Notices by Andrew (R.S Admin) (ajroach42@retro.social), page 35

@djsundog @thegibson @freakazoid
At least, that's what I was envisioning originally, and I think it aligns with what you described.

@freakazoid

So you have a tool that generates and signs the files. Sticks them in the right folder structure. You put those in whatever places you want.

The receiver retrieves them. They might retrieve the delta, or the whole folder depending on how it is being distributed.

They store and forward.

Then the receiver uses another tool to browse the updates/generate html or emails or an epub or whatever from the updates.

@thegibson @djsundog

@freakazoid
Ah. Okay, I think I follow.
@thegibson @djsundog

@freakazoid
That's the idea, yeah. The transport mechanism can be basically anything.

A folder that's being served over http. An "outbox" folder that will rsync to other machines on the LAN. A thumb drive that'll get dropped in a mailbox.

The files don't care.

@djsundog @thegibson

You can't unsend an email.

@estoricru @68km

I can describe briefly the things I'm trying to build if you think you'll have time to try and whip something together faster than I can. (which, I'm a not great dev and also very busy, so you probably can.)

Want that I should do that?

@freakazoid @thegibson @djsundog Alright, as I read through this thread more I really like where this is going.

@qwazix No one needs to open the plaintext files directly.

The idea is to separate the content from the display from the distribution.

The display layer takes the plain text files and parses them in to [rss|html|email|twitter|etc] and displays them to the user.

Working in plaintext, or rather in markdown, makes development easier, but that's all.

@freakazoid @thegibson @djsundog I like this a lot, but I'm concerned because it means that you're always required to distribute the full chain of messages, which might add up to several MB?

Ahhh, lots of messages.

Catching up.

@Username_Here_ASAP Go follow @Curator

@Username_Here_ASAP @skeezypeach @Ethancdavenport mastodon.art

@freakazoid @djsundog @thegibson That is a lot closer to the vision I had, but I'm having a hard time wrapping my head around how Trust would work in that system.

@thegibson @emsenn @Ethancdavenport @hummingrain In spite of my distaste for Node and Electron, I'm going to try to get the server portion of Dat running on my raspberry pi.

If it works at any kind of a reasonable speed I'm probably going to be lobbying that we us Dat.

@hummingrain @Ethancdavenport @emsenn @thegibson That's kind of how Dat works.

I wish their software wasn't so heavy. If there was a good reference implementation of Dat/Beaker in something other than Node (preferably one that entirely divorces the browser from the server, honestly) we'd probably be set.

If I could browse Dat sites from my mobile device, or even just download the contents of a Dat file from my mobile device, I'd be like 80% to where I want to be.

@yam655
Neat! I'd love to learn more.

@nonlinear This looks neat.

It also looks entirely reliant on existing internet infrastructure with no mechanism for existing outside of that structure (the username is the URL, for ex.)

I want something that doesn't require the internet.

@thegibson @Ethancdavenport @hummingrain @emsenn Trying not to centralize, yeah?