I’ve heard about the app Tana, which everybody talks about.
Comming from logseq to notion, I feel like Tana offers best of both worlds.
What are some of the things that Logseq is better at than Tana?
I’ve heard about the app Tana, which everybody talks about.
Never heard of it before, but from a quick glance;
- not open source
- not local-first (does it even work offline?)
- looks like the format they use to store notes is proprietary, unlike logseq which uses a superset of Markdown
These new apps are making bold claims like “future of computing” for their notes app. I saw a new bookmarking app that’s doing something similar. That is a red flag to me.
One of the biggest draws for me to Logseq is that I own my data; literally. How does one get their notes out of Tana?
Tana and Logseq are completely different.
Tana is basically a Relational Database (think of Excel/Access) visualized in different ways like Notion and to input data you use an outliner UI like the one in Logseq.
Tana should be seen as an alternative to Notion more than to Logseq, that has more word processing features like headings.
In Tana these strings of text (that can be formatted only in bold, italic and underline) are named “nodes” and organized in a single giant tree.
In Logseq instead there are a set of pages stored as Markdown/Org files that are treated as the nodes of a graph. Then each page has its own tree of blocks.
So from the UI it seems Tana nodes = Logseq blocks, but conceptually Tana nodes are treated like Logseq pages and referenced around the giant tree.
In Logseq there is no hierarchy between pages or other structure (aside from the network of references). In Tana the tree structure for nodes is enforced.
While Logseq focuses on looking for (eventually long) paragraphs of text in blocks containing [[wikilinks]] and #hashtags using queries, Tana focuses on specify the relations between its nodes by placing refereces inside so called “fields”.
Fields work basically like properties:: in Logseq, but in Tana the concept is more developed. Logseq will eventually catch up in term of feature here and there is already some WIP in a branch called enhance-properties.
Additionally, Tana uses #hashtags not as tags but to assign “classes” to nodes. They call them “supertags”. By belonging to a class, a node get a default template of fields and other things.
Since fields can have a “type” i.e. you can choose what kind of input it accepts, you can set a field to allow only instances of a certain class, so only some specific nodes instead of arbitrary text.
Tana has also a query builder, different ways to visualize blocks (indented list by default, table, kanboard, tabs) and an overview of the so called “schema” i.e. the system of classes and fields mentioned above.
Then Tana as a lot of advanced options regarding supertags.
And that’s it, this is Tana, nothing else. No word processing features, no Markdown/Org, no PDF annotations, no flashcards, no whiteboards and especially: no plugins, no custom CSS and JS, no macros, no custom commands, no local storage, no offline mode and proprietary code and data format.
So consider Logseq like an all-encompassing integrated environment that you can hack to do whatever you want but that still didn’t develop some features to their full potential, namely properties, query building and alternative block views.
Instead Tana is an experiment to use an outliner structure to organize what will be treated as records of a RDB, to be visualized in multiple ways like in Notion and using queries. The concpet is already there in Logseq’s properties but not much developed.
At the moment Tana is just a proof of concept of this idea and combos that could make sense in the future would be:
- Obsidian + Notion
- Obsidian + Tana
- Logseq + Tana
but exchanging data between apps like these is hard and once Logseq develop its properties to catch up with Tana, the best option will just be:
Thank you for your well written reply. I really hadn’t expected anyone to write such a great, I think quite objective review.
As for PDF annotations and whiteboards, I think that that is better to be done using some dedicated app.
Do you think Logseq will get something similar, I don’t think it’s in the Roadmap.
Also, you seem to be knowledgable about logseq so if you don’t mind can I ask:
Why is Logseq beccoming more of an all-in-one, I don’t have problems with that, it’s just that I think with that It’s stopping the development of innovative features that would allow for better knowledge management.
But then they wouldn’t be well integrated into Logseq, defeating the purpose of having all information in one place. Maybe you should play more with these features and see how well they are integrated with the outliner.
It’s already in enhance-properties branch on GitHub. See:
Because exchanging data between apps and make them work well together in general is hard.
Logseq is FOSS and everyone can contribute. It has very general features that can be adapted to many different workflows. It has a powerful plugin platform plus user custom CSS and JS. When it comes to FOSS, it’s important to provide an actual platform users can experiment with, it will attract more contributors and make the project evolve much faster than any proprietary product with only a team working on it.
That is true, but I had used them, yet I don’t find them as good as other apps, that focus solely on that. Even the current newer whiteboard, is lacking a lot of features that would make it a great note taking experience, so It’s easier to just put PDFs of notes in Logseq. About PDFs, it’s useful, but I’m using it just using to be able to read PDFs, so I can’t comment on that one.
That is generally true, but I think note taking isn’t one of those things. However for annotating books, It’s probably very good, but I haven’t checked it yet.
I don’t know much about FOSS, and open source projects. But, as far as I know, don’t all the changes have to be approved?
To finish up, I forgot to ask before. Will Logseq get something like supertags?
Of course and nothing stop you from using them too, but if you want your PDF annotations referenced directly in your notes, Logseq does a pretty good job. Instead whiteboards are to organize spatially mainly Logseq entities like blocks and pages, so it’s different from any other canvas application that works with its own entities.
The policy can change a lot from one project to another but Logseq is one of the most welcoming ones I know: as soon as your code meets quality standards and makes sense it will be accepted. In case a feature is very specific and doesn’t make sense to be included in Logseq itself, you probably will be able to develop a plugin instead. It’s hard to imagine a more open platform than Logseq, you can hack it on many levels as a user, as a Web developer or as a Clojure developer.
Ben developed a lot of plugins and now he is working on one inspired by Tana supertags:
But I think at a certain point Logseq may have something even better because what you can do with supertags makes sense a lot in Logseq but the implementation should be better designed to integrate in Logseq UI/UX in my opinion.
Okay thank you, you have convinced me, I’m sticking with Logseq. Just what do you mean with
Do you mean that: https://twitter.com/logseq/status/1605224589046386689?
I see that is really helpful, not directly for making notes, but for linking them and having better control over them. Awesome.
Just now I’m seeing how much functionality of Logseq I’m not using.
Thanks for your help and opening my eyes(that sounds far more religious than I wanted).
I’ll mark your answer as solution, as I think that helps you.
This is the correct answer.
I would try a shorter one with to me tana looks like block based graph (nodes) crossed with data structures (supertags). That’s why some people refer to it as if notion and roam had a baby.
There are some cool things you can do with it. I have an account and have played around with it but am nervous of how proprietry it is as per @alex0’s writeup.
At this point i am sticking with Logseq as owning your notes is a big deal to me.
It’s also important for me, because I can make a bot that resembles me at least a little if I know I’ll die..
But, seriously I do mean it would be nice to create a bot using this data.
Excellent summary of both apps.
There is an additional good feedback on Discord
Such a great in-depth review, tyvm
I find the hype surrounding Tana more daunting than inviting. It says everywhere that it’s the tool “everyone” is talking about. And then it turns out that only those who claim that everyone is talking about it are talking about it. Like they’re getting paid to make nice weather for Tana. The professional evangelists who chase after every hype, and when there isn’t any, create hype themselves.
As long as Logseq gives the right answer to some of the good features in Tana, they can do whatever marketing and it still should be in a good position. I summarized some of my pain points in What I am missing from Tana, if interested. In the end the fundamental difference between both products for me is: one is proprietary,online and the other one FOSS, local-first.
There is no question a tool needs enough features to be useful. But there is a breaking point between enough and gratuitous. Most people don’t care to move beyond a simplistic workflow into more complex ones for getting things done. Thus, it’s like the days when features sold DVD players but everyone basically put in the disc and just hit play.
Loqseq has a sufficient set of features to be useful as is, even if it never added another. Other tools will eclipse it, certainly. But Tana will eventually have a price tag. Logseq won’t. It’s local-first. You own your data and your data is human readable.
In my experience, the tool rarely determines a person’s productivity. The person does.