Notices by Rimu (rimu@mastodon.nzoss.nz)

@HolosSocial I like the approach you're taking with implementing E2EE, seems solid to me!

@ricferrer @julian @rimu@piefed.social @evan Here's a video of web+ap url handling in PieFed

https://peertube.wtf/w/7VwgZJ9UkH3REbcngaKn4Z

I'm not claiming it's the **best** solution but thought it was worth trying out.

@ricferrer @julian @rimu@piefed.social @evan Instead of putting web+ap links everywhere, PieFed just silently rewrites urls in post bodies so they go to the local copy of each post, if we have it.

In the threadiverse every instance has every post so this works pretty well.

Nice to see #Mastodon copying PieFed's geographical-based instance choosing mechanism.

https://piefed.social/auth/instance_chooser

A lot of people made new years resolutions about using social media less. So I'm adding an optional daily time limit to #PieFed!

One nice thing about the #fediverse is we don't need to optimize for addiction.

The warning pops up with every page load so just clicking through it once won't work. It's for your own good!

If you could access your #PieFed
instance using your Usenet client of choice (e.g. Thunderbird), would you?

#nntp #fediverse

@tchambers Mastodon just backfills on demand (when someone views a post) so their solution just papers over the issue.

When receiving a post from outside Mastodon already does a GET for links in a post in order to construct the preview card (using open graph metadata) so they could, at that point, realize that it is a fedi link, backfill the link to constrict a local copy of the post (if necessary) and then change the original post link to that new copy.

One would think.

@tchambers Yes, with limitations.

You can test it on https://crust.piefed.social as that is the bleeding-edge instance - it's not formally released yet.

On the threadiverse nearly everything is federated to nearly everyone so nearly all instances have most content. This makes it a simple matter of detecting links that go to fediverse instances and rewriting them to go to the local copy.

It wouldn't work so great with links to Mastodon content because chances are we wouldn't have it locally.

@wjmaggos Something like this is a good start.

https://piefed.social/post/1489573

In the past I've coded lots of things for blocking various things, which is all about excluding what we don't want.

I'm finding it much harder to think of ways to attract and promote things we DO want.

In an ideal utopia we'd be making social media that makes people feel better. Not neutral or the same as before and certainly not worse than before. I don't think we can get there by excluding bad things, we need to be including/nurturing/promoting good things too.

#fediverse #fedidev

#PieFed now has friendly looking urls for posts, like this one:

https://crust.piefed.social/c/nottheonion@lemmy.world/p/58005/top-white-house-official-selling-arlington-home-after-activists-write-chalk-messa

Now you can tell at a glance that the link goes to a post in a community called 'not the onion' and is something about an official selling their house.

Much better than https://crust.piefed.social/post/58005

#fediverse

@julian nothing good...

@tchambers @julian @silverpill I put together a prototype, using web+ap.

Here’s a video demo - https://peertube.wtf/w/r8M4KpS6yLd9H2r2oPAs9M

The workflow is a wee bit clunky, with the browser popping up things more than I'd like, but it works.

#PieFed 1.2 has just been released.

- Event posts
- 3D rotating tag cloud
- Built-in translation
- Set a reminder on any post/comment
- Log in with LDAP
- and so much more.

Details and links to examples at https://piefed.social/post/1278967

#fediverse

@tchambers @js I'm going to look into it.

Hard to imagine how this could work while still allowing usage by regular people AND avoiding scrapers that pretend to be regular people. But let's see.

@tchambers @js

Ok I've looked into it.

This is just robots.txt on steroids in the sense that it's entirely opt-in and only binds law-abiding actors. It has no answer to the badly-behaved scrapers that ignore robots.txt and overwhelm our instances.

Having said that it will still be great to have a way to bill the 'good' crawlers and I appreciate the lightweight and simple methods they propose. It might work.

Or it could just mean the incentive for crawlers to spoof user-agents is higher...

@renchap @tchambers It is opt in, yes.

The master list of PieFed instances is retrieved from the fediverse.observer API. Then an API endpoint on every instance in the list is queried to see if they opted in and to get information about the instance that fediverse.observer does not have.

This happens once per day.

Instance admins can choose to filter out any other instance and defederated ones are automatically filtered.

@tchambers Cheers :)

I'm hoping to maximize the dispersal of new user across all the instances so wanted to keep other options visible from the beginning.

Can't wait to roll this out more widely. A key part of it is that it's built-in to every #PieFed instance, not a separate web site. So it's power comes from everyone being on board.

https://crust.piefed.social/auth/instance_chooser

Does anyone know which #fediverse platforms return HTTP status codes according to https://www.w3.org/wiki/ActivityPub/Primer/HTTP_status_codes_for_delivery ?

Is there a table / list documenting the differences, anywhere?

#ActivityPub

@sashin PieFed automatically adds archive.ph links to posts that link to paywalled articles.