Frequently Said Things

I launched LO late-January, 2020. After the first launch, most people's reaction was "Something just doesn't work." And a little later, most people's reaction was "It's so complicated that I can't figure it out." Then people started talking about other things.

So now I've put together what people have said to me frequently:

1. The structure of knowledge is a graph, not a tree

This is true. However

1) The tree structure is a kind of approximation of the real structure, and this is good.
The actual structure of knowledge is certainly closer to the graph than the tree. However, there is some trade-off here, between usability and implementing the exact structure. Unconnected cards are very simple and easy to understand (traditional approach). The graph is closer to the actual structure of knowledge, but it is difficult to handle (actually, I think it's annoying to handle even mind maps digitally). The tree structure is in the middle, which is a good trade-off.

2) Graph-related features may be updated in the future.
The updated version would look like an expanded tree. If the app implements the function that one node can have multiple nodes as a parent node, it would look like a tree, but it would in fact a graph. So the tree structure is also a good starting point for graphs.



2. I feel anxious leaving my data here

If someone uses LO, they would be locked in, and if LO is closed everything disappears. This is an important issue.
I want you to be locked into LO and LO will not close. But if I say it like this, everyone will leave me.

The solution I came up with is to create a feature to export contents to Anki, the most widely used SR program.
With that feature, even if LO is closed, users can move their data to Anki. Or if the user decides not to use LO later, he or she can move written contents to Anki.
(Of course, I don't want anyone to use this feature, but without it, people wouldn't even start using LO, so it's necessary even though I hate it.)

And another solution is to make the LO successful enough that you don't have to worry about it disappearing.

Feel free to send me your ideas and suggestions about improving this aspect of LO.



3. "Obit" means the news of death

I didn't know the meaning of obit at first. My intention was to reduce "orbit".
(and LearnObit-Rabbit, Learn oh beat, Learn a bit)
I think the meaning of obit fits well with the overall mood of LearnObit, so now I think it's good. Some weirdness is good.



4. LO's functions can be implemented with Anki's add on

LO actually started as Anki's add-on idea. You can still see the trace here: https://learnobitlegacy.firebaseapp.com/

Perhaps I could have made the add-on, with most of the functionality it has now, and only leaving the cards showing to Anki.
But there's one thing that really couldn't be implemented within Anki's limits.
I just thought it was annoying not to be able to give ratings to multiple items at once. But by using a tree structure, similar items that have the same expected ratings can be grouped together, and multiple items can be checked at once.
But then, when I want to check each individual item separately, I can't, which is a real problem.
This seemed to be impossible to solve by using Anki, and I eventually released LO as a standalone app.
Now the app, which became independent for this reason has gone a step further and expanded a lot.



5. What is the difference from Remnote?

Actually, Remnote is a product I found and was embarrassed. The key insight, combining SR with a tree-structured note-taking tool, is the same.
But the two apps are definitely different
1) The biggest difference between LO and Remnote is that Remnote doesn't have the Tree2Chunk algorithm. This means that when showing cards, it only shows one item per card.
2) some features are only available in LO:
ChatMemo, Spreadsheet, Multi window, Drawing, Fidge spinner, etc.
3) Overall, the usability of LO is better.
Of course, I'm the one who built LO, and it fits me, so maybe that's the reason why I feel this way. But I do.
4) And most importantly, I don't think they're committed.
Remnote has not had many updates over a long period of time. Perhaps they lost interest or are busy living a daily life. I don't know. But I am committed, and that is the difference.



6. How are the contents of the tree converted as a card?


The last node should be a question and answer pair. It is just shown as a question & answer pair(with what nodes are the node's ancestor). The question part is the question, and the answer part is the answer. That's it. It is very simple. Fundamentally, the tree structure is just for convenience to make taking notes doable*.

This is just a little plus. However, it is possible to extract the properties from the tree, such as dependencies or similarities, and make these pairs slightly different when they are shown, so I'm doing so (e.g., what items should be chunked together, subtracting certain items when the user gets it wrong with another certain item).

*: Actually, making the note-taking experience doable is a very important issue (so it's not "just"). This app tries to solve not only problems with SR, but also with note-taking. The current note-taking experience is so shitty.