In-reply-to » @prologic @bender @eapl.me I think opening another file is a bad idea because it adds complexity to the clients, breaks the single feed and I think keeping legacy clients will be more complex to add new features in the future. A modern approach is important. I'll be honest, I'm a bit tired of the fight around the direct message. Perhaps, we can remove it as an extension and use the alternative @prologic . My suggestion apparently doesn't like to the community. I have no problem with remove it.

it should be
# 2025-04-13T11:02:12+02:00 !<dm-echo https://dm-echo.andros.dev/twtxt.txt> U2FsdGVkX1+QmwBNmk9Yu9jvazVRFPS2TGJRGle/BDDzFult6zCtxNhJrV0g+sx0EIKbjL2a9QpCT5C0Z2qWvw==

⤋ Read More
In-reply-to » @prologic @bender @eapl.me I think opening another file is a bad idea because it adds complexity to the clients, breaks the single feed and I think keeping legacy clients will be more complex to add new features in the future. A modern approach is important. I'll be honest, I'm a bit tired of the fight around the direct message. Perhaps, we can remove it as an extension and use the alternative @prologic . My suggestion apparently doesn't like to the community. I have no problem with remove it.

my main itch with the DMs extensions is that these messages are intended to be private, not public information. That’s why other extensions make sense, but DMs are another kind of feature.
TwiXter, Mastodon, FB and some other services usually hide the DMs in another section, so they are not mixed with the public timeline.

I find the DM topic interesting, I even made an indie experiment for a centralized messaging system here https://github.com/eapl-gemugami/owl.
Although, as I’ve said a few times here, I’m not particularly interested in supporting it on microblogging, as I don’t use it that much. In the rare case I’ve used them, I don’t have to manage public and private keys, and finally none of my acquaintances use encrypted email.
Nothing personal against anyone, and although I like to debate and even fight, it’s not the case here. This proposal is the only one allowing DMs on twtxt, and if the community wants it, I’ll support it, with my personal input, of course.

A good approach I could find with a good compromise between compatibility with current clients and keeping these messages private is ‘hiding’ the DMs in comments. For example:
# 2025-04-13T11:02:12+02:00 !<dm-echo https://dm-echo.andros.dev/twtxt.txt> U2FsdGVkX1+QmwBNmk9Yu9jvazVRFPS2TGJRGle/BDDzFult6zCtxNhJrV0g+sx0EIKbjL2a9QpCT5C0Z2qWvw==

⤋ Read More
In-reply-to » @prologic @bender @eapl.me I think opening another file is a bad idea because it adds complexity to the clients, breaks the single feed and I think keeping legacy clients will be more complex to add new features in the future. A modern approach is important. I'll be honest, I'm a bit tired of the fight around the direct message. Perhaps, we can remove it as an extension and use the alternative @prologic . My suggestion apparently doesn't like to the community. I have no problem with remove it.

@andros@twtxt.andros.dev how often do you send a private message on the Fediverse? How often do you send PGP/SMIME encrypted emails? Are there other tools that are more suitable for the task? If implementing direct/private messages on twtxt scratches an itch (you know, that hobbyist itch we all get from time to time), then don’t give up so easily. Worse comes to worse, and your feed becomes too noisy, people can simply unfollow/mute.

I really don’t care about direct messages here, but I might be on that bottom 1%!

⤋ Read More
In-reply-to » @eapl.me When it is up and running, I promise to add it to the specification. I will also include some corrections. The nature of twtxt does not allow us to selectively hide clients. It's a problem not with DM, but with any extension. @prologic Yes, it is a security hole. All dm-echo messages are readable. I intend it to be a debugging tool. Maybe I can include a warning message. If many of you see that it is a serious problem, I can remove the links. @xuu It's already much better than Mastodon :P . Maybe we can remove the sender and receiver references with an intermediary register.

@prologic@twtxt.net @bender@twtxt.net @eapl.me@eapl.me I think opening another file is a bad idea because it adds complexity to the clients, breaks the single feed and I think keeping legacy clients will be more complex to add new features in the future. A modern approach is important.
I’ll be honest, I’m a bit tired of the fight around the direct message. Perhaps, we can remove it as an extension and use the alternative @prologic@twtxt.net . My suggestion apparently doesn’t like to the community. I have no problem with remove it.

⤋ Read More
In-reply-to » @kate @eldersnake @abucci -- I've already spoken to @xuu on IRC about this, but the new SqliteCache backend I'm working on here, what are your thoughts regarding mgirations from old MemoryCache (which is now gone in the codebase in this branch). Do you care to migrate at all, or just let the pod re-fetch all feeds? 🤔

I don’t think I’d personally be worried about migrating, just re-fetch. Sounds cleaner anyway?
Sorry I’m late to the party!

⤋ Read More
In-reply-to » @eapl.me When it is up and running, I promise to add it to the specification. I will also include some corrections. The nature of twtxt does not allow us to selectively hide clients. It's a problem not with DM, but with any extension. @prologic Yes, it is a security hole. All dm-echo messages are readable. I intend it to be a debugging tool. Maybe I can include a warning message. If many of you see that it is a serious problem, I can remove the links. @xuu It's already much better than Mastodon :P . Maybe we can remove the sender and receiver references with an intermediary register.

I think that having a dm-only.txt, twtxt-dm.txt or any other name for an alternative file is the most sensible approach. The name could be specified in the metadata.

⤋ Read More
In-reply-to » @eapl.me When it is up and running, I promise to add it to the specification. I will also include some corrections. The nature of twtxt does not allow us to selectively hide clients. It's a problem not with DM, but with any extension. @prologic Yes, it is a security hole. All dm-echo messages are readable. I intend it to be a debugging tool. Maybe I can include a warning message. If many of you see that it is a serious problem, I can remove the links. @xuu It's already much better than Mastodon :P . Maybe we can remove the sender and receiver references with an intermediary register.

@andros@twtxt.andros.dev maybe create a separate, completely distinct feed for DM? That way, clients do not need to do anything, only those wanted to “talk in private” follow themselves, using their very special dm-only.txt feeds. 😂

⤋ Read More
In-reply-to » We’ve been on a trip to another city this weekend and one thing struck me as really odd:

@movq@www.uninformativ.de You just have to dress yourself up in a dog custume to pee on a tree, fence post or house wall. :-P

Depending on the hour, town halls and also graveyards are typically good options for public toilets. But yeah, you have to find them first. And then, there might be the “Nette Toilette”: https://www.die-nette-toilette.de/ (Unfortunately, you can’t see on the website which towns and cities participate in that concept. :-()

⤋ Read More

We’ve been on a trip to another city this weekend and one thing struck me as really odd:

The lack of “public water” in our cities.

Almost no way of washing your hands or going to the toilet or whatever. You can’t even pee in a bush, because a) that’s illegal and b) there are no bushes!

(It’s always been that way, I just never noticed before. 🥴)

⤋ Read More
In-reply-to » @eapl.me You asked me for private keys for testing purposes. I have added it to the bottom of this page: https://dm-echo.andros.dev/ It will soon be running. It won't be long now.

@eapl.me@eapl.me When it is up and running, I promise to add it to the specification. I will also include some corrections.
The nature of twtxt does not allow us to selectively hide clients. It’s a problem not with DM, but with any extension.
@prologic@twtxt.net Yes, it is a security hole. All dm-echo messages are readable. I intend it to be a debugging tool. Maybe I can include a warning message. If many of you see that it is a serious problem, I can remove the links.
@xuu@txt.sour.is It’s already much better than Mastodon :P . Maybe we can remove the sender and receiver references with an intermediary register.

⤋ Read More
In-reply-to » @eapl.me You asked me for private keys for testing purposes. I have added it to the bottom of this page: https://dm-echo.andros.dev/ It will soon be running. It won't be long now.

not a big deal as I can skip those messages, but again, it’s an extension, so older clients shouldn’t be affected by a new feature.

⤋ Read More
In-reply-to » @eapl.me You asked me for private keys for testing purposes. I have added it to the bottom of this page: https://dm-echo.andros.dev/ It will soon be running. It won't be long now.

I’m also thinking that some kind of tag might be needed to automatically hide twts from unknown extensions. For example our client doesn’t support DMs and always shows the !<nick url><encrypted_message> syntax which is meaningless.

⤋ Read More
In-reply-to » @eapl.me You asked me for private keys for testing purposes. I have added it to the bottom of this page: https://dm-echo.andros.dev/ It will soon be running. It won't be long now.

well, I suggested that in https://eapl.me/timeline/conv/k2ob6bq

The idea was to help those following the spec in https://twtxt.dev/exts/directmessage.Html, to replicate the steps and validate whether your implementation gives the same result.

BTW, you could add a link to the spec in the echo web.

⤋ Read More