- cross-posted to:
- hackernews@lemmy.bestiver.se
- cross-posted to:
- hackernews@lemmy.bestiver.se
You must log in or # to comment.
Tldr looks like a new community plugin discovery that requires plugins to be:
- source available
- hosted/mirrored on github
- owned by developers with Obsidian accounts
- pass automated scanning
- defined by permission manifests
Overall probably a good thing, though I wish they’d be open to featuring projects not shared with Github.
Most likely they have the tooling to automatically scan repos on GitHub specifically
Damn, I am so jaded by all the bad news that I saw this title and thought they are going to scrap plugins for something AI-related. They are not! It’s a plugin community page! Looks interesting.
The bad news is that this is a reaction to the recent incident where a Obsidian plugin contained malware and it became obvious that their plugin system is quite unsafe
Not a reaction, it has been in the works for a while.
Or kepano was lying and they built all this in a few days
I’ve been using LogSeq for a while, but there’s been a lot of positive news about Obsidian. I’ll have to check it out.
I think it is usually signaled pretty far ahead honestly when companies go rotten. Maybe it’s because I just generally choose foss lately but it seems obvious when you look at earlier choices for things like drm or licensing policies
sadly obsidian isn’t foss. i still like it though.
and yeah with the recent bambulab controversy in the 3d printing community it was noticeable years earlier if you just paid attention. but paying attention takes active effort too. being a discerning consumer/user/creator is difficult.
It’s not FOSS, but it uses markdown which is open, so even if Obsidian went down the path of enshitification, you could still (in theory) move your content over to an alternative. I know, I know, plugins, non-standard etc. etc…
I think Obsidian’s pages Just Work in Logseq, to the point you can use both programs concurrently on the same files
I use both for different use cases, so I’ve never used the same directory for an Obsidian vault and Logseq graph. Pretty sure they would both be able to read and write the markdown files, but would features like Logseq block references and queries work in Obsidian? How would Logseq treat Obsidian bases?
Yeah, this makes me trust them even though it is closed source. The data is mine and that is a big deal to the leadership there. I would never use it otherwise. And to me, that makes me want to even want to support their ecosystem and pay for actual services. It’s truly baffling sometimes how short-sighted most corps are these days. And in an age where alternatives are becoming both more numerous and often even better than their counterparts. Like Linux in general. It used to be a varying experience depending on your hardware. Even my laptop just got the correct drivers for buttons and brightness, fans and shit even though the makers made no effort at all. I’m actually optimistic about the entire cyberspace
Definitely a bait title.
I’ll stick with FOSS instead… Silverbullet.md in this use case.
I’ve been a (cautious, because not OSS and the inevitable enshittification timer) fan of Obsidian for years, and a data scientist friend switched from Notion to Logseq and raves about it. Is it that good?
Idk, I’ve never tried Obsidian so it’s hard for me to compare, essentially it’s a giant bullet-point database. You can add links or tags (they act the same) that act as references to pages, and opening those pages shows you all the references.
The idea is to write without thinking too much about it, and maybe going over it after you wrote to add links if you think it’s necessary. Personally, I wish I could join together OneNote (yes I know) and Logseq, because I like reasoning spatially and Logseq’s Whiteboards feature doesn’t click for me, but it’s good for writing down stuff in meetings and so.
The use cases are different for me, personally. There are some minor (on the surface), but major (depending on how you use the software) differences between them out of the box:
- Logseq is focused on daily journal pages and pages with lots of linkages, the idea being it keeps you on focused on tasks without feeling like you are spending too much time tinkering, organizing, or like you have to build out a perfect system right away. I like it for work because it lets me work “in my notes” rather than needing to work “on my notes” if that makes sense. I’m aware of similar Obsidian setups, but having it work out of the box in Logseq means many of the other design choices in the program are made with this workflow in mind.
- Every note is structured as a tree with hierarchical items (blocks) nested within one another. This means every block has a reference, so it’s easy to create and maintain links to different pages. Obsidian does support block references out of the box, but you have to insert them for every line you want to reference because it isn’t set up to enforce a tree structure by default.
- Tags and note links are interchangeable. I would actually say this is the main point against vanilla Logseq compared to Obsidian. There are plugins you can use to give tags different behavior to links though.
- The community seems smaller and there aren’t as many plugins. Many plugins don’t seem to be well supported or maintained, but they are usually pretty focused on solving one particular problem.
I use Logseq for work where being able to reference blocks is more useful (especially for task management), and Obsidian for personal projects where I feel a more free form PKM with customization options is nicer.
I see! Thanks for taking the time to respond. It’s useful to have usage examples.
I haven’t heard of this one! Thanks for the link, I’ll check it out.
Silverbullet is great. Plain markdown files combined with bidirectional linking, scripting for lists and content embedding and all in browser (so no sync to other clients needed). Even though since version 2 the complete sync from server to browser is a bit slow (and local https is a pain), I still prefer it to Obsidian.
I’ve been doing the SSL with Caddy and Let’s Encrypt via CertBot. Extra work but not too bad once you figure it out (and take notes since I forget by the time the renewal comes around :)
I still need to find time to set up auto renewal… One day
Thank you for your answer.
My problem is, that I would like to keep my (Docker) services like Silverbullet.md in my local network only and reach them exterally by Wireguard VPN.
Before https I reached my services with <ip>:<port>, which worked fine with VPN. As mentioned here, there are now only a few options to get https without open everything to the world:-
Set my local servers ip as localhost by ssh ( ssh -N -L 3002:localhost:3000 user@someip) which works, but needs to be added for every client and every access.
-
Add a hosts file entry (e.g. silverbullet.lan 192.168.1.123) on every client to access the local server and add a Caddyfile entry with a selfsigned certificate.
-
Put a local DNS server like dnsmasq or bind9 in my network and add local domains there (like silverbullet.lan) and use Caddy again for SSL. That’s the point where I haven’t figured out how to get it running over Wireguard VPN.
I don’t want to rely on external services like Tailscale.
I want to preface this with the fact that I am definitely NOT a networking expert so… don’t trust anything I say.
My situation is a bit different because I am using Tailscale, though I have it on the list to be replaced in the future.
When you Wireguard to your LAN, do subsequent DNS requests go through the VPN? Sounds like you’re looking into that route based on your third point above. If so, can you just add a static DNS resolution to your LAN router that points to your Caddy SSL terminator/reverse proxy? This assumes a static IP for your host.
That’s what I’ve done. On my router I’ve set a static DNS entry of
silverbullet.mydomain.com -> 10.0.0.101(where*.101is the static IP of my internal host/Caddy). This allows everything to resolve correctly when I’m physically attached to my LAN but also when connecting remotely via Tailscale.It may not be elegant, but it avoids the hassle / extra config of a local DNS server as well as the need to manage host routes on each device.
EDIT: My router is running OpenWRT but I think most consumer grade routers support static DNS routes… but I could be wrong.
-
