Sure, but I will reiterate my argument that this kind of workflow is aimed at curating the Library, which is not the default workflow for many Tana users. If you reference a node that exists e.g. in your Today page (which is what I frequently do when working off daily notes), it will not appear inside the Library, and all future references to that same node will appear in that day’s page. Personally, I never touched the Library in Tana, but my workflow differs from yours.
It’s also why I agree with your observations. But Logseq, like its sibling Roam Research, pages as containers for blocks are fundamentally without hierarchy, and this is by design[1], [2]. They are knowledge management tools, not outliners (even though they use outliners as their main text input interface). In fact, I would argue that compared to specialised outliners, they perform quite poorly but excel in other areas.
That said, I don’t quite understand why the DB version still includes pages as the top hierarchical unit. Maybe pages are necessary to maintain the fabled graph view, which has always been toted as one of the selling points of RR/LS/Obsidian etc., or maybe there’s another explanation I’m not aware of. Either way, I’ve always seen these tools as knowledge managers, not first-class outliners. So, I refer back to my original argument that it comes down to managing trade-offs. I’m sure you’ve felt this if you’ve considered switching to Tana.
But that is just one single use case, where you create an “inline” new node that creates a reference in your node and place the source within the Library.
You are correct, however, that is not my concern nor is it what i’m talking about. That same workflow is the same as it is today in Logseq, the choice of tying a reference back to a block or a page is on the user. With the DB version the page/block references have now been unified into a single lookup function so there’s no longer a distinction between the two, which makes it even more prominent. My issue is with the decentralized approach to storing node references. Logseq has a scattered/fragmented approach where page references can be scattered throughout the graph, but Tana focuses on using a centralized approach by storing all references in the library making it easier to see all your references, but also organizing them. They are nearly identical with Logseq, except Logseq only allows you to organize references through the use of properties, not in a traditional outliner method which is IMO a step backwards. Tools should be making things simpler, not more difficult.
The whole page thing could still be up in the air with the DB version. It’s changing rapidly with each day, so who knows. This is kind of my point in that Logseq started out as an outliner, but its moved away from being an outliner, they’ve actually removed a lot of references to the word “outliner” from their documentation for this very reason and its one of my growing concerns and why I have made this post. Folks can disagree or hate it, but this is the feedback section of the forum and this is my feedback as a 3+ year user and why i’m going to have to move away from Logseq. At this point i’m kind of in the same boat as Alex and moving back to using raw text files (Foam), along with Obsidian on the side for the Canvas feature when I need something with more of a visual approach. While Tana is great, I will not use any type of online service where my data is stored in their cloud.
That’s fair. I just brought here the additional information about Tana’s set of features regarding node creation and reference, since there is not just one single way.
Consider creating a new subject spin-off for a more fit discussion since the current thread hangs from DV Version and Future State from a 3+ year user.
Hey David - you and others are free to discuss this in detail if you like. I posted this in the Feedback section, and it was specifically intended as feedback for the LS team. I’ve responded to clear up any confusion around the points I’ve made where I think others were not grasping what I was trying to say, and I’m happy to provide further clarification if the LS team has any questions. But beyond that, I think I’ve said what I needed to.
Thanks for chiming in and sharing your thoughts - I’ll leave it here.