- doesn't describe handling spaces or other characters before the checkbox
- doesn't support UTF-8 checkbox emoji
- doesn't support nested lists
- specifies a specific space character when a character class would be better
- not clear what "item must not contain blank lines" means
- description indentation limit of four space characters is a problem for nested items
- description supporting blank lines (properly indented) would be useful for longer descriptions
- date should support an ISO or other standard date format
- timezone is necessary for events happening in specific timezones or across timezones
That's a good thing. Its also unicode, not utf-8. Your plugin can display it with a unicode character. Having a simple ASCII char set helps portability _a lot_. No benefit in supporting unicode code points.
> - doesn't support nested lists
If you can tick all items within a nested list, why bother writing it with dedicated TODO items? Additionally, you can just create more files for elaboration. Your file system is at your dispense. Additionally, that can be implemented in your plugin of choice. Goes with your first point in hand.
> - specifies a specific space character when a character class would be better
ASCII has no official character class. If your are refering to tabs; I don't think this is bad or good. Maybe you can follow uo on the use case..
> - not clear what "item must not contain blank lines" means
Probably that a blank line is used as a separator
> - description indentation limit of four space characters is a problem for nested items
I don't see why. The fixed amout of four characters obviously lines up to the start of the text from an item.
This helps readability a lot.
And is a simple standard for identation rules.
> - description supporting blank lines (properly indented) would be useful for longer descriptions
I disagree. A TODO list should be simple to grasp. Details and elaborations can be put in a dedicated directoy and refered to.
Also, it is a _very bad_ idea to work with invisible lines. Users are dump and quick to judge.
> - date should support an ISO or other standard date format
This is the most portable standard, but extended for human editing..
> - timezone is necessary for events happening in specific timezones or across timezones
I tend to agree. But the person hosting the list should probably be the one defining the time zone. But I haven't read anythinf about time zones in the primer. Nothing stops one to simply append it to a time. No need for the standard to define such.
I think it is well put and the first one which checks all the boxes for me.
Congratz to the author.
Sticking to ASCII in this day and age is nice and fun if English is your mother tongue, but it's a big middle finger to the rest of the world. Please don't.
You can use Unicode characters in item descriptions, so you can write these texts in Japanese, Finnish, or Greek. Only the “syntactical elements” (like checkboxes, priority, due dates) are made up from ASCII characters, to ensure that they are easy to type.
Perhaps the spec could allow an implementation to support a run-time option to recognize alternate characters for syntactic elements. This would have the downside that one person's data might not be shareable with another person's [x]it installation. But that's not necessarily a bug, especially to people who don't plan to share their todo lists.
This run-time-option-only approach would be much better than the "be liberal in what you accept and strict in what you emit" philosophy that tends to cause fragmentation on lots of different levels.
> Perhaps the spec could allow an implementation to support a run-time option to recognize alternate characters for syntactic elements.
Please don't. The equivalent in programming languages would be localizing the syntactic elements ({}, (), [], '.', etc). Its much simpler for a language ecosystem to have the same syntax regardless of the programmer's language.
As far as I know, international keyboards can still type these characters just fine.
You are right, I didn't wanted to imply that these lists can only be used by speakers fluent in english.
I wanted to imply that grammars of languages are best represented in ASCII.
There are languages written from right to left or top to bottom. No standard I know of are supporting such flexibility in syntax. And it shouldnt be necessary - if the <user text> items within the grammar support unicode but the keywords are ASCII only it can MORE easily adapted than supporting unicode....
-- I don't disagree with your basic point, but I like to be included :). There's a Lot of us around whose native / mother tongue is not English, but we prefer it for computing tasks (and frequently gaming / movies / entertainment too - Forums are littered with folks complaining of having to "enjoy" a horrible localized version of something, instead of the original English)
Thanks for sharing your notes. Nesting of items is not supported, although the idea has come up. It’s on hold right now – that’s mostly due to practical reasons, because it’s relatively hard to implement in tooling.
Allow me to clarify a few details regarding the spec:
- There can’t be any character before the checkbox. (An item “MUST start at the beginning of a line with a checkbox.”)
- A blank line is defined as “a line that is either empty, or that exclusively consists of blank characters”, and a blank character is defined as “a character from the Unicode Space Separator category (Zs)”. So the description text of an item (that can span multiple lines) cannot contain a line that’s all blank. Reason is that a blank or empty line separates item groups from each other, so allowing it to appear within descriptions could create visual ambiguity. (Even though it wouldn’t be ambiguous from a purely formal standpoint.)
- The format for dates is a subset of ISO-8601, except for the slashed variant (2020/12/31) that’s there for convenience, and the notation for quarters (2022-Q1).
It would be simpler to just use the ISO date format. Otherwise people have to implement their own date parser for your stricter format, which adds code (and therefore bugs) and increases probability that people will incorrectly implement the spec.
Why wouldn't you allow space before the checkbox? If somebody accidentally has a space because it's not obvious in their editor (esp. if items can be separated by two newlines) what happens? Undefined behavior? Your format looks a lot like Markdown so I wouldn't have assumed this was a spec requirement unless I looked closely (most people don't look at specs closely)
> It would be simpler to just use the ISO date format.
[x]it! is supposed to be edited (editable) by hand, and from an end-user perspective, I personally find ISO-8601 in its entirety relatively complex to grasp. The currently supported date formats aim to be a tradeoff between “flexible enough” but “not too complex yet”.
> Why wouldn't you allow space before the checkbox?
To enforce a tidy layout of the file. Whitespace before the checkbox (i.e. on the same line) is undefined behaviour. The idea is that editor plugins help users with compliance of the rules. For example, I think that all available [x]it! editor plugins would highlight it if there was illegal whitespace in front of a checkbox.
