RT @fortelabs@twitter.com
We really need a graph-based PKM app built by a serious company for mainstream users
• Roam is a religion
• Obsidian is a side hobby
• Logseq is an open-source project
• Tana is barely in alpha
If there was just one reliable app in this space I’d refer so much traffic to it
: 
Does this person really think being Open Source is a bad thing?
Anyway you seem to be talking about graph databases: why the user should care if an app uses a graph database or a relational database?
Instead in that thread it seems they are talking about apps having a graph view, that is cool but who cares, it's not a key feature.
@post Well that person is mostly trying to engagement-farm, but also in doing so has tapped into a common feeling that “graph”-based apps need to be easier to use.
@post In terms of why should users care, well, the same reason why a business that’s evaluating which database to use cares about different DB paradigms, but at the scale and scope of an individual’s data needs: the paradigms make tradeoffs in terms of how you can model data.
End-user databases like Notion, Airtable, Roam, Tana are domain-general, so the user is responsible for data modeling. If the user is modeling data for their domain, then paradigm absolutely matters to the user.
If I understand it correctly you are referring to the flexibility enabled by graph databases and how some PKM apps are taking advantage of it to let users define arbitrary schemas and ontologies.
But I think the people in that thread are thinking of graph-based apps as those where you link notes together in a network/graph and some of them even have a graphical visualization, like Logseq and Obsidian.
In both cases you are right that those apps need better UI/UX, despite you are referring to them with a different criterion.
@post I don’t think you need a graph database to let users define their own schemas and ontologies. You can do that with a more relational-style database, and that’s what Airtable and Notion found huge success with (among mainstream users, too).
LogSeq and Obsidian currently feature an incomplete featureset (requiring a plugin whereby users have to write queries in a non-standard query language) for doing this, and isn’t suitable for “mainstream” users.
@post And yes, the people in that thread are referring (perhaps somewhat incorrectly, as triple-stores are not exactly graph DBs) to apps like Roam, LogSeq, Obsidian, and Tana that let you link notes with each other and then query it using their built-in backlinks feature or via a non-standard query language.
They’re end-user databases, or at least aspire to be.
@post My assessment is that the design space for end-user database UX is constrained by the underlying query and storage engines used.
End-user databases necessarily need to be local-first, but options for local-first databases that replicate well with a remote graph database are a null set at the moment.
Local triple stores support the backlinked notes use case, but I doubt apps can go much further than whatever Tana’s built.
At the moment #Logseq uses #DataScript that runs in browsers/Electron so it uses #Datalog / #Datomic for its internal and user queries.
I believe they want to use something native like this in the future: https://github.com/cozodb/cozo
Also Logseq uses Markdown/Org files to store data and they are working on adding EDN files too.
Anything else they could do?
@post Thanks for sharing a link to Cozo! I’ve never heard of it before, but something like this would massively improve existing apps by being a full-fledged graph DB that’s embeddable across mobile devices and web browsers, while still being able to run on a server.