@bender@twtxt.net You said:

as long as those working on clients can reach an agreement on how to move forward. That has proven, though, to be a pickle in the past.

I think this is because we probably need to start thinking about three different aspects to the ecosystem and document them out:

• Specifications (as they are now)
  
• Server recommendations (e.g: Timeline, yarnd, etc)
  
• Client recommendations (e.g: jenny, tt, tt2, twet, etc)

@andros@twtxt.andros.dev nothing stands still, I agree. I think current twtxt has surpassed the initial specification, while still being relatively backwards compliant/compatible but, for how long?

As for new extensions (DM, for example), they should be OK as long as those working on clients can reach an agreement on how to move forward. That has proven, though, to be a pickle in the past.

The nice thing here is that any Ui/UX rendering for a “good user experience” is similar to what yarnd does for Youtube/Spotify/whatever embedding. Plus anyone can participate, even if they don’t really have a client that understand it, it’s just text with some “syntax” afterall.

💡 I had this crazy idea (or is it?) last night while thinking about Twtxt and Yarn.social 😅 There are two things I think that could be really useful additions to the yarnd UI/UX experience (for those that use it) and as “client” features (not spec changes). The two ideas are quite simple:

• Voting – a way to cast, collect a vote on a decision, topic or opinion.
  
• RSVP – a way to “rsvp” to a virtual (pr physical) event.
  

Both would use “plain text” on top of the way we already use Twtxt today and clients would render an appropriate UI/UX.

@kat@yarn.girlonthemoon.xyz At the core, you need an ngircd.conf like this:

[Global]
    Name = your.irc.server.com
    Password = yourfancypassword
    Listen = 0.0.0.0
    Ports = 6667

    AdminInfo1 = Well, me.
    AdminInfo2 = Over here!
    AdminEMail = forget.it@example.invalid

[Options]
    Ident = no
    PAM = no

[SSL]
    CertFile = /etc/ssl/acme/your.irc.server.com.fullchain.pem
    KeyFile = /etc/ssl/acme/private/your.irc.server.com.key
    DHFile = /etc/ngircd/dhparam.pem
    Ports = 6669

Start it and then you can connect on port 6667. (The SSL cert/key must be managed by an external tool, probably something like certbot or acme-client.)

I’m assuming OpenBSD here. Haven’t tried it on Linux lately, let alone Docker. 😅

@prologic@twtxt.net I also thought it was a client-server thingy at first and usually it is, I guess, there’s just this workaround:

If it is not possible to install Borg on the remote host, it is still possible to use the remote host to store a repository by mounting the remote filesystem, for example, using sshfs.

@prologic@twtxt.net no, it is not a “server-client thingy”.

Seem like it’s a server-client thingy? 🤔 I much prefer tools in this case and defer the responsibility of storage to something else. I really like restic for that reason and the fact that it’s pretty rock solid. I have zero complaints 😅

@javivf@adn.org.es having the extension listed means that it has been discussed and, usually implemented. Now, number 6 and 7 on the list as its stands today are not supported by any of the known clients. I believe their (again, 6 and 7) inclusion on the list has been precipitated, and lax.

@quark@ferengi.one No editing old Twts that are the root of a thread with replies in the ecosystem. Just results in a fork. Unless the client has an implementation that does not store Twts keyed by Hash.

@movq@www.uninformativ.de wouldn’t editing your own twtxts cause the same issue Yarnd (or any other client) has, which is breaking any replies to it? Under which conditions would this work the best? When copying the twtxt.txt file asynchronously? In my case I copy the twtxt.txt file to its web root right away, but I figure I could not do that, which would give me a set period of time to edit without worries.

@doesnm.p.psf.lt@doesnm.p.psf.lt It was always intended to have both Yarn.social and Salty.im integrate together. Yes. This includes having a set of specifications that anyone can write clients to.

I think I would encourage anyone in this community is to care less about supporting “legacy clients” and focus more on value-add whilst balancing the burden of client authors – which have very precious little “spare time” 🤣

@andros@twtxt.andros.dev I don’t see any “fighting” here. This is just good experimentation. Unfortunately there hasn’t really been enough time or effort by other “client authors” yet, me especially as I’ve been super busy with ya’ know my “day job” that pays the bills and refactoring yarnd to use a new and shiny and much better SqliteCache 🤣 – I certainly don’t think your efforts are wasted at all. I would however like @doesnm.p.psf.lt@doesnm.p.psf.lt encourage you to look at the work we’ve done as a community (which was also driven out of the Yarn.social / Twtxt community years back).

@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. 😂

Aside from fetching feeds every three minutes (which kind of adds mystery to this puzzle), I think there is something else going on with the client you are using, @andros@twtxt.andros.dev. Some of those twtxts are seconds apart, making me truly stumped. 😅

Hmm so looking at the swagger of the registry spec client it seems to just take a “page”.. That seems worse than doing an offset. Lol.

https://github.com/DracoBlue/twtxt-registry/blob/master/src/swagger.json

@lyse@lyse.isobeef.org I’m glad to hear that! Yay for more clients. 😊

@andros@twtxt.andros.dev Can you reproduce any of this outside of your client? I can’t spot a mistake here:

$ curl -sI 'http://movq.de/v/8684c7d264/.html%2Dindex%2Dthumb%2Dgimp11%2D1.png.jpg'
HTTP/1.1 200 OK
Connection: keep-alive
Content-Length: 2615
Content-Type: image/jpeg
Date: Wed, 19 Mar 2025 19:53:17 GMT
Last-Modified: Wed, 19 Mar 2025 17:34:08 GMT
Server: OpenBSD httpd

$ curl -sI 'https://movq.de/v/8684c7d264/gimp11%2D1.png'
HTTP/1.1 200 OK
Connection: keep-alive
Content-Length: 131798
Content-Type: image/png
Date: Wed, 19 Mar 2025 19:53:19 GMT
Last-Modified: Wed, 19 Mar 2025 17:18:07 GMT
Server: OpenBSD httpd

$ telnet movq.de 80
Trying 185.162.249.140...
Connected to movq.de.
Escape character is '^]'.
HEAD /v/8684c7d264/.html%2Dindex%2Dthumb%2Dgimp11%2D1.png.jpg HTTP/1.1
Host: movq.de
Connection: close

HTTP/1.1 200 OK
Connection: close
Content-Length: 2615
Content-Type: image/jpeg
Date: Wed, 19 Mar 2025 19:53:31 GMT
Last-Modified: Wed, 19 Mar 2025 17:34:08 GMT
Server: OpenBSD httpd

Connection closed by foreign host.
$ 

@andros@twtxt.andros.dev Hm, looks correct to me. The image to be displayed is a thumbnail and this links to the full-sized image. The thumbnail (JPG) is auto-generated from the full image (PNG), hence the two extensions.

What does look strange, though, is that your client came up with the hash pqsmcka, while it should have been te5quba. 🤔

@prologic@twtxt.net Agreed! But clients can hallucinate and generate wrong hashes aka Lies 🤣 Also, If you chheck your own twt on twtxt.net, it looks like a root twt instead of a replay.

Some satisfying icicle-breaking in our backyard: photos.falsifian.org/video/sM7G3vfS6yuc/VID_20250217_203250.mp4

I couldn’t resist taking home a prize:

Image

Image

It’s been snowy here in #Toronto.

Image

(I tried formatting the images in markdown for the benefit of yarn and any other clients that understand it.)

#photo #photos #winter

@arne@uplegger.eu nice work with the client.
I also see you are using the Yellow CMS for your website🍋

@prologic@twtxt.net I say we should find a way to support mentions with only url, no nick, as per the original spec.

• For @<nick url> we already got support
  
• For @<nick> the posting client should expand it to @<nick url>, if not then the reading client should just render it as @nick with no link.
  
• For @<url> the sending client should try to expand it to @<nick url>, if not then the reading client should try to find or construct a nick base on:
  
  1. Look in twtxt.txt for a nick =
     
  2. Use (sub)domain from URL
     
  3. Use folder or file name from URL

What should the advantage be to nick = _compared to just not defining a nick and let the client use the domain as the handle?

What is not intuitive is that you put something in the nick field that is not to be taken literary. The special meaning of _ is only clean if you read the documentation, compared to having something in nick that makes sense in the current context of the twtxt.txt.

@doesnm@doesnm.p.psf.lt So the user should then set nick = _@domain.tld in the twtxt.txt?

It seems more intuitive and userfriendly to just use: nick = domain.tld and have then convention for clients to render the handle as @domain.tld instead of @domain.tld@domain.tld

For a feed with no nick defined (eg. https://akkartik.name/twtxt.txt) it will also be simpler and make more sense to just use the domain as the nick and render it as @domain.tld

@prologic@twtxt.net What IRC client is that?

Yes it work: 2024-12-01T19:38:35Z twtxt/1.2.3 (+https://eapl.mx/twtxt.txt; @eapl) :D

The .log is just a simple append each request. The idea with the .cvs is to have it tally up how many request there have been from each client as a way to avoid having the log file grow too big. And that you can open the .cvs as a spreadsheet and have an easy overview and filtering options.

Access to those files are closed to the public.

@sorenpeter@darch.dk on 4 for gemini if your TLS client certificate contains your nick@host could that work for discovery?

@eapl.me@eapl.me here are my replies (somewhat similar to Lyse’s and James’)

1. Metadata in twts: Key=value is too complicated for non-hackers and hard to write by hand. So if there is a need then we should just use #NSFS or the alt-text file in markdown image syntax ![NSFW](url.to/image.jpg) if something is NSFW

2. IDs besides datetime. When you edit a twt then you should preserve the datetime if location-based addressing should have any advantages over content-based addressing. If you change the timestamp the its a new post. Just like any other blog cms.

3. Caching, Yes all good ideas, but that is more a task for the clients not the serving of the twtxt.txt files.

4. Discovery: User-agent for discovery can become better. I’m working on a wrapper script in PHP, so you don’t need to go to Apaches log-files to see who fetches your feed. But for other Gemini and gopher you need to relay on something else. That could be using my webmentions for twtxt suggestion, or simply defining an email metadata field for letting a person know you follow their feed. Interesting read about why WebMetions might be a bad idea. Twtxt being much simple that a full featured IndieWeb sites, then a lot of the concerns does not apply here. But that’s the issue with any open inbox. This is hard to solve without some form of (centralized or community) spam moderation.

5. Support more protocols besides http/s. Yes why not, if we can make clients that merge or diffident between the same feed server by multiples URLs

6. Languages: If the need is big then make a separate feed. I don’t mind seeing stuff in other langues as it is low. You got translating tool if you need to know whats going on. And again when there is a need for easier switching between posting to several feeds, then it’s about building clients with a UI that makes it easy. No something that should takes up space in the format/protocol.

7. Emojis: I’m not sure what this is about. Do you want to use emojis as avatar in CLI clients or it just about rendering emojis?

@bender@twtxt.net @prologic@twtxt.net I’m not exactly asking yarnd to change. If you are okay with the way it displayed my twts, then by all means, leave it as is. I hope you won’t mind if I continue to write things like 1/4 to mean “first out of four”.

What has text/markdown got to do with this? I don’t think Markdown says anything about replacing 1/4 with ¼, or other similar transformations. It’s not needed, because ¼ is already a unicode character that can simply be directly inserted into the text file.

What’s wrong with my original suggestion of doing the transformation before the text hits the twtxt.txt file? @prologic@twtxt.net, I think it would achieve what you are trying to achieve with this content-type thing: if someone writes 1/4 on a yarnd instance or any other client that wants to do this, it would get transformed, and other clients simply wouldn’t do the transformation. Every client that supports displaying unicode characters, including Jenny, would then display ¼ as ¼.

Alternatively, if you prefer yarnd to pretty-print all twts nicely, even ones from simpler clients, that’s fine too and you don’t need to change anything. My 1/4 -> ¼ thing is nothing more than a minor irritation which probably isn’t worth overthinking.

@sorenpeter@darch.dk I run Weechat headless on a VM and mostly connect via mobile or dwsktop. I use the android client or gliwing bear. Work blocks all comms on their always on MitM VPN so I cant in office anymore. So I just use mobile.

@Codebuzz@www.codebuzz.nl Speed is an issue for the client software, not the format itself, but yes I agree that it makes the most sense to append post to the end of the file. I’m referring to the definition that it’s the first url = in the file that is the one that has to be used for the twthash computation, which is a too arbitrary way of defining something that breaks treading time and time again. And this is the case for not using url+date+message = twthash.

I know no client support it (yet) - but it could be the future 😅

@2024-10-08T19:36:38-07:00@a.9srv.net Thanks for the followup. I agrees with most of it - especially:

Please nobody suggest sticking the content type in more metadata. 🙄

Yes, URL can be considered ugly, but they work and are understandable by both humans and machines. And its trivial for any client to hide the URLs used as reference in replies/treading.

Webfinger can be an add-on to help lookup people, and it can be made independent of the nick by just serving the same json regardless of the nick as people do with static sites and a as I implemented it on darch.dk (wf endpoint). Try RANDOMSTRING@darch.dk on http://darch.dk/wf-lookup.php (wf lookup) or RANDOMSTRING@garrido.io on https://webfinger.net

@doesnm@doesnm.p.psf.lt Agree. salty.im should allow the user to post multiple brokers on their webfinger so the client can find a working path.

@movq@www.uninformativ.de i’m sorry if I sound too contrarian. I’m not a fan of using an obscure hash as well. The problem is that of future and backward compatibility. If we change to sha256 or another we don’t just need to support sha256. But need to now support both sha256 AND blake2b. Or we devide the community. Users of some clients will still use the old algorithm and get left behind.

Really we should all think hard about how changes will break things and if those breakages are acceptable.

These collisions aren’t important unless someone tries to fork. So.. for the vast majority its not a big deal. Using the grow hash algorithm could inform the client to add another char when they fork.

@prologic@twtxt.net YES James, it should be up to the client to deal with changes like edits and deletions. And putting this load on the clients, location-addressing with make this a lot easier since what is says it: Look in this file at this timestamp, did anything change or went missing? (And then threading will not break;)

Diving into mblaze, I think I’ve nearly* reached peek email geek.

Just a bunch of shell commands I can pipe together to search, list, view and reply to email (after syncing it to a local Maildir).

EXAMPLES at https://git.vuxu.org/mblaze/tree/README

So far I’m using most of the tools directly from the command line, but I might take inspiration from https://sr.ht/~rakoo/omail/ to make my workflow a bit more efficient.

*To get any closer, I think I’d have to hand-craft my own SMTP client or something.

@prologic@twtxt.net

See https://dev.twtxt.net

Yes, that is exactly what I meant. I like that collection and “twtxt v2” feels like a departure.

Maybe there’s an advantage to grouping it into one spec, but IMO that shouldn’t be done at the same time as introducing new untested ideas.

See https://yarn.social (especially this section: https://yarn.social/#self-host) – It really doesn’t get much simpler than this 🤣

Again, I like this existing simplicity. (I would even argue you don’t need the metadata.)

That page says “For the best experience your client should also support some of the Twtxt Extensions…” but it is clear you don’t need to. I would like it to stay that way, and publishing a big long spec and calling it “twtxt v2” feels like a departure from that. (I think the content of the document is valuable; I’m just carping about how it’s being presented.)

why can we both have a format that you can write by hand and better clients?

Some more arguments for a local-based treading model over a content-based one:

1. The format: (#<DATE URL>) or (@<DATE URL>) both makes sense: # as prefix is for a hashtag like we allredy got with the (#twthash) and @ as prefix denotes that this is mention of a specific post in a feed, and not just the feed in general. Using either can make implementation easier, since most clients already got this kind of filtering.

2. Having something like (#<DATE URL>) will also make mentions via webmetions for twtxt easier to implement, since there is no need for looking up the #twthash. This will also make it possible to make 3th part twt-mentions services.

3. Supporting twt/webmentions will also increase discoverability as a way to know about both replies and feed mentions from feeds that you don’t follow.

#fzf is the new emacs: a tool with a simple purpose that has evolved to include an #email client. https://sr.ht/~rakoo/omail/

I’m being a little silly, of course. fzf doesn’t actually check your email, but it appears to be basically the whole user interface for that mail program, with #mblaze wrangling the emails.

I’ve been thinking about how I handle my email, and am tempted to make something similar. (When I originally saw this linked the author was presenting it as an example tweaked to their own needs, encouraging people to make their own.)

This approach could surely also be combined with #jenny, taking the place of (neo)mutt. For example mblaze’s mthread tool presents a threaded discussion with indentation.

@prologic@twtxt.net Thanks for writing that up!

I hope it can remain a living document (or sequence of draft revisions) for a good long time while we figure out how this stuff works in practice.

I am not sure how I feel about all this being done at once, vs. letting conventions arise.

For example, even today I could reply to twt abc1234 with “(#abc1234) Edit: …” and I think all you humans would understand it as an edit to (#abc1234). Maybe eventually it would become a common enough convention that clients would start to support it explicitly.

Similarly we could just start using 11-digit hashes. We should iron out whether it’s sha256 or whatever but there’s no need get all the other stuff right at the same time.

I have similar thoughts about how some users could try out location-based replies in a backward-compatible way (append the replyto: stuff after the legacy (#hash) style).

However I recognize that I’m not the one implementing this stuff, and it’s less work to just have everything determined up front.

Misc comments (I haven’t read the whole thing):

• Did you mean to make hashes hexadecimal? You lose 11 bits that way compared to base32. I’d suggest gaining 11 bits with base64 instead.

• “Clients MUST preserve the original hash” — do you mean they MUST preserve the original twt?

• Thanks for phrasing the bit about deletions so neutrally.

• I don’t like the MUST in “Clients MUST follow the chain of reply-to references…”. If someone writes a client as a 40-line shell script that requires the user to piece together the threading themselves, IMO we shouldn’t declare the client non-conforming just because they didn’t get to all the bells and whistles.

• Similarly I don’t like the MUST for user agents. For one thing, you might want to fetch a feed without revealing your identty. Also, it raises the bar for a minimal implementation (I’m again thinking again of the 40-line shell script).

• For “who follows” lists: why must the long, random tokens be only valid for a limited time? Do you have a scenario in mind where they could leak?

• Why can’t feeds be served over HTTP/1.0? Again, thinking about simple software. I recently tried implementing HTTP/1.1 and it wasn’t too bad, but 1.0 would have been slightly simpler.

• Why get into the nitty-gritty about caching headers? This seems like generic advice for HTTP servers and clients.

• I’m a little sad about other protocols being not recommended.

• I don’t know how I feel about including markdown. I don’t mind too much that yarn users emit twts full of markdown, but I’m more of a plain text kind of person. Also it adds to the length. I wonder if putting a separate document would make more sense; that would also help with the length.

@prologic@twtxt.net Do you have a link to some past discussion?

Would the GDPR would apply to a one-person client like jenny? I seriously hope not. If someone asks me to delete an email they sent me, I don’t think I have to honour that request, no matter how European they are.

I am really bothered by the idea that someone could force me to delete my private, personal record of my interactions with them. Would I have to delete my journal entries about them too if they asked?

Maybe a public-facing client like yarnd needs to consider this, but that also bothers me. I was actually thinking about making an Internet Archive style twtxt archiver, letting you explore past twts, including long-dead feeds, see edit histories, deleted twts, etc.

I wrote some code to try out non-hash reply subjects formatted as (replyto ), while keeping the ability to use the existing hash style.

I don’t think we need to decide all at once. If clients add support for a new method then people can use it if they like. The downside of course is that this costs developer time, so I decided to invest a few hours of my own time into a proof of concept.

With apologies to @movq@www.uninformativ.de for corrupting jenny’s beautiful code. I don’t write this expecting you to incorporate the patch, because it does complicate things and might not be a direction you want to go in. But if you like any part of this approach feel free to use bits of it; I release the patch under jenny’s current LICENCE.

Supporting both kinds of reply in jenny was complicated because each email can only have one Message-Id, and because it’s possible the target twt will not be seen until after the twt referencing it. The following patch uses an sqlite database to keep track of known (url, timestamp) pairs, as well as a separate table of (url, timestamp) pairs that haven’t been seen yet but are wanted. When one of those “wanted” twts is finally seen, the mail file gets rewritten to include the appropriate In-Reply-To header.

Patch based on jenny commit 73a5ea81.

https://www.falsifian.org/a/oDtr/patch0.txt

Not implemented:

• Composing twts using the (replyto …) format.
  
• Probably other important things I’m forgetting.

the stem matching is the same as how GIT does its branch hashes. i think you can stem it down to 2 or 3 sha bytes.

if a client sees someone in a yarn using a byte longer hash it can lengthen to match since it can assume that maybe the other client has a collision that it doesnt know about.

@sorenpeter@darch.dk hmm, how does your client handles “a little editing”? I am sure threads would break just as well. 😉

@prologic@twtxt.net I wouldn’t want my client to honour delete requests. I like my computer’s memory to be better than mine, not worse, so it would bug me if I remember seeing something and my computer can’t find it.