Zettelkasten, Evergreen Notes, digital garden, second brain, etc.

847 views
Skip to first unread message

bimlas

unread,
Sep 19, 2020, 2:39:20 PM9/19/20
to TiddlyWiki
I’ve been busy with these topics lately and as I look at, this topic is becoming more and more interesting for many, so I’m opening a topic where we can talk about these. Although this is the TiddlyWiki group, keep in mind that Zettelkasten was originally implemented on paper, so the topic of conversation is an implementation that can be implemented anywhere, regardless of software.

The purposes of Zettelkasten and similar knowledgebases are:
  • Easily recall your long-forgotten thoughts
  • Clarify your knowledge of a particular topic
  • To notice contradictions and shortcomings during the dialogue with the Notes
  • Create new ideas from your existing knowledge
Related topics:
I think TiddlyWiki empty edition is exactly the software you need for a Zettelkasten-type note collection:
  • Displays "physically" separate notes that we can handle as if they were made of paper: you can leave all of them on "the desk" that you want to manage right now and you can put the rest in "the drawer"
  • It basically works by not support for renaming tiddlers, which fits the Zettelkasten mindset (permanent titles)
  • Backlinks can be easily displayed in the tiddler info bar
  • It does not include table of contents, which may be due to link-based navigation
  • Tags are basically only used for grouping (if ToC-type usage is ignored)
  • The selected text can be extracted into a separate tiddler and replaced with a link (refactoring of notes)
The main principles of the "second brain" kind of knowledgebase in my opinion are:
  • Write notes about everything to make sure that thoughts and experiences are not lost, in addition, the wording helps to understand the topic
  • Add a unique ID to your notes so you can clearly identify them even with a text search (so you get a list of backlinks)
  • There should be only one well-defined idea on a note
  • Because of link-based navigation, use links instead of text search to look for a topic
Wording helps with understanding, thus instead of copying someone’s writing, opinion, solution, describe it in your own words, because only then will you become aware of what you actually understood while you have to articulate, “have to say back”.

If you just copy, you gather the information without actually interpreting it, but while writing the text, you need to interpret your thoughts so that you can put them on paper in a meaningful way. This will shed light on the dark spots, points you don’t really understand yet, and it will encourage you to gather more information on that topic (even during a dialogue with your own notes), so it will lead to real understanding. It will become your habit that instead of scanning and reviewing the text quickly, you will actually start reading the text, interpreting what is described, so your reading efficiency will also improve.

It is necessary to use IDs in paper form to identify the notes because you have no other option. In digital form, each of your notes actually has a unique ID (path, in-database ID, URL), but if you use this, you will depend on the implementation (vendor lock-in): if you want to migrate to another system where links are marked differently (e.g. the title identifies the note instead of a generated ID), it will be difficult to migrate (if it can be solved at all). Therefore, it is worth using a notation system that works in all circumstances (even with a simple text search), e.g. the form "2020-09-19_20-24". There is no line break in the unique ID (which would make it difficult to find with a text search), it does not depend on character encoding (so it even works in filenames, it does not need to be modified in URLs because it does not contain accented letters), so it is a universal solution.

However, since this name is not verbose, it is a good idea to display the title of the note along with its ID. It is a software-dependent solution, but the best solution is, for example, if the note can be identified by an ID, but we also display the title in the text in the search results.

A note should be on a signle topic and be as short as possible; if it’s already too long, it probably includes more topics.

When you put headlines in a note, it is a clear indication that the note is about multiple topics (e.g., a Mammal note has a Dog and Cat headline). Put these paragraphs in a separate note and just place a link in their original location (note refactoring) so that if you use text search, you're sure to find what you're looking for - if the text you're looking for were under a headline within a note, you wouldn't find it in the search results (e.g. you would search for the word "dobberman" and only Mammals would appear in the results, you probably wouldn’t consider it, but you would already notice the title Dog).

Creating connection between notes requires nothing but links and backlinks (see details) because they can be used to implement tags (details), fields (details), and table of contents (details). It follows that when using a wiki-like annotation method, we can search for a topic by following links instead of text search, because we can find more accurate results that we think are as opposed to having to search a list of notes that contain those words (must to read details about the weakness of full text search).

Backlinks are only useful if they show really relevant content. For example, on Wikipedia, the list of backlinks to Wiki page contains a bunch of unrelated information:
  • The Aeronautics page mentions "Wiki" in the footnote in connection with a link
  • Batman also mentions it in the footnote
  • I could list, but this is the case in most places
So I think we need to choose when to refer to a word and when not. We need to stick to the Zettelkasten principles to first describe what we want, but don’t place links in it until we’re completely done with the note: we visit the related notes and only link to the really relevant places. For example, if I write a note about squirrels, it would not be worthwhile to link from that note to the note about mammals itself, but only to the note that collects rodents. While this is more time consuming than linking to key words on an ad-hoc basis, a list of backlinks will be really useful if you don’t have to sort between them because they are all really closely related to that particular note.

One of the advantages of this kind of note-taking, for example, is that it took me about 15 minutes to write it all down. In fact, all I had to do was translate my previously written and thought-out notes into English and make small changes to it.


I ask everyone who is also interested in the topic to join the conversation.

Saq Imtiaz

unread,
Sep 19, 2020, 4:20:39 PM9/19/20
to TiddlyWiki
Hi Bimlas, 

Just a quick note to say that I quite enjoy reading your thoughts on this. As someone who doesn't really take notes, I don't have much of substance to add. However if I were taking extensive notes the idea of an approach that relied on full text search or links rather than fields/tags would quite appeal to me.

Oh, I couldn't get your link about the weakness of full text search to load for me but this works:

Cheers,
Saq

Saq Imtiaz

unread,
Sep 19, 2020, 4:23:05 PM9/19/20
to TiddlyWiki
PS: I should clarify that I misspoke when I mention the appeal of fulltext search and instead meant to refer to the appeal of the entirety of the note being self-contained in the text as opposed to other fields.

TW Tones

unread,
Sep 19, 2020, 7:57:46 PM9/19/20
to TiddlyWiki
Bimlas,

Thanks for this clears and to the point information. "You have my ear". 

A few comments
  • Building non trivial links through thoughtful note taking is a wise approach. 
    • This reminds me of before Google Search, "curated lists/directories", and Australian one is still going https://www.blackstump.com.au/, the key is a Human curates, not an algorithm
  • To me knowledge and information capture and management does require thoughtful reflection, 
    • as I have argued elsewhere, when you are entering something that is, the best occasion, because you are aware at the time or context and place more than you will read later, failure to capture key information at this point will result in loss
  • I believe we could use your guidelines for a "Zettelkasten" system in tiddlywiki, but if these notes were clearly identified and the required practices clear, there is no reason, in my mind, that other organisational or network principals can't be applied to the same wiki and content. There is often value in hybrids and tiddlywiki allows it. You can build restricted searches etc... 
One method perhaps available in tiddlywiki but not "Zettelkasten" is the idea of full text titles. 
  • Assuming you keep to a single point as in  "Zettelkasten" there is no harm using the text, as a the title to a tiddler, the key is unique by the content/title, and if it were not you need to know.
  • All TiddlyWiki title tools come into play
In a related area I am considering Title Generation
  • My thoughts go beyond those in the topic already
  • The idea is there are a range of ways for tiddler titles to be proposed from the content, context, tags and more.
  • Once a tiddler gains a unique ID its trivial to rename (and the relink plugin is a big help)
  • My idea is an AI style title generator but this could be also applied to extracting and storing keywords or captions
    • Eg words with 7+ characters could be listed from which you tick those you consider appropriate keywords for the text you entered and these perhaps inspire more
    • You could have a limited set of categories appear according to context and content
Regards
Tony/Tones

TW Tones

unread,
Sep 19, 2020, 8:17:38 PM9/19/20
to TiddlyWiki
Bimlas,

To response to the original thread, requesting books on note taking I would like to suggest something a little out of left field.
The Checklist Manifesto apart from its own inherit value, it suggests away to develop new habits and learning, like curation methods

There is far more to this that one could consider by the title, it demonstrates not only the value, but when and how to use them. It covers a lot of real world cases and disasters as a result. Key points I remember;
  • Check lists should not be every step, but every step that is critical or likely to missed ONLY
  • Effective checklists are "second brain" components where you can safely delegate, supporting quality assurance
  • Checklists should evolve and there a methods we can implement within tiddlywiki that supports this
  • Checklists can be adaptive, they could even respond to search, and analytics data eg; you never open a set of sub steps, because you recall them, so in time they are hidden, and visa versa.
  • The dismissal of checklist by experts, who think they know everything, and refuse to follow an "effective checklist", have killed people.
  • Korean Piolet's must use English, because the Korean language has a "deference" to your "superiors", that co-pilots could not tell the captain when they were wrong. This shows how a check list can act as an independent authority that hierarchies (or fatigue) can not overpower. 
Regards
Tony

On Sunday, 20 September 2020 at 04:39:20 UTC+10 bimlas wrote:

TiddlyTweeter

unread,
Sep 21, 2020, 8:01:38 AM9/21/20
to TiddlyWiki
bimlas

That is very good. Very needed.

It is BIGGER than TW.

The net GENERALLY is way crude. Your points are generic, not TW specific.

I think you could BOTH post here (like 3-6 responses) and elsewhere (more than 6).

I DO like idea of using TW to DEMONSTRATE different approaches (it is agnostic) but also that YOU get read by a much larger audience. 
Your kind of thinking needs a bigger context IMO.

Best wishes
TT

bimlas

unread,
Sep 21, 2020, 8:47:46 AM9/21/20
to TiddlyWiki
TiddlyTweeter,

Thank you for the kind words, but these thoughts are not quite mine, but of Niklas Luhman, famous for Zettelkasten, for example. I just experienced and agree with them, so I “preach” about them. But let’s see what others have to say about these thoughts!

Reply all
Reply to author
Forward
0 new messages