Thanks for pointing me to Typora and VNote, @Peter! I'll check out at least Typora. Looks appealing
But the one thing I really, really don't want to miss anymore after my experience with Roam, is backlinks. Automatic backlinks to be precise. It's really what makes a difference when building a zettelkasten. To me at least.
Plus - I guess I have to add - the ability to link to each and every paragraph (block) in Roam.
Both features taken together make it so much easier to weave a web of zettels (or should I say thoughts?). Because I don't have to think about where to put a thought. Should it become a separate page? Should it be nested inside some existing page? Is an effort to set up a backlink justified? That's all questions I don't feel burdened with anymore.
Zettels to me have become something I write down like anything else over the course of a day. They are anchored naturally to a date, i.e. to a context. But that does not tie them up. They show up in a search, I can reference them from anywhere, I can refactor my "daily documentation" (of which zettels are a part) at any time very easily (extract a zettel into a separate page, inline a zettel into an existing page, "rename" a zettel).
I don't mind to get all this with a tool which stores my "daily documentation" in local .md files in my dropbox - as long as I get the same functionality for weaving my web.
I'd even argue that should not be difficult to build. No need to host it on the web. No need to scale it for thousands of users. I'd be happy to do it myself - but I'm not the UI guy for today's fancy desktop programming technologies.
Simple Markdown editors don't do the job for me, though. They are not smart about the purpose: building a second brain.
Markdown is a great start - but without a data model underlying Markdown texts I don't think they will be powerful enough. At least for my requirements.
Roam is doing a good job in the regard. It pretty much looks like a Markdown editor - but that's only a user interface on top of a database, I'd say. And I guess that kind of software architecture is the key to Roam's features.
@ralfw : it certainly was not my intention to get you off Roam, so to speak. I pointed out Typora and VNote in case you were interested in a markdown editor.
Like I said, VNote is the more advanced of the two, with more features. Regarding backlinks, if you do an Advanced Search in VNote, e.g. looking for note number 123987, you get a list with all the notes that have that number in their text, i.e. through links. So, apart from the original note with number 123987 in that list, all the other notes listed are effectively notes backlinked to note 123987.
I don't quite understand your comment about the simple markdown editors. I have seen 1 tutorial about Roam and know it is very different from a markdown editor, a lot more complex. All I can say is that Typora is "just" a markdown editor, whereas VNote is a markdown editor cum note-taking manager. Instead of checking out Typora I'd recommend checking out VNote, if you are interested. Don't do it to be kind to me or not to hurt my feelings. I shan't be hurt if you aren't interested in either of the two: I know those who use Roam are absolutely crazy about it.
Again, I am not trying to get you to drop Roam. Also, I have no commercial or other interests in either Typora or VNote, so I am not "talking my book". I was just after an exchange of ideas
Don't worry, @Peter, I didn't think you were trying to get me off of Roam.
I took a look at Typora and like it for its simplicity. And it's great it lets me so easily include images from the clipboard even with storing them right where I want them in my local folder hierarchy.
And maybe some other time I'll check out VNote.
But what you mentioned as a feature of VNote (and The Archive, if I remember correctly), that you can search for some note ID and find all the notes containing it, that's exactly what I'm not (!) looking for.
I don't want to deal with IDs anymore. And I experienced how convenient it is to not even have to think about files (or pages or documents) anymore.
When you experience something your mind is able to connect to any part of that experience. You don't need to break it up into certain chunks. It does not need to have a particular granularity.
Looking at a computer I might remember a certain book about micro processors I read when I was in my early 20s. That book then reminds me of my bookshelf in my room at home with my parents. And that bookshelf in the context of the micro processor book then might evoke a memory of me and my father working on a electronic tinker kit when I way maybe 7 years old. And that fond memory then might lead to an even earlier moment us looking at the stars.
Looking at this chain of memories from the other end: When we were watching the stars I did not need to think about how to break up and store the then current experience. What I saw, heard, felt... all that was woven into my brain/body in a way so that future experiences could relate to it in any way, to details or the overall experience, to a smell or a sight.
That's how I'd like to work today when I log a thought, and experience. The only thing I want to do is, write it down, paragraph by paragraph. And then rest assured that I can connect to any part of it at any time.
The atoms of such recording to me are the paragraphs. That I find convenient. They come naturally to me as units of text with a certain meaning.
Paragraphs can be put in sequence and can be nested (in Roam). Thus they end up in a context. Linking to an already exiting paragraph pulls something into another context I'm currently in - and lets me jump to the context the former paragraph was written in.
Adding a tag to a paragraph or isolating it into a page of its own to me are optimizations. If I can do that to add even more context/meaning, that's great. But the basic building block to me right now are just paragraphs not otherwise attributed with meta data.
I did not know that this would become so important for me. But that's my current experience with Roam. That's what's making working with it fundamentally different.
That said I don't want to sell Roam here. It just happens to be the tool right now implementing a certain approach to recording "thoughts".
I don't care about products or manufacturers. I care about convenience. How much do I feel supported in pursuing my interests without getting distracted by infrastructure or technological idiosyncracies?
As it turns our, it's a journey to come ever closer what suits me best - probably never ending And from what you wrote about how you started with Evernote and now are using two tools, it's somewhat the same for you. Let's wait and see what 2022 will bring...
This is just a meta-commentary to say that this conversation between @ralfw and @Sascha is an excellent example of what @ctietze said here on the event of the anniversary: this community is very strong because of our ability to have debates such as these in an enthusiastic yet civil manner. As someone who has no opinion here (being a strict anti-dogmatist about theory), I read through this thoroughly just to see two people have a disagreement in such an agreeable way.
But what you mentioned as a feature of VNote (and The Archive, if I remember correctly), that you can search for some note ID and find all the notes containing it, that's exactly what I'm not (!) looking for.
I don't want to deal with IDs anymore. And I experienced how convenient it is to not even have to think about files (or pages or documents) anymore.
Looking at this chain of memories from the other end: When we were watching the stars I did not need to think about how to break up and store the then current experience. What I saw, heard, felt... all that was woven into my brain/body in a way so that future experiences could relate to it in any way, to details or the overall experience, to a smell or a sight.
That's how I'd like to work today when I log a thought, and experience. The only thing I want to do is, write it down, paragraph by paragraph. And then rest assured that I can connect to any part of it at any time.
The atoms of such recording to me are the paragraphs. That I find convenient. They come naturally to me as units of text with a certain meaning.
To me, this is the crux of your thinking and how you want to build and expand your knowledge base. And whether it is called a Zettelkasten or anything else is not important; besides, it is your right to call it whatever you wish. What is important is whether the tool works for you in the best possible way.
Paragraphs can be put in sequence and can be nested (in Roam). Thus they end up in a context. Linking to an already exiting paragraph pulls something into another context I'm currently in - and lets me jump to the context the former paragraph was written in.
That is nothing exceptional, it is possible with other apps like The Archive, VNote too, surely?
I did not know that this would become so important for me. But that's my current experience with Roam. That's what's making working with it fundamentally different.
That said I don't want to sell Roam here. It just happens to be the tool right now implementing a certain approach to recording "thoughts".
I don't care about products or manufacturers. I care about convenience. How much do I feel supported in pursuing my interests without getting distracted by infrastructure or technological idiosyncracies?
I agree. And to put that in my context, my knowledge base consists of 2 parts:
Geopolitics
Concepts-Thoughts-Ideas
The 1st is mostly (historic) facts-based, with interpretations, and not really a Zettelkasten tool.
The 2nd one, however, is a Zk-type tool, and I have often wondered how, apart from the uni-directional links, I could get a more dynamic, interactive tool like you describe above. I looked at Roam and dismissed it as more of a task manager than a dynamic note-taker/manager. Your comments show that I was wrong, have given me new food for thought and perhaps I should look at it again
The only other drawback of Roam I see at this stage is future portability. If the dev decides to quit, you can of course continue to use it. But what if you wanted to switch: what format are your notes in and can they be read into another app? With markdown or plaintext notes there is no problem, but with Roam notes?
Paragraphs can be put in sequence and can be nested (in Roam). Thus they end up in a context. Linking to an already exiting paragraph pulls something into another context I'm currently in - and lets me jump to the context the former paragraph was written in.
That is nothing exceptional, it is possible with other apps like The Archive, VNote too, surely?
Is that possible? I'd love to see, how to do that.
Here's what I mean: I write a text and structure it into paragraphs. Or think of it as a hierarchical list of bullet points.
some thought here
some nested thought
another one
a second thought there
a nested one pertaining to the second thought
And so on...
And all these paragraphs can be referenced from other paragraphs in other "documents" (.md files or pages)? Without adding any meta-information like a paragraph number? And that produces a link back to the paragraph?
Here's how it's done in Roam:
No paragraph is out of reach for reference. No provisions have to be made while writing a paragraph.
That's the convenience I'm looking for. It keeps me from constantly switching between the real and the meta
The only other drawback of Roam I see at this stage is future portability. If the dev decides to quit, you can of course continue to use it. But what if you wanted to switch: what format are your notes in and can they be read into another app? With markdown or plaintext notes there is no problem, but with Roam notes?
You can export all your Roam "notes" to .md files at any time. In fact I'm doing that almost every day now while Roam is in its early stages of development. Just to be safe It's a matter of 10 seconds.
@ralfw said:
And all these paragraphs can be referenced from other paragraphs in other "documents" (.md files or pages)? Without adding any meta-information like a paragraph number? And that produces a link back to the paragraph?
No, that is not possible, you are right.
No paragraph is out of reach for reference. No provisions have to be made while writing a paragraph.
That's the convenience I'm looking for. It keeps me from constantly switching between the real and the meta
I agree, that is actually how it should be. I have not had that real Eureka feeling from any tool (yet?).
You can export all your Roam "notes" to .md files at any time. In fact I'm doing that almost every day now while Roam is in its early stages of development. Just to be safe It's a matter of 10 seconds.
In principle I am against having to export to be able to pick up my notes. Then again, it could be an acceptable compromise if a tool is worthwhile.
As for those 10 secs., I should imagine that depends on the number of notes, or rather, on the size of note(s) / file(s) to be exported.
BTW, I notice that you need to sign in to the site: does that mean Roam is web-based only, i.e. no app on your computer?
@cobblepot said:
I asked CWS about this and believe that they currently can, although user encryption is a long-term goal.
Thank you for that feedback, @cobblepot.
It is a real pity that they do not have a desktop version because from what I have seen over the past 2 days this app could blow all the other Zettelkasten-type note-taking apps out of the water, and I for one would certainly want to get stuck in, so to speak.
But to entrust one's notes to an unencrypted 3rd party is simply not acceptable to me, and I am sure to a lot of other people. And that has nothing to do with having something to hide.
I am surprised someone like @ralfw accepts it.
Since you‘re addressing me personally, @Peter, let me just say:
I‘m taking the topic of data security not lightly. But after much thought I have come to certain conclusions. One is, that it‘s not a matter of black and white.
For storing my passwords I‘m taking different measures than for storing my zettels, for example.
I‘m stuck in a country which has shut down public life in an instant. Politicians are demanding to track infected people. What‘s next? Forced testing for infection?
Let me say, these days I‘m not really concerned about zettels in some database which potentially people could read and derive from what I‘m reading and thinking. Obviously they don‘t need that to force me and millions others into a standstill.
I‘m not worried about a government (agency) (or company) which spies on my data. I‘m worried about those who don‘t need that anymore.
That said, please let‘s not turn this thread into a discussion about security consciousness etc. I‘m happy to read you took a second look at Roam and seem to like it more. And I agree: a desktop version storing data locally would be even cooler.
Storing data locally and syncing them between devices doesn‘t sound all to difficult too me. But for whipping up a „prototype“ it might have been easier to go web-first.
Two days ago I had never heard of Zettelkasten
(nor Luhmann, I am ashamed to admit).
I discovered Sonke Ahrens book entirely by chance and, even though I have only read about 10% of the book, I have already jumped down the rabbit hole of devouring all the blogs, medium articles, youtube videos, and anything else related to Zettelkasten I could get my hands on. It's one of those incredible moments in life when you don't realise there is a solution to your problem until you suddenly trip right over it whilst walking (or in my case, browsing Amazon!). In this case, the 'problem' was finding a better solution to my personal knowledge management and book writing systems.
2 days later and I now have accounts on Notion and RoamReseach - but am yet to create my very first Zettel (and I have also learnt that using this noun can be a contentious point!)
This morning (day 3 of my Zettelkasten journey) I discovered this thread, and even though I am still a complete Zettelkasten newbie (ZettelKinder?) I completely devoured the entire back and forth debate between @ralfw and @Sascha with delight. Thank you both for the interesting read! What a clash of titans!
I have also thoroughly enjoyed reading all about @ralfw 's journey from The Archive > Notion > Roam Research. The way you write and explain things is really fantastic.
Wish me luck on my maiden Zettelkasten voyage - I will no doubt be seeking more expert guidance from both of you and all the other wonderful other forum users I've encountered here over the past 2 days.
@ralfw could you briefly show us the process of moving Zettels from Notion to Roam? I use Notion and I am not thinking of moving now, but it's good to have a plan B if things go south with it.
Sorry, @alxfazio, I don't have any experience importing into Roam.
I imported from The Archive into Notion. And with Roam I started a "Zettelkasten 2.0" from scratch with links back to Notion.
@ralfw said:
Sorry, @alxfazio, I don't have any experience importing into Roam.
I imported from The Archive into Notion. And with Roam I started a "Zettelkasten 2.0" from scratch with links back to Notion.
@ralfw thanks for your reply. It would be interesting to see even a simple experiment of migrating a couple of Zettels, to see if it’s feasible or not.
Probably useful for you as well, if you ever decide to commit to Roam and migrate everything
@alxfazio I did an experiment with importing from The Archive into Roam, and it turns out there is a limit of 10 files at a time. Which makes it unfortunately a non-starter, as I have... a lot more than 10 zk files.
@mediapathic said: @alxfazio I did an experiment with importing from The Archive into Roam, and it turns out there is a limit of 10 files at a time. Which makes it unfortunately a non-starter, as I have... a lot more than 10 zk files.
@mediapathic thanks for that. Did you preserve tags, links and backlinks succesfully? Or does it need a lot of manual amending?
@mediapathic said: @alxfazio I did an experiment with importing from The Archive into Roam, and it turns out there is a limit of 10 files at a time. Which makes it unfortunately a non-starter, as I have... a lot more than 10 zk files.
Just another reason to keep your stuff software-agnostic.
Also, as much as I really like Roam and how it is pushing the boundaries, this recent development underscores why having control of your data on your own computer is important.
About a week ago I've given Roam another try and this time I found it much more interesting. But it seems that it might get too cluttered over time, especially considering people tend to save everything in it - from Zettels to Todos, meeting notes, tracking events by dates, etc.
@achamess said:
Also, as much as I really like Roam and how it is pushing the boundaries, this recent development underscores why having control of your data on your own computer is important.
There's nothing unusual. Roam is in beta and they've gained too many users over the last few months that it's causing issues for them (abysmal loading times are the proof to it). Closing registration is a reasonable thing to do in such a situation.
But I do agree, having a local version would've been much better for most users.
@sigod wrote:
About a week ago I've given Roam another try and this time I found it much more interesting. But it seems that it might get too cluttered over time, especially considering people tend to save everything in it - from Zettels to Todos, meeting notes, tracking events by dates, etc.
I hope the guys of Roam solve this issue. I think it is a methodological issue. Roam looks like it nudges you to a cluttered mode.
I committed to digging into Roam Research for awhile, and have come to the conclusion that Zettelkasten-as-practiced-in-The-Archive and Roam encourage you to focus on two different levels of the abstraction hierarchy.
In Zettelkasten, as I understand it, objects (eg, “dogs”) are represented as tags. They don't have a fixed representation, but we might eventually create a summary page for them.
Concepts, on the other hand, connect two objects, or make a claim about an object (eg, “dogs have hair”, “dogs are better than cats”). Then within that concept, we provide an explanation, or support for the claim. These concepts become zettel notes, with a defined ID, title, and card.
In Roam, Pages seem to lend themselves to representing objects. So objects have a more definite presence from the beginning. Concepts which make claims, or which connect two objects, then either get conflated into the object hierarchy as Pages, or get nested under the object page as Blocks.
I think this leads Zettelkasten-Methoders to focus on Claims/Concepts as the fundamental units of thought. And I think it leads Roam-ists to focus on Objects as the fundamental units of thought.
Maybe one way to put it would be that Roam focuses on the nodes in the network, while Zettelkasten focuses on the links.
This seems like a good lens through which to tease out the advantages/disadvantages of each system.
Also, it might be possible to work against this Roam-ish tendency, by using Blocks for Claims/Concepts, and then focusing on Blocks instead of Pages. That seems to be the way that most power-users are trending. The only reason this isn't straight-forward for me is that it either means squeezing a lot into a block, or elevating extraneous details into blocks themselves.
@micahredding said:
I committed to digging into Roam Research for awhile, and have come to the conclusion that Zettelkasten-as-practiced-in-The-Archive and Roam encourage you to focus on two different levels of the abstraction hierarchy.
Hm... I don't think, Roam encourages any focus. Right to the contrary! And that might make it difficult for people to use it.
I think it helps to distinguish between "physical" representation and concept.
The "physical" representation in Roam is on two levels:
Page
Block inside of page
Content can be addressed (linked to, referenced) on both levels. But only pages can be addressed by name, while blocks are addressed by content (even though that gets shortened to an unintelligible address).
(I'm neglecting the block hierarchy through nesting/indentation. It does not introduce another form of addressing.)
In Zettelkasten, as I understand it, objects (eg, “dogs”) are represented as tags. They don't have a fixed representation, but we might eventually create a summary page for them.
Given the "physical" representation I have to decide how to use it for whatever I want to record in my Zettelkasten.
A page is only suitable if I can name my content explicitly. That is easy for people or places, for example, or whatever has a name already. Whenever you record the name of an author it's a no brainer to write the name as a page reference, e.g. [[Stephen King]].
But that's equally easy for a concept, e.g. [[Confirmation Bias]] or [[Conspiracy Theory]] or [[Emotional Blackmail]].
It's the name, I'd say, that let's you choose between page or block, not whether something is an object or a concept.
Using a tag or brackets to reference a page, I think, is a matter of convention and practicality. With brackets you don't have to think about whether a name contains whitespace, for example. Just write the name in a natural way. I'd always go for [[Stephen King]] instead of #StephenKing.
Also my feeling is that tags are more associated with additional annotations and categorization. They are not "primary" names.
It's great that Roam provides you with a page for each tag so you can add further information. But still, to me, tags are not "primary" names. I use them to "color" content or load it up with addition aspects.
Concepts, on the other hand, connect two objects, or make a claim about an object (eg, “dogs have hair”, “dogs are better than cats”). Then within that concept, we provide an explanation, or support for the claim. These concepts become zettel notes, with a defined ID, title, and card.
Pages generate a namespace, an ever growing set of names (page titles). That's what we see when we look at the hierarchy of folders and files in a file system on our computers. But the "page space" in Roam is not hierarchical. All pages are created equal - but they are connected by references. Hence, Roam presents them as a network.
The question to me is not "Is it an object or concept that I want to record?", but "Do I want to pollute the namespace?" Or even more mundane: "Can I come up with a reasonable title for what I want to record in the first place?"
Why not create pages [[Dog]] and [[Cat]] to record whatever seems worthy to record about these animals. And then create another page for the relationship between them, e.g. [[Cats and dogs hate each other]], to record whatever seems worthy to record about the relationship.
Sure, instead a reference to cats could just be made on the dog page (or the other way round) with some info on the relationship between the two. But maybe the relationship is so complicated it warrants a page of its own.
And then the abstraction. Another page for [[Mammal]] which references the cat and dog pages. Or if you don't know yet if that's important to you, just add #mammal to the cat and dog pages.
Roam let's you do this. And that to some is a problem ;-)
It does not force you into a structure.
In a file system things are easy: just one hierarchy, and information can only be recorded in files.
But in Roam you weave a web.
You can do that, too, in The Archive: with only one addressable unit, a file.
You can do that, too, in Obsidian: with files and folders and links to headlines inside of files.
But in Roam tags are not something special but just page references of a different kind. By tagging you create pages.
And Roam let's you address recordings by content.
That's great because it helps avoiding "premature structuring", e.g. structuring your content before you actually know how it best should be structured.
After you set up the cat and dog pages and come to think about the relationship between the two, you might just go to the dog page and add "[[Cat]]s hate dogs and vice versa."
That's a concept - but not yet one you think is important enough for a page of its own.
But then later one day you remember and want to refer to it. You know, you recorded something like that, but you're not sure if it had been important enough for a page of its own. So you type [[hate first and check if there is a page on this concept. And if not you type ((hate - and find the above block.
You then can decide whether to reference the block - or refactor the dog page by extracting the relationship to its own page.
I think this leads Zettelkasten-Methoders to focus on Claims/Concepts as the fundamental units of thought. And I think it leads Roam-ists to focus on Objects as the fundamental units of thought.
Roam, to me, is empty of such focus. It provides you with "physical" representations and a way to weave a web of connections between them including auto backlinks.
It's completely up to you how to use these features.
Like with physical paper slips and a pen. It's up to you how to use them.
But it's obvious that a paper slip is something different than a paragraph on a paper slip. You can address both by giving them an id, e.g. the paper slip gets the id 12b2a4 and on that slip you number each paragraph, e.g. 12b2a4.1, 12b2a4.2, 12b2a4.3.
Then, to be able to access a paper slip or paragraph quickly, you order the paper slips physically by id in a wooden box.
Nevertheless the "weight" of a paper slip is always greater than that of a paragraph due to it being an object.
Paper slips are just handy to cut up an ever growing amount of recordings. They make it easy to insert new recordings between existing ones. No need to retype a whole page just to insert a paragraph in the middle. They are like atoms you can rearrange and through reference even combine.
But in the end, paper slips like files are relics from the past. As containers and addressable units they are not important anymore. With modern databases like Roam uses it's feasible to make every paragraph addressable.
The question then becomes more and more: Which two information nuggets benefit from being put next to each other physically in an ever growing text?
And which information do I want to address by name instead of content?
Roam even makes it easier for you with the first question: You can have the same paragraph "physically" (e.g. in an editable form) in multiple places.
Bottom line: I don't see any preference in Roam to record objects, relationships, concepts etc. in a certain way given the two structural elements page and block and the three different references page ref, tag, block ref.
How you record is a matter of what you want to express. You have to find your own conventions - and then let them evolve and let your Zettelkasten change and grow over time.
Thank you for a very inspiring discussion in this thread.
I am not deep into the formal definitions, understanding and structures of Zettelkasten.
I found out about Zettelkasten after discovering Personal Knowledge Management tools like Roam Research and Obsidian - that showed me a way to build knowledge and understanding gradually, without being tied down by having to guess my structure first.
What appeals to me - besides not needing to classify and tag myself to death when putting down my thoughts (I use Roam Research at the moment) - is the way I can GRADUALLY expand and refine my knowledge, ideas, understanding about something.
Then I found out about Zettelkasten and how that has been part of the inspiration for how many people use Roam Research.
What appeals to me about the latest PKM tools (with backlinks and infinite hiearchy of blocks) and also about the Zettelkasten concept, is this whole thing about the gradual approach - not always with a specific end in mind. Because ideas and direction also changes with increased understanding.
To me, it is about exploration and expanding my view - more than executing something with a set target.
I can see that many have used Zk as a tool or foundation for writing and publishing.
In my personal view, the publishing part is not so vital. It is one use of it, yes.
But for me, the key thing is simply that it carries an INTENT and has some sort of direction to it. (As was already said). It has to do with reaching some sort of insight, or creating some kind of effect or achieving some kind of potential.
This, for me, is what distinguishes it from mere note-taking or "collecting data".
And then the other interesting part - because of the interconnection of Zk's - is the magic of how new associations arise. It is not only about setting out on a course to solve some specific topics towards a known end. Like building a bridge.
In a way, a weird paradox. The character (for me) about Zk's is that they carry intent and direction. Then again, they are not cut in stone, like a normal project would be. The direction and the intent becomes refined as progress occurs.
Comments
Thanks for pointing me to Typora and VNote, @Peter! I'll check out at least Typora. Looks appealing
But the one thing I really, really don't want to miss anymore after my experience with Roam, is backlinks. Automatic backlinks to be precise. It's really what makes a difference when building a zettelkasten. To me at least.
Plus - I guess I have to add - the ability to link to each and every paragraph (block) in Roam.
Both features taken together make it so much easier to weave a web of zettels (or should I say thoughts?). Because I don't have to think about where to put a thought. Should it become a separate page? Should it be nested inside some existing page? Is an effort to set up a backlink justified? That's all questions I don't feel burdened with anymore.
Zettels to me have become something I write down like anything else over the course of a day. They are anchored naturally to a date, i.e. to a context. But that does not tie them up. They show up in a search, I can reference them from anywhere, I can refactor my "daily documentation" (of which zettels are a part) at any time very easily (extract a zettel into a separate page, inline a zettel into an existing page, "rename" a zettel).
I don't mind to get all this with a tool which stores my "daily documentation" in local .md files in my dropbox - as long as I get the same functionality for weaving my web.
I'd even argue that should not be difficult to build. No need to host it on the web. No need to scale it for thousands of users. I'd be happy to do it myself - but I'm not the UI guy for today's fancy desktop programming technologies.
Simple Markdown editors don't do the job for me, though. They are not smart about the purpose: building a second brain.
Markdown is a great start - but without a data model underlying Markdown texts I don't think they will be powerful enough. At least for my requirements.
Roam is doing a good job in the regard. It pretty much looks like a Markdown editor - but that's only a user interface on top of a database, I'd say. And I guess that kind of software architecture is the key to Roam's features.
Let's wait and see... The evolution will continue
@ralfw : it certainly was not my intention to get you off Roam, so to speak. I pointed out Typora and VNote in case you were interested in a markdown editor.
Like I said, VNote is the more advanced of the two, with more features. Regarding backlinks, if you do an Advanced Search in VNote, e.g. looking for note number 123987, you get a list with all the notes that have that number in their text, i.e. through links. So, apart from the original note with number 123987 in that list, all the other notes listed are effectively notes backlinked to note 123987.
I don't quite understand your comment about the simple markdown editors. I have seen 1 tutorial about Roam and know it is very different from a markdown editor, a lot more complex. All I can say is that Typora is "just" a markdown editor, whereas VNote is a markdown editor cum note-taking manager. Instead of checking out Typora I'd recommend checking out VNote, if you are interested. Don't do it to be kind to me or not to hurt my feelings. I shan't be hurt if you aren't interested in either of the two: I know those who use Roam are absolutely crazy about it.
Again, I am not trying to get you to drop Roam. Also, I have no commercial or other interests in either Typora or VNote, so I am not "talking my book". I was just after an exchange of ideas
Don't worry, @Peter, I didn't think you were trying to get me off of Roam.
I took a look at Typora and like it for its simplicity. And it's great it lets me so easily include images from the clipboard even with storing them right where I want them in my local folder hierarchy.
And maybe some other time I'll check out VNote.
But what you mentioned as a feature of VNote (and The Archive, if I remember correctly), that you can search for some note ID and find all the notes containing it, that's exactly what I'm not (!) looking for.
I don't want to deal with IDs anymore. And I experienced how convenient it is to not even have to think about files (or pages or documents) anymore.
When you experience something your mind is able to connect to any part of that experience. You don't need to break it up into certain chunks. It does not need to have a particular granularity.
Looking at a computer I might remember a certain book about micro processors I read when I was in my early 20s. That book then reminds me of my bookshelf in my room at home with my parents. And that bookshelf in the context of the micro processor book then might evoke a memory of me and my father working on a electronic tinker kit when I way maybe 7 years old. And that fond memory then might lead to an even earlier moment us looking at the stars.
Looking at this chain of memories from the other end: When we were watching the stars I did not need to think about how to break up and store the then current experience. What I saw, heard, felt... all that was woven into my brain/body in a way so that future experiences could relate to it in any way, to details or the overall experience, to a smell or a sight.
That's how I'd like to work today when I log a thought, and experience. The only thing I want to do is, write it down, paragraph by paragraph. And then rest assured that I can connect to any part of it at any time.
The atoms of such recording to me are the paragraphs. That I find convenient. They come naturally to me as units of text with a certain meaning.
Paragraphs can be put in sequence and can be nested (in Roam). Thus they end up in a context. Linking to an already exiting paragraph pulls something into another context I'm currently in - and lets me jump to the context the former paragraph was written in.
Adding a tag to a paragraph or isolating it into a page of its own to me are optimizations. If I can do that to add even more context/meaning, that's great. But the basic building block to me right now are just paragraphs not otherwise attributed with meta data.
I did not know that this would become so important for me. But that's my current experience with Roam. That's what's making working with it fundamentally different.
That said I don't want to sell Roam here. It just happens to be the tool right now implementing a certain approach to recording "thoughts".
I don't care about products or manufacturers. I care about convenience. How much do I feel supported in pursuing my interests without getting distracted by infrastructure or technological idiosyncracies?
As it turns our, it's a journey to come ever closer what suits me best - probably never ending
And from what you wrote about how you started with Evernote and now are using two tools, it's somewhat the same for you. Let's wait and see what 2022 will bring... 
This is just a meta-commentary to say that this conversation between @ralfw and @Sascha is an excellent example of what @ctietze said here on the event of the anniversary: this community is very strong because of our ability to have debates such as these in an enthusiastic yet civil manner. As someone who has no opinion here (being a strict anti-dogmatist about theory), I read through this thoroughly just to see two people have a disagreement in such an agreeable way.
I love this place.
@ralfw, you said:
To me, this is the crux of your thinking and how you want to build and expand your knowledge base. And whether it is called a Zettelkasten or anything else is not important; besides, it is your right to call it whatever you wish. What is important is whether the tool works for you in the best possible way.
That is nothing exceptional, it is possible with other apps like The Archive, VNote too, surely?
I agree. And to put that in my context, my knowledge base consists of 2 parts:
The 1st is mostly (historic) facts-based, with interpretations, and not really a Zettelkasten tool.
The 2nd one, however, is a Zk-type tool, and I have often wondered how, apart from the uni-directional links, I could get a more dynamic, interactive tool like you describe above. I looked at Roam and dismissed it as more of a task manager than a dynamic note-taker/manager. Your comments show that I was wrong, have given me new food for thought and perhaps I should look at it again
The only other drawback of Roam I see at this stage is future portability. If the dev decides to quit, you can of course continue to use it. But what if you wanted to switch: what format are your notes in and can they be read into another app? With markdown or plaintext notes there is no problem, but with Roam notes?
Is that possible? I'd love to see, how to do that.
Here's what I mean: I write a text and structure it into paragraphs. Or think of it as a hierarchical list of bullet points.
some thought here
a second thought there
And so on...
And all these paragraphs can be referenced from other paragraphs in other "documents" (.md files or pages)? Without adding any meta-information like a paragraph number? And that produces a link back to the paragraph?
Here's how it's done in Roam:
No paragraph is out of reach for reference. No provisions have to be made while writing a paragraph.
That's the convenience I'm looking for. It keeps me from constantly switching between the real and the meta
You can export all your Roam "notes" to .md files at any time. In fact I'm doing that almost every day now while Roam is in its early stages of development. Just to be safe
It's a matter of 10 seconds.
No, that is not possible, you are right.
I agree, that is actually how it should be. I have not had that real Eureka feeling from any tool (yet?).
In principle I am against having to export to be able to pick up my notes. Then again, it could be an acceptable compromise if a tool is worthwhile.
As for those 10 secs., I should imagine that depends on the number of notes, or rather, on the size of note(s) / file(s) to be exported.
BTW, I notice that you need to sign in to the site: does that mean Roam is web-based only, i.e. no app on your computer?
Yes, Roam currently is web-based only. You can use it in your browser on the desktop or on a mobile device. For me that's fine.
Last question @ralfw : how private is that? Can Conor White-Sullivan and his team freely browse your notes?
I don't know.
I asked CWS about this and believe that they currently can, although user encryption is a long-term goal.
Thank you for that feedback, @cobblepot.
It is a real pity that they do not have a desktop version because from what I have seen over the past 2 days this app could blow all the other Zettelkasten-type note-taking apps out of the water, and I for one would certainly want to get stuck in, so to speak.
But to entrust one's notes to an unencrypted 3rd party is simply not acceptable to me, and I am sure to a lot of other people. And that has nothing to do with having something to hide.
I am surprised someone like @ralfw accepts it.
Since you‘re addressing me personally, @Peter, let me just say:
I‘m taking the topic of data security not lightly. But after much thought I have come to certain conclusions. One is, that it‘s not a matter of black and white.
For storing my passwords I‘m taking different measures than for storing my zettels, for example.
I‘m stuck in a country which has shut down public life in an instant. Politicians are demanding to track infected people. What‘s next? Forced testing for infection?
Let me say, these days I‘m not really concerned about zettels in some database which potentially people could read and derive from what I‘m reading and thinking. Obviously they don‘t need that to force me and millions others into a standstill.
I‘m not worried about a government (agency) (or company) which spies on my data. I‘m worried about those who don‘t need that anymore.
That said, please let‘s not turn this thread into a discussion about security consciousness etc. I‘m happy to read you took a second look at Roam and seem to like it more. And I agree: a desktop version storing data locally would be even cooler.
Storing data locally and syncing them between devices doesn‘t sound all to difficult too me. But for whipping up a „prototype“ it might have been easier to go web-first.
@Peter , not sure if this is of interest, but you could check out Org-Roam
Two days ago I had never heard of Zettelkasten
(nor Luhmann, I am ashamed to admit).
I discovered Sonke Ahrens book entirely by chance and, even though I have only read about 10% of the book, I have already jumped down the rabbit hole of devouring all the blogs, medium articles, youtube videos, and anything else related to Zettelkasten I could get my hands on. It's one of those incredible moments in life when you don't realise there is a solution to your problem until you suddenly trip right over it whilst walking (or in my case, browsing Amazon!). In this case, the 'problem' was finding a better solution to my personal knowledge management and book writing systems.
2 days later and I now have accounts on Notion and RoamReseach - but am yet to create my very first Zettel (and I have also learnt that using this noun can be a contentious point!)
This morning (day 3 of my Zettelkasten journey) I discovered this thread, and even though I am still a complete Zettelkasten newbie (ZettelKinder?) I completely devoured the entire back and forth debate between @ralfw and @Sascha with delight. Thank you both for the interesting read! What a clash of titans!
I have also thoroughly enjoyed reading all about @ralfw 's journey from The Archive > Notion > Roam Research. The way you write and explain things is really fantastic.
Wish me luck on my maiden Zettelkasten voyage - I will no doubt be seeking more expert guidance from both of you and all the other wonderful other forum users I've encountered here over the past 2 days.
Tom.
A ZettelKinder from London, UK
@ralfw could you briefly show us the process of moving Zettels from Notion to Roam? I use Notion and I am not thinking of moving now, but it's good to have a plan B if things go south with it.
Sorry, @alxfazio, I don't have any experience importing into Roam.
I imported from The Archive into Notion. And with Roam I started a "Zettelkasten 2.0" from scratch with links back to Notion.
@ralfw thanks for your reply. It would be interesting to see even a simple experiment of migrating a couple of Zettels, to see if it’s feasible or not.
Probably useful for you as well, if you ever decide to commit to Roam and migrate everything
@alxfazio I did an experiment with importing from The Archive into Roam, and it turns out there is a limit of 10 files at a time. Which makes it unfortunately a non-starter, as I have... a lot more than 10 zk files.
Good to know. Thanks for the info. Indeed that's kind of a bummer. But then Roam still is deep in beta.
@piotr would you mind sharing a snippet of you year/month/week tag?
Do you store them in the file name or in a header?
@mediapathic thanks for that. Did you preserve tags, links and backlinks succesfully? Or does it need a lot of manual amending?
Just another reason to keep your stuff software-agnostic.
I am a Zettler
Couldn't really say. The 10 files I selected at random for my experiment were not a useful subset for understanding that. Sorry.
Also, as much as I really like Roam and how it is pushing the boundaries, this recent development underscores why having control of your data on your own computer is important.
About a week ago I've given Roam another try and this time I found it much more interesting. But it seems that it might get too cluttered over time, especially considering people tend to save everything in it - from Zettels to Todos, meeting notes, tracking events by dates, etc.
There's nothing unusual. Roam is in beta and they've gained too many users over the last few months that it's causing issues for them (abysmal loading times are the proof to it). Closing registration is a reasonable thing to do in such a situation.
But I do agree, having a local version would've been much better for most users.
I hope the guys of Roam solve this issue. I think it is a methodological issue. Roam looks like it nudges you to a cluttered mode.
I am a Zettler
I committed to digging into Roam Research for awhile, and have come to the conclusion that Zettelkasten-as-practiced-in-The-Archive and Roam encourage you to focus on two different levels of the abstraction hierarchy.
In Zettelkasten, as I understand it, objects (eg, “dogs”) are represented as tags. They don't have a fixed representation, but we might eventually create a summary page for them.
Concepts, on the other hand, connect two objects, or make a claim about an object (eg, “dogs have hair”, “dogs are better than cats”). Then within that concept, we provide an explanation, or support for the claim. These concepts become zettel notes, with a defined ID, title, and card.
In Roam, Pages seem to lend themselves to representing objects. So objects have a more definite presence from the beginning. Concepts which make claims, or which connect two objects, then either get conflated into the object hierarchy as Pages, or get nested under the object page as Blocks.
I think this leads Zettelkasten-Methoders to focus on Claims/Concepts as the fundamental units of thought. And I think it leads Roam-ists to focus on Objects as the fundamental units of thought.
Maybe one way to put it would be that Roam focuses on the nodes in the network, while Zettelkasten focuses on the links.
This seems like a good lens through which to tease out the advantages/disadvantages of each system.
Also, it might be possible to work against this Roam-ish tendency, by using Blocks for Claims/Concepts, and then focusing on Blocks instead of Pages. That seems to be the way that most power-users are trending. The only reason this isn't straight-forward for me is that it either means squeezing a lot into a block, or elevating extraneous details into blocks themselves.
Really interesting to see these dynamics.
Hm... I don't think, Roam encourages any focus. Right to the contrary! And that might make it difficult for people to use it.
I think it helps to distinguish between "physical" representation and concept.
The "physical" representation in Roam is on two levels:
Content can be addressed (linked to, referenced) on both levels. But only pages can be addressed by name, while blocks are addressed by content (even though that gets shortened to an unintelligible address).
(I'm neglecting the block hierarchy through nesting/indentation. It does not introduce another form of addressing.)
Given the "physical" representation I have to decide how to use it for whatever I want to record in my Zettelkasten.
A page is only suitable if I can name my content explicitly. That is easy for people or places, for example, or whatever has a name already. Whenever you record the name of an author it's a no brainer to write the name as a page reference, e.g.
[[Stephen King]].But that's equally easy for a concept, e.g.
[[Confirmation Bias]]or[[Conspiracy Theory]]or[[Emotional Blackmail]].It's the name, I'd say, that let's you choose between page or block, not whether something is an object or a concept.
Using a tag or brackets to reference a page, I think, is a matter of convention and practicality. With brackets you don't have to think about whether a name contains whitespace, for example. Just write the name in a natural way. I'd always go for
[[Stephen King]]instead of#StephenKing.Also my feeling is that tags are more associated with additional annotations and categorization. They are not "primary" names.
It's great that Roam provides you with a page for each tag so you can add further information. But still, to me, tags are not "primary" names. I use them to "color" content or load it up with addition aspects.
Pages generate a namespace, an ever growing set of names (page titles). That's what we see when we look at the hierarchy of folders and files in a file system on our computers. But the "page space" in Roam is not hierarchical. All pages are created equal - but they are connected by references. Hence, Roam presents them as a network.
The question to me is not "Is it an object or concept that I want to record?", but "Do I want to pollute the namespace?" Or even more mundane: "Can I come up with a reasonable title for what I want to record in the first place?"
Why not create pages
[[Dog]]and[[Cat]]to record whatever seems worthy to record about these animals. And then create another page for the relationship between them, e.g.[[Cats and dogs hate each other]], to record whatever seems worthy to record about the relationship.Sure, instead a reference to cats could just be made on the dog page (or the other way round) with some info on the relationship between the two. But maybe the relationship is so complicated it warrants a page of its own.
And then the abstraction. Another page for
[[Mammal]]which references the cat and dog pages. Or if you don't know yet if that's important to you, just add#mammalto the cat and dog pages.Roam let's you do this. And that to some is a problem ;-)
It does not force you into a structure.
In a file system things are easy: just one hierarchy, and information can only be recorded in files.
But in Roam you weave a web.
You can do that, too, in The Archive: with only one addressable unit, a file.
You can do that, too, in Obsidian: with files and folders and links to headlines inside of files.
But in Roam tags are not something special but just page references of a different kind. By tagging you create pages.
And Roam let's you address recordings by content.
That's great because it helps avoiding "premature structuring", e.g. structuring your content before you actually know how it best should be structured.
After you set up the cat and dog pages and come to think about the relationship between the two, you might just go to the dog page and add "[[Cat]]s hate dogs and vice versa."
That's a concept - but not yet one you think is important enough for a page of its own.
But then later one day you remember and want to refer to it. You know, you recorded something like that, but you're not sure if it had been important enough for a page of its own. So you type
[[hatefirst and check if there is a page on this concept. And if not you type((hate- and find the above block.You then can decide whether to reference the block - or refactor the dog page by extracting the relationship to its own page.
Roam, to me, is empty of such focus. It provides you with "physical" representations and a way to weave a web of connections between them including auto backlinks.
It's completely up to you how to use these features.
Like with physical paper slips and a pen. It's up to you how to use them.
But it's obvious that a paper slip is something different than a paragraph on a paper slip. You can address both by giving them an id, e.g. the paper slip gets the id
12b2a4and on that slip you number each paragraph, e.g.12b2a4.1,12b2a4.2,12b2a4.3.Then, to be able to access a paper slip or paragraph quickly, you order the paper slips physically by id in a wooden box.
Nevertheless the "weight" of a paper slip is always greater than that of a paragraph due to it being an object.
Paper slips are just handy to cut up an ever growing amount of recordings. They make it easy to insert new recordings between existing ones. No need to retype a whole page just to insert a paragraph in the middle. They are like atoms you can rearrange and through reference even combine.
But in the end, paper slips like files are relics from the past. As containers and addressable units they are not important anymore. With modern databases like Roam uses it's feasible to make every paragraph addressable.
The question then becomes more and more: Which two information nuggets benefit from being put next to each other physically in an ever growing text?
And which information do I want to address by name instead of content?
Roam even makes it easier for you with the first question: You can have the same paragraph "physically" (e.g. in an editable form) in multiple places.
Bottom line: I don't see any preference in Roam to record objects, relationships, concepts etc. in a certain way given the two structural elements page and block and the three different references page ref, tag, block ref.
How you record is a matter of what you want to express. You have to find your own conventions - and then let them evolve and let your Zettelkasten change and grow over time.
Thank you for a very inspiring discussion in this thread.
I am not deep into the formal definitions, understanding and structures of Zettelkasten.
I found out about Zettelkasten after discovering Personal Knowledge Management tools like Roam Research and Obsidian - that showed me a way to build knowledge and understanding gradually, without being tied down by having to guess my structure first.
What appeals to me - besides not needing to classify and tag myself to death when putting down my thoughts (I use Roam Research at the moment) - is the way I can GRADUALLY expand and refine my knowledge, ideas, understanding about something.
Then I found out about Zettelkasten and how that has been part of the inspiration for how many people use Roam Research.
What appeals to me about the latest PKM tools (with backlinks and infinite hiearchy of blocks) and also about the Zettelkasten concept, is this whole thing about the gradual approach - not always with a specific end in mind. Because ideas and direction also changes with increased understanding.
To me, it is about exploration and expanding my view - more than executing something with a set target.
I can see that many have used Zk as a tool or foundation for writing and publishing.
In my personal view, the publishing part is not so vital. It is one use of it, yes.
But for me, the key thing is simply that it carries an INTENT and has some sort of direction to it. (As was already said). It has to do with reaching some sort of insight, or creating some kind of effect or achieving some kind of potential.
This, for me, is what distinguishes it from mere note-taking or "collecting data".
And then the other interesting part - because of the interconnection of Zk's - is the magic of how new associations arise. It is not only about setting out on a course to solve some specific topics towards a known end. Like building a bridge.
In a way, a weird paradox. The character (for me) about Zk's is that they carry intent and direction. Then again, they are not cut in stone, like a normal project would be. The direction and the intent becomes refined as progress occurs.