String – Structured Notes: Logseq

Wed September 21, 2022 Andima

There are many ways to create notes. You can use simple note apps, keep a wiki and you don’t have to be too quick to consider the good old paper economy a trace.

Another way is to use “Outliners” or – if you want to translate – the “Outlines Editor”.

In addition to commercial solutions such as Roam Research and Obsidian, which are often mentioned and which, strictly speaking, do not belong to third parties, there are also open source solutions. At this point it should be about Logseq under AGPL v.3.0.

With classic outlines and thus also with Logseq, you can create your notes in the form of bullet points, i.e. outline points. Depending on the application, this can be an advantage or a disadvantage. The only thing that helps you is to test and check if this type of note taking fits your own work style or not. However, Logseq also supports Document Mode to generate text that works more conveniently.

One advantage of these tools is their ability to link individual points together, since a backlink is inserted automatically, unlike a wiki. Or to put it in a less complicated way: you can see on a point (or page) who is referring to that point.

Here, Logseq stores notes in text form, and you can choose whether you prefer writing your notes in Markdown or in Org mode. Using a common text format is an advantage if you later want to switch from Logseq to another tool.

However, the warning is appropriate here. For correlation, Logseq saves hash values ​​or information about whether the enumeration point is folded or not, so there may be a certain snag when changing tools.

And while we’re on the topic of reminders. Logseq has a setting to send anonymous data remotely, which unfortunately is turned on by default. If you don’t want that, you have to turn it off under “Settings -> Advanced”.

What I also don’t want to hide is that Logseq is still in beta. Sometimes this means that you have to rebuild the graph – name “notes project” – after the update, which isn’t particularly tragic. With very large graphs, I can imagine this could take a while.

Like I said, you create your notes in the form of points. You can build hierarchies through indentations, although I didn’t notice that they were depth-limiting in any way.

Individual blocks can be distinguished like pages. Aliases can be created for tags.

Pages can also be organized hierarchically, although this is not recommended. Less for technical reasons, but because technically identical links and tags are preferable, as they give more freedom.

This gives you many options for organizing your notes and the automatically generated backlinks can be very useful.

So there are whole points here; 5 tuxedos.

How do I gently teach this to one or the other? Maybe not at all. So, if your computer has only 512MB or if your RAM is too good to use, you should stop reading now. Believe me, you are just upset. I warned you!

Because the basis of Logseq is the electron! Now he is out. And yes there are worse things.

You can use tags for that! And yes, there is darkness, too.

Apart from that, we are dealing with a three-column layout with the main editor in the middle. On the left is the navigation bar and on the right there is a column in which you can view another editor window, for example, or a freely editable table of contents. The left column can be hidden as well as the right column.

Otherwise the editor is just an editor and does what it is supposed to do. There are also keyboard shortcuts, but I haven’t seen a way to customize them yet.

Here I give 4 out of 5 unenthusiastic tux.

The best knowledge database in the world is useless if you can’t find anything in it. Accordingly, good research is an important point.

Logseq presents this with some caveats.

Finally, different search options are available. The simplest way to search is a dialogue-based full-text search, which can be easily done using a keyboard shortcut, among other things. cmd + k can be invoked. Advanced search parameters are not supported here, so it’s really just a standard search.

There is also the option to select a search anywhere in the editor, with two different search combinations to choose from. Then the search result is displayed directly under the meta and is automatically updated when the page is accessed. So you can create a separate page to search for frequently.

With a simple search syntax, it is relatively easy to perform searches of medium complexity. If you want to get the most out of the search functionality, you have to use the more complex syntax. However, you can really find everything with it.

Awarding points is a bit tricky here, because there is actually very good research available. On the other hand, the required formula is not self-explanatory. (Unless you have already dealt with the Datomic database…)

To clarify better, using an example from the official documentation, where all pages marked with the word “programming” are searched:

{:title "All pages have a *programming* tag"
 :query [:find ?name
       :in $ ?tag
       :where
       [?t :block/name ?tag]
       [?p :block/tags ?t]
       [?p :block/name ?name]]
 :inputs ["programming"]
 :view (fn [result]
       [:div.flex.flex-col
        (for [page result]
          [:a {:href (str "#/page/" page)} (clojure.string/capitalize page)])])}

Give 3.5 tux suits.

As mentioned earlier, Logseq supports backlinks. In addition, we already know that you can link from one block to any other block.

Of course you can also link to a page.

It has already been mentioned that you can create aliases for tags.

Another important thing to know is that links and tags in Logseq are semantically similar, although syntactically different.

The tags are written in the usual garden fence notation (double cross): #IBinOneTag

The internal links are set in square brackets: [[IchBinEinLink]]

Tags cannot contain spaces, links do not contain spaces.

Personally, I use links and tags in continuous text whenever I specifically want to tag a block.

Long story short: 5 out of 5 tux

Logseq’s export options can be described as average if you exclude the use of plugins.

You can export individual pages as markdowns and you can export the entire graph as html, here, unless configured otherwise, only those with general Tagged pages are exported.

Additionally, you can opml exports.

The pdf can be exported via the plugin, but I’ve never tested it.

Import includes the ability to transfer roam and OPML search files to Logseq.

Previous Markdown or Org-mode files can be stored at the file level and included in Logseq by re-indexing.

If I rate only the natively built import and export functions, I’d give Tuxe 3 out of 5.

Logseq comes with a variety of functions. Add-ons can be modified using plugins, although the selection is not yet as large as with Obsidian and perhaps as with Org-mode.

Chances are to create flash cards, you can, of course, mark tasks and give them a history, although there is no notification function here without a plugin.

Graphics can be created in Logseq by integrating the TypeScript application Exaledraw Implemented, slide creation is made possible through the integration of reveal. js made available.

There is a calculator available, and of course there is source code that highlights for code examples and you can embed videos from online sources and run them inside Logseqs. Here you can also set transition marks, which is useful for notes on videos, as you can go directly from the note to the corresponding point in the video.

You can embed PDFs, read them in Logseq, and can highlight text segments in a PDF and provide a reference that can be used in a Logseq page as a jump tag to a PDF, which in turn facilitates notes; For example, the PDF annotation function, which does not save the comment directly to a PDF file, but is useful for additional search and linking within Logseq.

PDFs can be added directly to Logseq or you can also use PDFs provided as Zotero attachments, because Logseq offers Zotero support. However, I have never tested this.

From now on, you can define templates, for example to define a fixed structure for notes. This can be used, for example, to determine how to take notes on a book and what block attributes should be added to the section. Like page attributes, block attributes are just predefined meta information like “alias::” to specify an alias for the page name or you can define it yourself. In this example, it could be “Author::”, “Publisher::”, “Book Title::” and “Edition::”.

There is little I would personally miss even without the plugins. Personally, I miss the built in task reminder function, because the current solution is somewhat confusing. It will also be useful to have an overview of the calendar.

So there is one less tuxedo and 4 suits remain.

Multi-platform support hardly leaves a dry eye. Only BSD users and users of exotic operating systems look at the tube, otherwise Windows, Linux, macOS, Android and iOS versions are available.

You can easily reward 5 penguins.

As mentioned earlier, Logseq works based on a file in text format. This means that any file synchronization can eventually be used for synchronization. I initially used Nextcloud sync, which resulted in errors after a while. It looks like Nextcloud and Logseq syncing started to crash because Nextcloud was still syncing the file while Logseq was still making changes. What exactly is the problem here, however, I didn’t investigate further, as I was trying to come up with a different solution anyway.

My solution is called Git. The Logseq desktop app offers built-in Git support and a time-based auto-commit function, which defaults to 60 seconds. With this solution, you have to push and pull yourself, that’s why I use Termux on my cell phone. The fact that autocommits aren’t available on mobile doesn’t bother me, since I’m running Logseq there in read mode anyway.

Additionally, Logseq developers are working on inline sync, which will likely be offered commercially so they can monetize. What the licensing model will eventually look like is not yet 100% clear. The most likely model at this point is a fully commercial one, eventually making the service commercially available and making the code open source under a license that allows for private, non-commercial use. This was at least – if I remember correctly – Ramses Odd (now Community Manager) mentioned it in one of his videos in a minor sentence. Unfortunately I could not find the source again.

In the end, I give 3 out of 5 Tuxe here.