I think you misunderstood my point. The fact that the Hover Editor is forced to use non-API functionality means that it will be very difficult to maintain the plugin, Hover Editor. It will frequently break because it is using functionality that is not guaranteed to remain stable. Thus even though there is a plugin that (to a small degree) resolves this, it is still important that this feature be either incorporated into the API or implemented as a native feature.
To your second point, I’m saying that using Hover Editor is quite distracting. It’s very different in my opinion from natively implemented transclusions as you can experience in Roam, Notion, and many other apps. It is a UI issue for the most part but simply having to interact through a separate window is quite distracting when I’m trying to edit an outline.
Not to mention that:
- It does not work on mobile, and I do at least a quarter of my note taking on mobile
- There are many features that don’t work with it, like “find”
- I basically exclusively use the keyboard when writing, using keyboard shortcuts and the cursor for navigation. I very rarely use the mouse. This is probably why the hover editor is very foreign to me. When I go to make a change to a transcluded block, with the cursor, it simply collapses back to the line of code that has the embedding. Then I have to go back up, then hover over with my mouse, etc etc.
Are any of these absolutely a disaster? No. But to me transclusions are really useful. And once you’ve used a really smooth implementation like Roam’s it’s hard to use this implementation and find it to be pleasant.
To be clear I love obsidian. It’s just one of the things I miss. And I strongly disagree that this is something that comes down to a philosophical issue between obsidian and roam. Transclusions do not require a database or a block-based system. The fact that they are able to very efficiently render the embedding, means that it can be edited. And the fact that embedding seems to be quite popular implies to me that it would be worthwhile to extend it. I think that most people would prefer embeddings to be editable and the vast majority of the rest dont mind either way. So if it’s not overwhelming technically I think it’s a no brainer in terms of being a non-controversial feature addition.
Not saying it’ll be trivial by any means. Just that I think it’s something that would be valuable to have and that few if any would be unhappy about