Extending the notion of siblings to notes

First, I want to make clear that if you agree with the feature request, I’m willing to create a design doc and implement it myself. But I would not do it unless I know it’s not time lost.

I’ll tell you what I want first, and then the problem it solves.

My solution

Essentially, I would want to be able to state that notes are siblings, and when we bury siblings, not only we bury the cards of the same notes, but we bury card of sibling notes. There are multiple ways to do it, but I believe the easiest solution would to have an optional field that is designated as sibling field. And if the sibling field is non-empty, each note with the same value as sibling field are also buried. I could even ask for multiple key, separated by comma, so that a note can be in multiple siblings relationship.

But I could also offer a way to do it from the UI: selecting multiple notes and having an internal field set so that we know they are siblings. Which admittedly would make more complex to show to the user which notes are siblings. And ask the question of what to do if a note N1 is sibling of N2 and you set it as sibling of N3. Should N2 and N3 become siblings or note. We would need a way to show the notes that are siblings of N1 in the UI. While, with my first proposal, it would be easier to use current search tools to find siblings.

My problem

One of the problem I have in anki, which can’t easily be solved with add-ons due to the lack of add-on on mobile, is that the notion of siblings is too restrictive. I very often have a lot of related cards that I really don’t want to see the same day. I end up creating very big note types, that tries to cover so many related informations, that should actually end up being in different notes.

Let me give you an example. I’ll take a very simple one from basic geometry. Let’s say I want to learn about squares, and I’ve the following theorem: “Let Q be a quadrilateral. If Q is a diamond and if Q has a right angle then Q is a square”. Then I’ll want the following related cards: remember the first hypothesis “Q is a diamond”, remember the second hyptohesis “Q has a right angle”, remember the conclusion “Q is a square”. But also “Show an example of quadrilateral which is not a diamond, which have a right angle, such that Q is not a square” and “Show an example of diamond without a right angle such that Q is not a square”. And also “show the proof that if Q is a diamond with a right angle then Q is a square”.

Admittedly, this is very basic level maths. But for actual algebra, analysis, this kind of questions where I want to learn the construction of the proof, where I want to learn examples showing why each hypothesis is relevant, is something that occurs all of the time. Naturally, I guess I should have a note type for proof, a note type to show why each hypothesis matter. A note type for the actual theorem… Instead, I currently have a single note type simply to ensure that I don’t see more than a card related to this theorem each day.

I could have hoped that after 9 years of using anki, my note type would cover all of the case. But currently, it does not. Because some theorems are of the form “A implies B implies C implies D” and if I have a card for the proof of “A implies B” I don’t have one for the proof of “B implies C” or “A implies C” or anything related to D either. Thanks to my add-on that creates card type from my own template language, I could make them easily, but I don’t think it would be wise to add more card types

2 Likes

I would like this too.

This has been asked for before:


A better way than using a plain field is to basically duplicate the tag system and call it “families” or something like that. Notes with the same family generate siblings.

But there should be a better way, one that avoids the user having to manage “family names”.

1 Like

Yes this idea was rejected by the official team.

I think the best solution is the add on, and syncing from the desktop first thing in the morning.

https://ankiweb.net/shared/info/1916261072

I have a similar problem for language learning. Let’s say that there’s a rule that has various aspects that I need to learn. I currently use a single, potentially very long note to create questions that present the basic principle and test my knowledge of its various aspects.

For example, if there’s a rule that says an infinitive in a particular form, say its ending is -xy, will have first person, present tense conjugations in one of two forms, say their endings will always be either -ab and -cd, I’ll have note that asks a question about the general rule and poses individual questions for each application of the rule, often going both ways. E.g., Q1: If a verb infinitive ends in -xy what could be its first person, present tense ending? A1: -ab or -cd. Q2: If a verb’s first person, present tense ends in -ab, what is its infinitive’s ending? A2: -xy. Q3: if a verb’s first person, present tense ends in -cd, what is its infinitive’s ending? A3: -xy.

The idea here is that I want to have all the questions relating to a concept in one location so that I can a) determine if I’ve actually asked all the questions I need to in order to learn the concept, b) easily find those questions if I want to modify them as well as c) avoid creating notes with duplicate questions.

So, your ideas could be useful to me. Although from your presentation of those ideas, it’s clear that I have to learn more about siblings.

I would like this and think that despite the official team´s rejection of this in the past due to concerns over misuse there are a number of valid use cases for this feature.

One other for language learning would be where you have notes for different word types, eg. a verb, a noun, an adverb, etc that have the same root. It is desirable that cards for notes with the same root are not shown in the same session.

Whilst I would not make such notes for all words, for me the point of doing this for a number of words is to embed the rules for how words are derived from the root or each other (eg. adding -ly in English).

The alternative of having a single humongous composite note type risks hitting some of Anki’s internal limits, and is painful to maintain.

Rather than associate multiple notes, you could put more related knowledge in each note.

There’s an add-on (Yukogurafu) to link together concepts with relations into facts, such that every little topic fits in one note. For example, to learn the geometric theorem ArthurMilchior suggested:

It might be a bit messy, but I ended up doing this (I don’t really want to use addons since they break or stop being mantained, etc.):

  • Created separate fields for cloze deletions:
    cloze_1, cloze_2, cloze_3, etc.

  • Added to the front template:

{{#c1}}{{type:cloze:cloze_1}}{{/c1}}

{{#c2}}{{type:cloze:cloze_2}}{{/c2}}

{{#c3}}{{type:cloze:cloze_3}}{{/c3}}

{{#c4}}{{type:cloze:cloze_4}}{{/c4}}

{{#c5}}{{type:cloze:cloze_5}}{{/c5}}

<hr>

{{cloze:cloze_1}}

{{cloze:cloze_2}}

{{cloze:cloze_3}}

{{cloze:cloze_4}}

{{cloze:cloze_5}}
  • The back template:
{{cloze:cloze_1}}
{{#c1}}{{audio_1}}{{/c1}}


{{cloze:cloze_2}}
{{#c2}}{{audio_2}}{{/c2}}

{{cloze:cloze_3}}
{{#c3}}{{audio_3}}{{/c3}}

{{cloze:cloze_4}}
{{#c4}}{{audio_4}}{{/c4}}

{{cloze:cloze_5}}
{{#c5}}{{audio_5}}{{/c5}}

This way I can put similar sentences in the same note, but when I answer any of them, the rest behave like proper siblings.

This is how it works in practice: Idioms

Graphs like that will get too complicated with more complex information. That system also doesn’t work nicely with formatting, media, and customization (the text has to be “incomplete”). Your example doesn’t even even test the two hypotheses separately.

This looks interesting. Re complexity, that’s something the user can control, I imagine? Ideally, the graph would deal with all possible interrelationships. But since this isn’t practically (and perhaps even theoretically) feasible, it’s up to the user, it seems to me, to carry things through to a stopping point acceptable to her/him. The danger of an add-on becoming an orphan is also a concern, but I think that this add-on is nonetheless worth exploring.