The main reason #Logseq is still my favourite #PKM app despite very good alternatives like #SiYuan, #Affine and #AnyType is queries.

I think one can't really appreciate Logseq until they integrate queries in their workflow.

But I hate that Simple Queries syntax is too limited and I need Advanced Queries syntax (i.e. #Datalog ) too often.

The #JsonLD context is _declarative_ as a set of directives, more like #maven or #datalog than #ant or #python.

The (frustrating) exception to this is how it has a linear-order precedence and the ability to unset things in later contexts, but that aside

If you think of it this way, then it _isn't_ a type and you can't really load it as one productively _per se_, but it may be a _series of typed statements_ (yes, that has a type, but so does a maven build file in that sense, here have a cookie)

"Been experimenting with a thing that visualizes program execution by interpreting the program in #datalog." from Pete Vilter: https://twitter.com/vilterp/status/1667985154613100546

As a still practicing academic, I don’t quite know how to formulate my tech interests just yet — so here’s a disorganized hashtag list of such things I tend to like thinking or learning about:

#csv #linux #archlinux #bsd #datalog #dlv #naturallanguageprocessing #functionalprogramming #rustlang #rustlings #python #javascript #portability #unicode #lexicography #obsidian #postgresql #gis

Learning a #programming language that challenges your notion of what programming is might be one of the most rewarding things once it "clicks" and you start seeing the bigger picture. Whether that's #Clojure, #Rust, #Haskell, #Datalog or whatnot — doesn't really matter. They're *systematic*, and when you see the system, that's empowering.

@masi

I don't know which functionalities but I can imagine. I really like software that can be remodeled by the user like #Emacs with #Lisp and I see #Logseq is on the same track with #ClojureScript, Hiccup and #Datalog for queries, it's just that the ecosystem is not yet mature but has the potential to bridge the average #PKM userbase and (Emacs) hackers 🙂

@johnny @pedersen

Finally, notice that #Logseq query language is #Datalog, an alternative to SQL that works great with graph databases, is popular in Machine Learning and it is way more powerful than SQL.

If I remember correctly the syntax by Dataview plugin for Obsidian is not a standard query language.

@anderseknert @charlieegan3 @chat_with_matt In fairness, I do agree that it is a lot like inverting your brain and a lot of engineers—many who want to get a job done who are often avoidant of SQL and have no experience with #datalog—struggle with it at first ^^

There's a steep learning curve w/ limited resources to get you where you need to be in a hurry.

On the other side I can come up with about eight million reasons I don't like doing the sorts of things that #Rego does in #Go or #Python.

@charlieegan3 @chat_with_matt

Originally based on #datalog, the language is indeed foreign to most. Once it clicks, there's no going back though :)

@hrefna recently described it better than I could: https://hachyderm.io/@hrefna/109819678091325981

Yet another new #Datalog-inspired #authorization policy language, this time from AWS: https://onecloudplease.com/blog/cedar-a-new-policy-language

An introduction to #Datalog - https://blogit.michelin.io/an-introduction-to-datalog/ #dailylinks

I wish that there were more appreciation for #prolog (and #datalog) in our industry.

Like I'm really pleased to see languages like #rego and developments along those lines, but I get so frustrated by the continued reinventing C or BASIC syntaxes for what are fundamentally declarative problems.

@gizmo

Dataview is a plugin for Obsidian, the idea is similar but properly implemented in #Logseq. Logseq is that at its core, turning text notes into a modern database.

#Datalog is the query language (like SQL but way more powerful) that Logseq use internally but it is also available to the user with so called Advanced Queries.

And being able to use #Hiccup (an alternative syntax to HTML) to style results means you as a user can remodel Logseq to your needs. It's a very innovative approach.

#Logseq is #FreeSoftware , its source code is available under #AGPL license.

It turns #Markdown or #OrgMode notes into a #GraphDatabase that can be queried using #Datalog , it supports #HTML and #Hiccup even to style query results. And this is just the surface.

@gizmo

Basically a rules engine let's me make declarations like "I am seeing this message, what do I want to do with it for handling" without writing eight million individual handlers.

If this were a serious project I would go out and grab someone else's and call it a day (or maybe #rego), but because this is a toy and I'm experimenting I'm going to create a little DSL—based on #datalog—and see if I can use that to implement some ideas in this space. I'll call it #APRL or #ActivityPubRulesLanguage.

@andric

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?

@nate @philoup @eikedrescher

This is why I love #Logseq : it stores data as Markdown or Org files but when running it turns them into a graph database you can query using #Datalog from inside the app itself or via CLI. It can even expose API as a local HTTP server so that other apps can create ad-hoc GUI or programmatically edit your notes/tasks. Its data model is so flexible that let you store notes, tasks and a relational database of arbitrary objects where you define keys and values.

When heroku killed their free tier I decided to retire https://www.learndatalogtoday.org which had been happily running mostly unattended since 2013.

But after some kind words & a little sponsorship it’s back up. Hopefully aws t2.micro is as reliable as heroku was.

#datalog #datomic #clojure

I don't know if the language is just so unintuitive or if it's just me. Has anyone ever worked with #datalog queries? My brain hurts.

@post @andric I understand your advocating.

I consider Logseq, especially the #Datalog language for queries, a tool not designed for the masses that are not programmers like me. I would call that also an experiment for a niche target.

Time will tell us.