office. Then he added arrows to the days he forgot to pay using a ballpoint pen. He talked a bit about periods in the year and used his finger as a pointing device.
I don't like that week numbers are (documented as) mandatory. I wouldn't see myself grepping on a week number. I'd end up constantly looking up the correct week number in order to annotate each event properly, with no benefit in return. Then again, could write a macro/plugin thingy for one's editor of choice, to fill in the week number for you (ditto for day of week, which I would at least find quite useful, but which would similarly be a right pain to have to fill in manually all the time).
Also: claims to be "one thing per line", yet allows and encourages multiple events on one line (as long as they're on the same day). This is a calendar - is an event not the main "thing" we're dealing with?!
I use a template [1] that has the dates (including week numbers) there already. That way, I only need a text editor.
In my work, similar events happen on the same week numbers in different years. The importance of week numbers probably depends on your area of work.
For me, one thing is a day. That's also the context I need when reviewing and planning. But if I remember correctly, previous HN thread on calendar.txt had a suggestion where someone had an event-per-line format, if you prefer that.
* [ ] YYYY-MM-DD
* [ ] 1800: Dinner with parents
* [x] Walk the dog
* [ ] YYYY-MM-DD
* [ ] 0900: Interview new recruit
* [ ] Prepare interview questions
* [ ] Upload result to HR system
This worked really well for me and was easily searchable. One benefit would be that it is easily trackable in Git.
These days I have a little A6 lined notebook and manually list tasks there. Each page is a new day and the tasks are listed similarly. The only modification is that sometimes I put some letters to theme a task, i.e.:
This is a much nicer format than the one in the linked article.
I think my approach would be to generate a file in Markdown, then import that into a good Markdown editor. You could generate it each week or month from a template, making it easier to add repeating tasks like exercise and dog walking.
I think the format I suggested would parse in markdown, but I created it pre-markdown.
For me the best parts are:
1. You could write a parser very quickly.
2. It's very easy to edit and add to as a human.
3. It can easily be tracked in git or via a diff-based tool.
I did already think about the repetitive tasks, a long time ago I used to have a tool that would look at the past tasks and recommend future tasks (with appropriate distributions). One example was booking a dentist appointment every 6 months, even if I forgot to schedule something it would remind me to book one in, dismiss it or push it back.
This format is terrible for working with line-based tools; you need to go an unknown number of lines back to get the date or the time.
Although, TBH, the OP's suggestion of one-line-per-day puts too many items on the same ling. I would have gone with one line per event. Yes, there's more repetition, but you need no out-of-line context, you have to do a lot less parsing, and you can look at a fixed number of _columns_ and get the beginning of the description of each event, which is nicer for manual searching.
I use a small paper planner that I keep in my pocket at (nearly) all times. People are taken aback and ask why I don't use something more modern, but why should I? This works better, for me, than anything else I ever tried.
It's unfair to computers how well paper planners worked. Among with todo.txt, paper calendars were an inspiration and a benchmark for writing calendar.txt.
Nested list of checkboxes in markdown is exactly how I organize my work days - I write it I every morning, and then review it every afternoon before I power down. It’s basically job-journaling, it helps fix everything in my head day-to-day.
To mark the parsing of the day and to note that any incomplete tasks that I intend to keep have been re-planned to elsewhere. Any marked days are then there for historic purposes only.
> These days I have a little A6 lined notebook and manually list tasks there. Each page is a new day and the tasks are listed similarly.
Periodically my task system fails (the Taskwarrior 3 disaster being the most recent) and this is pretty much what I end up using. But then I forget a bill or something important because there are no bells or buzzers keeping me as a useful working unit. Then I erect the previous system, or some variant, and go with that until some inevitable breakage sends me back to paper. So it goes...
I'm pretty sure the entire world uses ical files. It's a bit of an annoying format to parse, though. The calendar.txt file format seems to be made for people to interact with it directly through a terminal (which the target audience probably already had open).
The first online store/cart I wrote (in php 3 or 4, around 2001), I didn't know sql so I worked backwards into inventing a database from scratch using text files. This was great for a couple years until someone put an asterisk into the name of a product. Asterisk was my row delimiter. lol.
Seriously, with the number of invites, slots, to juggle. Might be better to work again like a dev and with text files in an IDE. Would be funny to send invites with raising a PR. Merging, reviewing, ... what a beautiful mess.
Seems useful for very basic calendar requirements. I'm not seeing much about timezones or anything at all about DST so I'm guessing this format just ignores those for convenience. Saying "times are local to events" is great for physical events but when you're dealing with an international call, you probably want to specify that.
Using a full stop as a separator seems rather limiting, something less likely to appear in an event description such as a vertical pipe would make more sense. Now you'd need some kind of weird logic to write an event titled "read top 10 news.ycombinator.com articles". Using @ as a special character also means you can't store "email support@localhost.com" as an event title.
And, of course, everything is hard coded in English, using English style time notation.
This seems like a fine solution for a personal file format but everyone will probably have to modify it to fit their own needs. If I were to use it, I'd violate the "spec" all over the place by time and date notation alone. This could be fixed by adding some kind of header, but I doubt any full application will ever support a format like this so it's hardly a problem.
If I were to use this as a file format, I'd add headers to store things like language, default time zone, ltr vs rtl, and alter the separator characters. Adding something like a title, an author, and the moment of last edit might also be useful.
I'd personally also probably store events as separate, duplicate lines. That way, you can easily add an event to the bottom of the file without having to find/replace an existing date (or generate an entirely new line). Using basic POSIX tools you can easily get the events back to a single line without making scripts too difficult to read. Assuming culture and other headers match, you'd be able to import another calendar file by simply appending the event lines.
Interesting ideas for formalization and metadata, jeroendh. Some comments on those:
Calendar.txt is indeed a personal calendar. So feel free to violate the spec as much as you need or want to.
My main use for calendar.txt is editing with text editor, and thus, I have not given emphasis to automatic modification.
Time notation is ISO-8601 date and week number. In my view, ISO-8601 is an international standard and not tied to a specific country. I find it convenient to read and sort.
I use local time for physical events. For international calls, I just write down the timezone "13:00Z" or "11:00EET" (update: or mark it in my local time).
At-sign "@" for context was chosen to match todo.txt, as I also use that. As the calendar is personal, it does not need to be fully machine readble. But for your example on email addresses, you could still do it: @WORK email bob@example.com. The context starts with an at "@", email does not.
Full stop "." is easy to type on a cell phone. It's true that a rare char would be more suitable for automatic parsing.
One thing I really would like in a calendar app is some sort of "change log", and this gets somewhat close to that, assuming you track the file in git.
For example, I recently scheduled a dentist appointment 6 months from now. Unless I scroll through the calendar, or specifically search for it, there is no easy way to find that I added that event.
As you guessed, roland35, I do track my calendar.txt in git.
One main benefit from git is also to work as an insurance against (user) errors, and make the whole process much more transparent. When I have used calendars with automatic synchronization, I would have enjoyed a "change log" to make sure that my timezones and meetings are not mistakenly modified by software.
For your dentist example, I would probably just use 'grep' or find in my text editor. After all, six months of days is very little text.
I like text based formats for things. But I find that you need software to help with the maintenance and presentation of such things. Todo.txt and plain text accounting are two examples of this; sure you could edit the plain text but a software system just makes it better to use, especially on mobile.
With that in mind I don't see how this is a replacement for caldav. Sure looking at the plain text of a caldav file is worse, especially one generated by a computer, but at least all expected calendar functionality is included. Though the sprawling RFC is hardly simple to follow or implement.
I like this, but need more detail in my log and have too many things in a given day to have them all on a line. Had the following idea, and will try it out at some point. Call it calendar.md
Use calendar.txt format and method with the following changes:
1) Use markdown, with a top level heading of Calendar (so inclusion is easier) and the portion
2) Use :tag: instead of +tag. Tags can be run together (:tag:tag2:). This helps with Org mode compatibility
3) Third level heading for each event in day, following same format as calendar.txt
4) text under heading is for notes about the event
5) Searching and seeing info on event in day, or summary about day is no longer easy with grep. This is the biggest drawback from not using calendar.txt. Overcome by writing a tool mgrep that is specifically designed to search markdown files in a Markdown aware way (search headings or specific level of headings, show all headings under matching heading or just one level under, show all content under matching headings, search text and show either lines or section text is in optionally along with ancestry of headings).
6) Create CalendarMDMode, minor mode designed to facilitate calendar.md use and editing within Emacs, requires OrgMode, things like shortcuts for new date, new event, in-editor use of mgrep, etc.
142 comments
[ 3.2 ms ] story [ 160 ms ] thread- one line per event, so good luck finding multi-day events like “Grandma is visiting”;
- rigid metadata (dates, week numbers, weekdays) stored right next to the editable data (events), so copy-pasting errors are inevitable;
- the most important feature of the real calendar software (reminders) is thrown out;
- grepping is really not how most people interact with a calendar.
If you’re ready to ditch reminders, attachments, locations, use the paper diary planner. At least it won’t let you screw the dates with botched copy-paste.
Update: also, sorting by date must be done manually, my god.
I believe software is a form of literacy, not engineering. And I believe everyone should be literate it brings great benefits socially and individually
And if this works for him/her it’s a calendar on a piece of paper stuck to the fridge. Great, it works for them and some people might find it good for them too
But yeah.
“One day I will get organisssseed”
The format allows for variable granularity and ranges. If Grandma were visiting for a week, it would be fine:
Right now, the range (start-end) can only be hours, but changing that could fulfil your requirement, e.g. > grepping is really not how most people interact with a calendar.I don't think the creator ever suggested for one minute that this is a calendar for "most people"! Most people don't use Linux, macOS, or a command line.
> sorting by date must be done manually, my god.
is not much of a hardship. At least it's possible, unlike a typical GUI app that doesn't support sort.I use +plustags for multiday and recurring events. So for each line +grandma is visiting, I would add the tag +grandma.
I take similar approach with my courses, +tt for pentesting (from the word in Finnish). I found that for me, creating and validating recurring and multiday events was easier for me. Of course, your mileage may vary.
Smikhanov, you found copy pasting challenging. For me, copy pasting from some dedicated calendar software was a challenge. Copy-pasting with calendar .txt makes it easier for me to keep date, week number, weekday and the event together. And your comment on paper planners was on the spot, I wanted to catch some of their benefits, transparency and reliability with calendar.txt.
The Unix philosophy is often recited as "do one thing, and do it well". This does one thing, but doesn't do it well at all.
It's conflating a format for recording calendars with a syntax the user needs to write.
That syntax looks quite brittle and not very intuitive. One day = one line would get unwieldy fast. It doesn't localise well and there's no obvious way to implement recurring events beyond daily actions. (From what I can see at least?)
You could build a client on top of that but then, you lose the benefit.
It's a neat idea and it's good to see someone share a simple concept that works for them. But I don't see it working for most people.
disagree. having the data in a simple format makes it syncable, trackable, and guarantees that even if you don't have access to the software you're used to, you can always edit it by hand as long as you can access the file.
you can build some software on top of it; or better yet, build sync plugins for existing software, and benefit from nicer UX without giving up the benefits.
> Unix philosophy. It's one day, one line. You can grep (only show lines).
Surprisingly, grep is able to output several lines, so even if you were to use a (say) more sensible format of one line per event, grep could still output all lines, and display them.
I use plain text accounting for my business, and it's lovely to be able to enter bookkeeping data with a plain text editor and keep it all under version control, but the next step is using a program such as hledger or beancount to work with that data.
I'd love to work with plain text calendar software, but it would still need to do things such as provide a mechanism to work on my desktop and mobile device. It would need to handle recurring events automatically, and it would need to allow me to invite others to events and ideally to track when they've accepted or rejected my invites.
I'd need to be able to update, cancel or propose to move events in such a way that others would be updated.
I wish I knew of a good command line tool to interact with CalDav or similar servers, or that I could maintain my calendar in a file format and then handle the synchronization automatically.
Maybe calendar.txt could play a role in that, but on its own, it's not quite enough for me.
[0]: https://xkcd.com/927/
I use calendar.txt on a cell phone with Markor text editor. For my desktop, I just use micro editor on terminal, the same I use for most text editing tasks.
Bookkeeping using plain text files sounds both interesting and challenging.
It has quite a bit of a following; I personally would never switch back to some more "complete" application. I believe all the most popular implementations are listed here: https://plaintextaccounting.org/
Can you recommend a tool and open your accounting workflow, marttt?
Are there any pain points, like filling tax forms or handling travel paperwork?
pta, however, does this numbering [2], so for filing taxes in my country, its journal format seems more straightforward in case auditors should have any interest in it. I also quite like its terseness (1 transaction = often 1 line, as compared to 3 in a ledger-cli journal -- gives a better overview and is somewhat better to parse further with sed or awk when needed [3]). Finally, it is a single Perl script, thus very portable and lightweight.
All that said, pta doesn't seem to have much of a following, as compared to ledger-cli, which, obviously, is a much more mature project.
There was also an interesting discussion regarding pta vs ledger, involving the author, OpenBSD dev Ingo Schwarze [4] -- well worth a read regarding the general plain text accounting philosophy also. Interesting stuff.
1: https://mandoc.bsd.lv/pta/
2: https://cvsweb.bsd.lv/pta/accounts.example.en?rev=1.3&conten...
3: https://cvsweb.bsd.lv/~checkout~/pta/journal.example.en?rev=...
4: https://www.undeadly.org/cgi?action=article;sid=202009281234...
Dpifke, beancounter looks nice, big and well tought out. I bookmarked the docs/ to have a more troughout look at it later.
Update: Had a more detailed look on Beancount.
The end-of-search post is here, with the final choice being GnuCash:
<https://lwn.net/Articles/925782/>
HN discussion (200+ comments): <https://news.ycombinator.com/item?id=36021197>
You'll find at least some of the other posts in the series, and related articles, here:
<https://duckduckgo.com/?q=site%3Alwn.net+accounting+software...>
There is a task equivalent, which also supports caldav sync
https://github.com/pimutils/todoman
According to mpcjanssen: "The suspending of the applications by Google #1102 and the continuously fixing breakage by Android API changes have reduced my Simpletask motivation to a very low level." [1].
[1] https://github.com/mpcjanssen/simpletask-android/issues/1110
This tools is a successor to vdirsyncer, which filled a similar role but with several limitations.
I’ve working in Pimsync for many months now. The basics all work fine, but still working in some extended feature. Feedback is most welcome at this stage.
https://pimsync.whynothugo.nl/
There might be value in checking that out.
Personally I haven't used it in years, but the latest release is from this year so it seems active.
---- [0] https://dianne.skoll.ca/projects/remind/
The details of how I had it all wired up are now lost to memory, but it seemed to accomplish a similar goal to calendar.txt while supporting standard calendar file formats and sync with Google Calendar.
[0] https://wyrd-calendar.gitlab.io/wyrd/
[1] https://github.com/jspricke/remind-caldav
It sounds super dumb but it works so well.
The main feature is that I can categorize the events (personal, family, work, friends, etc.) and share individual URL’s with other people. Admitedly, I didn’t try solving this problem with Google Calendar, Apple Calendar, or any other calendar service, mostly because I wanted to own my data and also learn a bit more about CalDav itself.
For a personal calendar, I see no reason to make it any more secure than an obscure URL.
http://todotxt.org/
https://www.neilvandyke.org/todotxt/
In some ways it worked well, but there were a few drawbacks, and eventually I switched to native calendar programs on desktop and mobile.
Drawbacks I personally felt:
* In the text file, recurring tasks didn't show up when I looked into the future, such as to schedule an appointment.
* Calendar invites over email weren't integrated, so I had to enter and update those manually anyway. (Though one advantage over the native calendar programs I use now is that todo.txt doesn't force the appointment headline someone else wrote into my calendar view, and refuse to let me edit it in my local copy.)
* I had to keep editing dates on tasks manually, every day, for my "current day view" of top of the file to work with priorities.
* No visual calendar views with the tools I was using.
* No device sync with the tools I was using (though possible).
* There are only so many ways in which I was willing to be a weirdo at once, and this one didn't make the cut.
I don't discourage anyone from trying todo.txt or calendar.txt. Just a heads-up of some things you might want to find solutions/workarounds for.
3 days ago, 7 points, 0 comments https://news.ycombinator.com/item?id=43169019
1004 days ago, 202 points, 93 comments https://news.ycombinator.com/item?id=31574125
It also works nicely together with calendar.txt for me. I can 'grep' trough both to see all todos and events related to a project or a +tag.
His response was WOAH, I've always wanted this! He then send the whole thing to the printer and put it on the wall in his office. Then he added arrows to the days he forgot to pay using a ballpoint pen. He talked a bit about periods in the year and used his finger as a pointing device.
https://web.archive.org/web/20250216110151/https://terokarvi...
It's like this, but better. I actually get notifications from my calendar text file. I can set some of them as alarms.
https://www.gnu.org/software/emacs/manual/html_node/emacs/Fo...
Also: claims to be "one thing per line", yet allows and encourages multiple events on one line (as long as they're on the same day). This is a calendar - is an event not the main "thing" we're dealing with?!
In my work, similar events happen on the same week numbers in different years. The importance of week numbers probably depends on your area of work.
For me, one thing is a day. That's also the context I need when reviewing and planning. But if I remember correctly, previous HN thread on calendar.txt had a suggestion where someone had an event-per-line format, if you prefer that.
[1] https://terokarvinen.com/2021/calendar-txt/calendar-txt-unti...
I just create files in a folder with the `yyyy-MM-dd-hh:mm` format. I use `XX` for recurring events.
I use `at` for reminders/notifications.
These days I have a little A6 lined notebook and manually list tasks there. Each page is a new day and the tasks are listed similarly. The only modification is that sometimes I put some letters to theme a task, i.e.:
I think my approach would be to generate a file in Markdown, then import that into a good Markdown editor. You could generate it each week or month from a template, making it easier to add repeating tasks like exercise and dog walking.
For me the best parts are:
1. You could write a parser very quickly.
2. It's very easy to edit and add to as a human.
3. It can easily be tracked in git or via a diff-based tool.
I did already think about the repetitive tasks, a long time ago I used to have a tool that would look at the past tasks and recommend future tasks (with appropriate distributions). One example was booking a dentist appointment every 6 months, even if I forgot to schedule something it would remind me to book one in, dismiss it or push it back.
Although, TBH, the OP's suggestion of one-line-per-day puts too many items on the same ling. I would have gone with one line per event. Yes, there's more repetition, but you need no out-of-line context, you have to do a lot less parsing, and you can look at a fixed number of _columns_ and get the beginning of the description of each event, which is nicer for manual searching.
I think that it's far more important that it is human readable, if it becomes too tricky to edit manually, then just go with a database.
[1] https://unix.stackexchange.com/a/236754
Periodically my task system fails (the Taskwarrior 3 disaster being the most recent) and this is pretty much what I end up using. But then I forget a bill or something important because there are no bells or buzzers keeping me as a useful working unit. Then I erect the previous system, or some variant, and go with that until some inevitable breakage sends me back to paper. So it goes...
Using a full stop as a separator seems rather limiting, something less likely to appear in an event description such as a vertical pipe would make more sense. Now you'd need some kind of weird logic to write an event titled "read top 10 news.ycombinator.com articles". Using @ as a special character also means you can't store "email support@localhost.com" as an event title.
And, of course, everything is hard coded in English, using English style time notation.
This seems like a fine solution for a personal file format but everyone will probably have to modify it to fit their own needs. If I were to use it, I'd violate the "spec" all over the place by time and date notation alone. This could be fixed by adding some kind of header, but I doubt any full application will ever support a format like this so it's hardly a problem.
If I were to use this as a file format, I'd add headers to store things like language, default time zone, ltr vs rtl, and alter the separator characters. Adding something like a title, an author, and the moment of last edit might also be useful.
I'd personally also probably store events as separate, duplicate lines. That way, you can easily add an event to the bottom of the file without having to find/replace an existing date (or generate an entirely new line). Using basic POSIX tools you can easily get the events back to a single line without making scripts too difficult to read. Assuming culture and other headers match, you'd be able to import another calendar file by simply appending the event lines.
Calendar.txt is indeed a personal calendar. So feel free to violate the spec as much as you need or want to. My main use for calendar.txt is editing with text editor, and thus, I have not given emphasis to automatic modification.
Time notation is ISO-8601 date and week number. In my view, ISO-8601 is an international standard and not tied to a specific country. I find it convenient to read and sort.
I use local time for physical events. For international calls, I just write down the timezone "13:00Z" or "11:00EET" (update: or mark it in my local time).
At-sign "@" for context was chosen to match todo.txt, as I also use that. As the calendar is personal, it does not need to be fully machine readble. But for your example on email addresses, you could still do it: @WORK email bob@example.com. The context starts with an at "@", email does not.
Full stop "." is easy to type on a cell phone. It's true that a rare char would be more suitable for automatic parsing.
Thanks for the ideas!
For example, I recently scheduled a dentist appointment 6 months from now. Unless I scroll through the calendar, or specifically search for it, there is no easy way to find that I added that event.
One main benefit from git is also to work as an insurance against (user) errors, and make the whole process much more transparent. When I have used calendars with automatic synchronization, I would have enjoyed a "change log" to make sure that my timezones and meetings are not mistakenly modified by software.
For your dentist example, I would probably just use 'grep' or find in my text editor. After all, six months of days is very little text.
Editing this on mobile sounds very annoying.
`YY-MM-DD:HH:MMa (Optional repeat cron definition) "Event string"`
With that in mind I don't see how this is a replacement for caldav. Sure looking at the plain text of a caldav file is worse, especially one generated by a computer, but at least all expected calendar functionality is included. Though the sprawling RFC is hardly simple to follow or implement.
Use calendar.txt format and method with the following changes:
1) Use markdown, with a top level heading of Calendar (so inclusion is easier) and the portion
2) Use :tag: instead of +tag. Tags can be run together (:tag:tag2:). This helps with Org mode compatibility
3) Third level heading for each event in day, following same format as calendar.txt
4) text under heading is for notes about the event
5) Searching and seeing info on event in day, or summary about day is no longer easy with grep. This is the biggest drawback from not using calendar.txt. Overcome by writing a tool mgrep that is specifically designed to search markdown files in a Markdown aware way (search headings or specific level of headings, show all headings under matching heading or just one level under, show all content under matching headings, search text and show either lines or section text is in optionally along with ancestry of headings).
6) Create CalendarMDMode, minor mode designed to facilitate calendar.md use and editing within Emacs, requires OrgMode, things like shortcuts for new date, new event, in-editor use of mgrep, etc.
7) Attempt to add CalendarMD support to Helix, which is my daily notes editor, using the as-yet unlanded Scheme based plugin system (see https://github.com/helix-editor/helix/pull/8675 )
I have considered similar tools that make use of MarkDown syntax, I feel that there are many tools waiting to be written here.
[1] https://github.com/yshavit/mdq
Based on documentation, you can mdq for the chapter whose heading is "## Foobar"; or extract all lists in chapter "### Xyzzy".
Thanks, abound!
Update: mdq discussion https://news.ycombinator.com/item?id=43152704