I like that a UI for datalog queries is first class. I have two questions: 1. Can I connect a REPL to a running instance of this application and/or execute queries from a Clojure program? 2. What license is this project available under?

I am honestly not sure about the license. Leaning towards MIT. Open to feedback! I am not sure I know enough about REPLs to answer your first question. Here are some thoughts: 1. As a general rule, I want the user to be able to do as much as possible. Full ownership. If we can make connecting a REPL to a running instance a thing, we should/will. 2. The database files are local. You should certainly be able to access the DB from another Clojure program and use #CG3AM2F7V's API. 3. Running arbitrary datalog queries will be enabled from the UI itself. 4. I have to think more about how much of the 'system' should/can be accessed through #C015LCR9MHD. One idea: run a custom query, then a clojure transformation pipeline, display the result in the UI. 5. Once there's a backend API, you can certainly interact with the database that way. Does that make sense?

I mean, it's going to be open source, so you can always start the app from source files? 🙂 I know people connect to production instances, I just haven't done it and don't know what it takes.

That answers my questions! Thanks for the feedback. You might be interested in this discussion: https://clojureverse.org/t/repls-on-production-deployments/3476/4 The security model for a local app is probably much simpler than worrying about validating and proxying REPL connections to a "production" VM.

speaking of repl, i really would like to see a note app can achieve some degree of editor agnostic just the same rationale as described in clerk https://github.com/nextjournal/clerk#%EF%B8%8F-rationale A traditional web based note app suffers all the same symptoms as rationales described in clerk. And it turns out it's actually rocket science to build a really robust text editor in web. Try to press a crtl-z in logseq or notion, does it revert back in block level? or local block text level? or even between pages, those things are really tricky to handle. Not to say things like vim mode or modal editing of with each ones preference. If there'd be some kind of repl that I can just connect and send 'note block' back and forth to the app in my favorite editor, that'd solve a lot of hard to solve tech problems by nature. Long form writing would be a delight, edit datalog query would be a delight, etc, etc..

First of all, let me say I'm very sympathetic to what you wrote @U0599HVJX70. It's a thought that pops up every once in a while. Don't know if you've seen https://www.dendron.so/? Though, that's VS Code specific, I think. Brimm doesn't model its data as text. This is one of the first decisions I made, even though I recognise the benefits. I wanted to deal with structured data and trying to fit that into plaintext seemed too limiting and awkward. @UFTRLDZEW laid it out nicely. I also think there is an opportunity to lean into SCI and https://github.com/nextjournal/clojure-mode to bring a sort of 'REPL for blocks' into the default Brimm UI. Having said that, Brimm leaning into the backend means it should be feasible to write editor extensions that do what you describe. I have dreams of parsing my codebase into Brimm blocks. Example: turning TODO comments into tasks, living side by side with the rest of that project's management, comments, etc.

@U0599HVJX70 you may also be interested in https://github.com/artempyanykh/marksman, which provides components of an editor-agnostic note-taking system using a novel LSP implementation for elements of Markdown.