@RikD: +1 for 6-Sigma. I did not work as one, but did get closer through Lean. The latter concept is marvellous, can be applied to almost any situation, incl. the private situation. Even when I go through a supermarket I try to do it in a Lean way, and get irritated when my wife goes back and forth because her shopping list is constructed in a haphazard way.
1 Like
Again, very interesting. I also found #tags low utility, given they grow quickly to a unmanageable size, and the availability of a search function. Itâs also good to hear the Obsidian manages a large number of notes well.
If you would indulge me further, a few more questions:
- It sounds like you have two main types of notes/Zettels - source Zettels of full text articles from original sources, and comprehensive summary Zettels that pull together ideas from different articles. Is that right?
- Do you take notes within your source Zettels, or use them strictly as a repository to reference when writing your summary notes?
- Do you link to and from your source Zettels into summary Zettels, or link between your comprehensive summary Zettels (i.e. do links cross the two categories)? I worry that if I link liberally within source Zettels (which outnumber summary Zettels by orders of magnitude) that the utility of the graph may degrade.
- How do you distinguish your source Zettels from your summary Zettels in Obsidian - archives, folders, naming convention, indices, or not at all? Simply using the search function?
- How devoted are you to the atomicity principal in your summary Zettels?
- Do you tend to revise your summary notes over time, or do you prepare them once in advance of your work obligations that need them and then move on?
I agree that every workflow will differ, and we all need to iterate our own approaches rather than take a one-size-fits-all approach to note taking - I found some aspects of . But, I notice some overlaps in our workflows (I described mine here and here, above in this thread). Iâve had a very limited, but functional, source saving system for years, but have recently been trying to iterate improvements using Obsidian. Your steps have helped give me some ideas to experiment with myself, and thatâs why Iâm probing more here.
Again, appreciate all your comments!
Aside: we are kindred spirits on this. From one to another, beware pushing efficiency in the supermarket too hard! 
1 Like
Thanks for the warning, you are right. I do take my foot off the accelerator, telling myself that Lean was not really meant for that. Itâs possible to drive oneself and those around nuts with it 
Sorry for the late reply. Some other issues to solve first 
- I âtreatâ all my Zettles the same. For me (personal opinion) they are all just âconceptsâ/pieces of information that one day (or not) could be linked to one another.
- I donât take notes in concept Zettles. They are âthe sourceâ. Again this is personal. I donât want to evangelise my personal workflow.
- Links are placed in the âsummaryâ Zettles allthough you could see them also as Concepts when you really want to. Luhmanâs workflow seemed to be that he didnât put more than 3 references to other Zettles in his notes. I tend to go for 5 as a maximum. Links to other concepts can be much higher. Certainly when you are writing research papers. With links I personally refer to what you call Summary Zettles. I would not put more then 5 of those links in one Zettle.
- Indeed search function. Obsidian is the most powerfull (together with Sublime Editor which I personally use the longest now). I literally crashed Zettler and Joplin with my bigger vaults (10k Zettles seems to be the max for both). Zettlr and Joplin do not handle big numbers of files well. The tool litteraly becomes very slow and unusable. Obsidean seems to handle this very well. And no distinction between source and summary as already mentioned. Thinking about how search can help me âupdateâ the source Zettles with internal links between one another. E.g. I have done some research about Company X. I would do a global search for âCompany Xâ. You probably are writing a comprehensive summary article but it would be nice to see that in the Graph too. Therefor you would need to replace Company X by [[Company X]] in the sources Zettles. Not sure if that is a good approach. All advice welcome there
- How devoted? Very as I follow the max link rule of that is what you meant ;-). Try to keep summary Zettles as small as possible. Even after some years of active Zettelkasten research I have a good workflow (even if I say it myself), but I still struggle with some small things.
Hope to having helped you with this.
If not: shoot new questions. WIll try to answer them ASAP.
5 Likes
@RikD: some good points here.
I have been working on my zettelkasten for 6 years, incl. setting it up. I started 1st with Daniel Luedeckeâs ZKN3 app, but soon found it too limited, not flexible, and not offering enough features.
I shanât describe all my travails, suffice to say that I went to Evernote, OneNote, CherryTree, TiddlyWiki, and now Obsidian.
The way Obsidian works, and its features that keep evolving, makes it a real joy to work on my zettelkasten. I also have a different set of notes, with long to very long notes, and even there Obs is a pleasure. And many more useful features are on the way, absolutely mouth-watering.
The reason I say all this is because of the point I want to make: I have noticed that since I started using Obs some 3 months ago, I increasingly think in terms of links. My mind is much more alert to seeing connections between notes, and connections or duplications I had not picked up before.
I even discover relationships between things I hear and see that have nothing to do with my zettelkasten !
I have not tried to analyse the reason(s), I am just happy to have got to what I would describe as the next level of ârelationship awarenessâ, which I believe (perhaps wrongly, perhaps not) is due to Obsidian.
Totally valid point youâre making here.
The more you use the Zettelkasten approach the better you get at it. Luhmann seemed to be an absolute example of that.
I do believe Obsidian is certainly a positive enabler, offering you all the tools and gimmicks that make it possible to âseeâ links and other connections.
That will set you to make more Zettles and your second brain will become bigger and bigger and will help you become as good as Luhmann in creating new content.
I am still very far from that point. Certainly working towards that.
Great to see some likeminded souls who are trying to reach this goal too.
Hope Obsidian will definitely become the number one tool achieving that!
1 Like
Thanks again - all interesting comments to consider. Iâm tinkering with my own note taking approach and will report back as I have anything new/innovative to offer.
My observation, from my own habits, is that getting into Zettelkasten it is easy to pile up tons of material quickly, which may not be processable ever (if one keeps collecting at that pace). As a result, most of my deliberate thinking on how to use Obsidian, including for analyzing trends in economies, markets, geopolitics, are around the idea of sustainability of the practice:
- Do these habits make work towards my professional writing obligations more productive and insightful, or are they creating additional archival work that is merely additive rather than multiplicative?
- Can I maintain this practice for a couple hours a day, for years? How enjoyable is the process itself, relative to my earlier systems of gathering information, generating insight, and composition?
- What is the optimal amount of source (literature) notes to gather daily within Obsidian vs just reading and moving on? Your comments on scaling of search are useful in helping understand Obsidianâs strengths as a Markdown archive vs other options.
- What about daily creation of true Zettels, my own original creations (whether long or short)? Do I optimize summarizing anything I can each day, and assume serendipity will occur to make that eventually useful? Or create Zettels (short compositions) that are very goal oriented towards weekly/monthly reporting, and again assume value will amass over time via linking and search? How much do I focus on editing old Zettels versus merely linking to them from newer Zettels that expand on old ideas?
- Those last few bullets could be summarized to: Will I dread opening Obsidian because I see so many âin processâ notes in my backlog, many eventually orphaned? Or instead get bogged down trying to write only a few notes per day that are more certainly useful long term?
1 Like
Again all valid remarks.
My âArchiveâ vault is created and maintained automatically so no âdaily stressâ there. You could argue if this is a good practice. I tend to lean towards âYesâ as the information often isnât retrievable on the WWW after a few years and the search capabilities of the source is rather questionable ( )
For your last bullet:
I (personal opinion) would go for writing whatever comes up in a daily-note. The workflow that I try to maintain is overviewing the notes and [[linking]] interesting stuff to new Atomic Notes (mentioned in the thread about block linkingâ. Those atomic Notes (read Zettles) do not need to have âcontentâ yet. But if you can already fill them with the general idea that is great.
The graph directly gives you on idea on the âorphenageâ status. Question for me is if âorphanageâ is good or bad (look at my comment on gravition towards something or away from someting overtime).
What I am going to be experimenting with is the Obsidian option of the random note. That creates a random âstrollâ through your second brain which sometimes could spark new thing (and sometimes not which is great).
Thanks again for your insights. This is really a great community to be in!
Some thoughts from me on archiving articles.
I donât have a job were I need be able to read and categorize so many articles as you guys.
However I work in IT and am a geek in so I still read a lot.
As RikD (I think) mentioned it can be quite a problem that content on the internet just disappears.
It doesnât even have to be that many years.
Since I noticed this being a problem for me I started to use my Wallabag instance (think Pocket but selfhosted) in the reverse way.
I read an article and then send it to Wallabag for archiving.
When I take notes I try to link to the article in my Wallabag instance.
This way I should always have a readable version of the article (if Wallabag was able to get it) and Wallabag will store the original link for me.
A problem for is that my Wallabag server is something that I have to keep running in order to be able to read the articles.
Which might not be possible in the long run.
Iâm therefore currently thinking about an automated approach which would let me send articles to Wallabag and then exports them to Markdown, Plaintext, HTML or whatever other plaintext format I could use that would make sense.
Another thing I experimented with is storing the websites directly as a HTM or HTML file but that doesnât work so well on mobile and Iâm not sure how well those files would work in a future browser.
IMO something in plaintext would be the best and most future proof.
Linking to all those articles, well thatâs a whole different topic.
2 Likes
I would be interested in what results you have, because Iâm in a similar situation. Iâm an analyst so Iâm reading several dozen articles and reports each day and itâs very hard to manage the information. It seems like a tough balance to determine if something is âworthy enoughâ to get a note in Obsidian.
I do like your idea how linking things to a question that needs answering. Iâve been experimenting with a similar thing which I just refer to as âThemesâ. For example creating one like âUS-China trade warâ which might have some information in the note, but is mostly a linking point for other less in-depth notes. Something like a tariff getting adjusted probably isnât âworthy enoughâ on its own to be worth saving, if I connect it to the US-China trade war theme, then itâs easy to refer to back lately.
Hey @Nebucatnetzer,
That is how I do keep my stuff. I do convert articles directly into .md-files though. Some of my older archives are HTML (which is text also )
With .md one looses a little bit of lay-out but that is in my opinion not what knowledge is about. The information needs to be there
Today I did some work related research on my biggest vault (85k+ articles). Managed to get the demanded information out in a formatted document under five minutes. Added some comments and (hopefully) useful anchor-points to get the discussion going and start evolving to a good basis for an in-depth discussion.
2 Likes
How do you manage the pictures in an article?
A plugin like MarkDownload keeps images as references to web addresses, so they will display fine so long as those images addresses donât change. For my usage this is fine as I mostly am downloading to archive the text.
How do you save them on mobile? If I may ask so directly 
.
I use Firefox share to desktop function, and then save there later.
Good question!
Pandoc lets you make a standalone html (selfcontained) with everything included.
For .md I havenât found an automatic flow yet to go from URL to ready to use files.
Still some manual stuff to get the links OK (to the pictures). For me that is not an option.
When I found an automatic workflow I will certianly share!
1 Like
I keep them in any app that helps me keeping URL for a âread laterâ.
That list can then be used to make your .md-files on your laptop.
Thinking of it:
You could create an .md (Dropbox, OneDrive, âŚ) with those URL links and work your way from there. Daily notes .md is also a good starting point.
You wouldnât even have to use Pandoc to get a self contained HTML file.
There are various plugins which let you save a website as an MHTML file.
E.g. https://addons.mozilla.org/en-CA/firefox/addon/save-page-we/?src=search for both Firefox and Chrome.
However Iâm not sure how good those HTML files will work at a later point in time.
As for the automation, if someone doesnât mind using proprietary services like Dropbox, etc. Iâm sure itâs possible to come up with a workflow in IFTTT to convert a website to Markdown and save it into a Dropbox account.
Hence the use of Pandoc 