Notices by silverpill (silverpill@mitra.social), page 28

@erlend Comparisons between centralized services and nodes in a distributed network are not meaningful. Comparisons between a specialized relay service and a full social media server are also not meaningful. And finally, it is not ActivityPub that is expensive, but Mastodon. Other implementations are cheaper, some are 10-100x cheaper

@maegul @fediverse Some ActivityPub implementations already work as social media browsers. For example, my server can interact with microblogs, but also forums, blogs, events etc. The more activity / object types are supported, the closer software is to a browser.

@nimda I found another probable cause of that. The fix has been included in 2.25.1

https://mitra.social/objects/0190a13a-7352-ecdf-3ee5-5501a6a17de2

Mitra v2.25.0 includes a functioning FEP-ef61 gateway, which portable actors can use to communicate with the rest of Fediverse.

I'm currently using a command line FEP-ae97 client, but it is very primitive, so before migrating to a nomadic identity I need to develop a better client.

One nice property of FEP-ae97 is that clients and servers are not very different. In theory one can even take any existing server like Mastodon and convert it into a client, because client-to-server and server-to-server APIs use the same data format. Servers send activities to inboxes located on other servers, but clients need to send activities to outboxes located on gateways.

This is exactly what I'm considering doing with #MItra. A special build that will work as an offline-first desktop client. Alternatively, FEP-ae97 can be implemented in mitra-web frontend, this option is less interesting, but will not require installation.

RE: https://mitra.social/objects/01909e2f-3cbd-7ffb-a0ea-4b3f09fd1aad

@frogzone

>inflation bug

https://www.moneroinflation.com/

>i'm unaware of any layer 2 for monero that would help it scale

Several solutions were proposed, but none of them are being worked on because there is no urgency. Unless Monero growth becomes explosive (unlikely), the flexible blocksize is a good enough scaling solution.

>bitcoin is as private as a person makes it

Not really, because all transactions are public. You can only make your transaction history harder to analyze, but this is not easy.

L2 systems can be private by default, so not all hope is lost.

>btcpayserver (L2)

btcpayserver is a payment processor, not an L2 system. Perhaps you meant Lightning Network? It borrows some ideas from Tor, but it doesn't operate over Tor by default.

@tadano @Tadano

>Done https://github.com/captainepoch/husky/issues/75

I attempted to fix the problem with replies in v2.25.0

@frogzone There's no such option, but I'm not against including it.

Bitcoin itself lacks privacy, so using it for payments is not a good idea, but recently some interesting L2 solutions started to appear, like Cashu. If there will be enough demand I can support them.

@nimda Yes, detroitriotcity being unresponsive is a possible cause but I agree that 107 timeouts is probably not enough. Do you have any messages with ERROR level in the log?

@sun @arcanicanis @PurpCat @frogzone @Morghur One of the first features I added to Mitra! Token metadata was stored in IPFS and contained a link to tokenized post.

https://codeberg.org/silverpill/mitra-contracts/src/branch/main/contracts/Adapter.sol#L96-L116

@sun NFTs were quietly removed several months ago. Everyone hated them.

(But the code can still be found in the repo.)

@arcanicanis @PurpCat @frogzone @Morghur

@streams @mikedev For the last week or so I was experiencing federation issues between my server and fediversity.site (couldn't verify integrity proofs on activities).

I can see this post, though.

@frogzone

>when it comes to decentralized hosting, for example if your server goes down, then i've heard good things about IPFS and tahoe-lafs

Have a look at this: https://codeberg.org/fediverse/fep/src/branch/main/fep/ef61/fep-ef61.md. With this federation protocol improvement we can have redundancy in ActivityPub (i.e. if your server goes down you can switch to another one without losing your identity). It will nicely work with I2P too.

>it can be linked to a bitcoin or monero wallet (or lightning node?

Monero and Ethereum, but Ethereum will be removed very soon.

@arcanicanis @PurpCat @Morghur

@nimda Other instance operators reported similar issues. If you see a lot of "request timeout" messages in your log, increasing queue size probably won't help.

@shreyan Most of what you are talking about in "ActivityPub" section is only true for Mastodon. #ActivityPub is much more capable.

>identity is extremely tied to your initial server
>Your data is not really portable

ActivityPub specification doesn't say that identity must be tied to a server. Identity can be tied to a secret key or a DID too, and users with different identity types can even communicate with each other.

Specifically, DID-based identity scheme has been proposed in FEP-ef61. Data portability easily follows from it.

So, with ActivityPub one can have everything that ATProto and Nostr offer, and much more.

@fedify This value of actor's icon property may be not valid:

According to ActivityStreams Vocabulary, it should be either Object or Link, but in this case it is simply an URI of an image file (I guess it has xsd:anyURI type?).

FEP-c7d3 (Ownership) has been updated

https://codeberg.org/fediverse/fep/pulls/349/files

>Object identifiers are grouped together into protection domains called "origins". This concept is similar to the "web origin" concept described in [RFC-6454], and origins of object IDs are computed by the same algorithm.

@nomad now has inbox and can receive activities from other servers. Outbox forwarding was also implemented. I have successfully sent messages from FEP-ae97 client to users on Pleroma and Mastodon servers.

I think this proves that full data portability is possible with ActivityPub, and that it can be implemented in a backward compatible way.

#fep_ef61

@pomstan I'm saying that Mitra uses custom logic when it serves web client, so you theory about actix and default handler is probably not correct.
I don't know why it doesn't serve your static files without adding trailing slash. The configuration example works for me and apparently for everyone else too.

Could you tell me what are you trying to do?

@0x1C3B00DA @newt +1 for using ActivityPub IDs.

Somebody should write a FEP

@xianc78 @gabriel @servo As far as I know, Rust was created to build Servo.