Finding a place to share my thoughts

May 01, 2026

I want to try writing about some thoughts or at least write up about what I know. I'd also want these thoughts to be linked in some form. I don't think something which presents itself as a web log (blog) is suitable for me. As it's information that I might need to update on months later. I'm rather "wiki-brained", I'm calling it that. As I'm used to editing pages using MediaWiki. I'd use MediaWiki for that, but I think it only works well because it's a wiki. Although extremely easy to export/import, MediaWiki isn't entirely portable to other formats (templates, modules, styling). But Markdown is pretty standard and can still be understood without a proper viewer.

Maybe I'd just think to just write a bunch of Markdown documents in a cloud drive and call it a day. It still tends to be that. But ATProto exists and as far as I'm concerned, it just stores public documents (records). I tried using git before, maybe it works, but then I'd be working with empty git messages and end up not using revisions. Lack of revisions doesn't concern me.

I did read about someone using GitHub issues for notes(where?). That's something I could understand doing. I love using GItHub issues, but the thought about all of those issues going "poof" is a bit scary. Sure, GitHub is decently reliable, and I don't think I'm the type of person who would get kicked from it. But I shouldn't have a plan to back up my online issues to a SQLite database in which I'm not sure what to do when I need it. Maybe I could try creating an empty repository on Tangled and use that? But actually, I'm not sure what changes are to come for issues on Tangled. Generally, a record stays the same, nobody can force a record to upgrade to whatever new schema an application wants to use. So, it's probably fine. But writing is a done-with problem. There's standard.site, WhiteWind and probably various others.

About those records.

_'s avatar
_
8mo

who was going to tell me that atproto wasn't a sync protocol and it's actually just a backup protocol

Depending on the purpose, there's various methods to obtain records, either existing ones or ones which are being announced to the network. This is also a pain point. I've seen some ATProto-based Astro content collections which just... seem easily abusable? Usually, I'd think for building a site, either a .car, record listing or firehose/Jetstream ingestion would suffice. But I've seen it as a record lookup with no caching. Services usually have rate limits, and I don't want other applications on my server to be knocked off because a bot wanted to look at non-existent records. I'm not fond of naming and shaming, so I'm not going to list where I observed this.

  • I'd list a site building CLI tool that I think I might've found on Tangled. But I'm not sure where that is now. oops