@eaplmx@twtxt.net Nop but a lot at least.

@prologic@twtxt.net yep. In that experiment I was only interested by real people, who talk/contacts with others I can follow.

one other thing cool that you have not noticed but i want to share. Is that the first iterative version was taking 8 minutes to go on all twtxt / download/ parse. The new version using heavily go routines takes around 10 seconds.

Thanks @lyse@lyse.isobeef.org,

1. even a switch case would be better
2. done quickly
3. done quickly
4. yes BUT if I check only mentions I will loose the header following one, (that are not always in the same style: cf : https://niplav.github.io/twtxt.txt ), I can do both and only loose the redirection BUT I need to be bored again to do that change (1 hour I ll say)
5. (easy) when i will be bored again it will be on the pipe.

Thanks

Actually I should black list search as well.

@mckinley@twtxt.net my bad, I ’ve blacklisted it in the code for next time (github-commit-fix)

@prologic@twtxt.net you didn’t capture the best of the form :

@abucci@anthony.buc.ci Very nice thinking :D

@prologic@twtxt.net what is spam in twtxt ? I mean I just saw 2 hotel advertise so far.

@lyse@lyse.isobeef.org @mckinley@twtxt.net @eaplmx@twtxt.net the code you requested to follow all the twtxt mentions (https://github.com/tkanos/we-are-twtxt on the folder search)

• I adjusted the code to take for the rank, taking in account only the mentions and not weighted by the number of tweets)
• the files generated will be in your home folder

@mckinley@twtxt.net here is the dump https://github.com/tkanos/we-are-twtxt (on the tarball all-twtxt.tar.xz)

@mckinley@twtxt.net @lyse@lyse.isobeef.org I clean m’y code today and release it tomorrow or tonight.

Where are you ?

A good lightweight and written in go alternative : gotosocial (still in alpha)

One of my friends did a very good joke, I was talking with him about privacy on internet, and he replied to me :

• What are you talking about, you use firefox on Linux in a village of maximum 1000 people, you don’t even need to use cookies.

:D

@brasshopper@twtxt.net Oh please implement everything you want. It is always enjoyable to implement things.

Yarn already have most of it :

• Latest tweet : Discovery tab has it (https://twtxt.net/discover)
• Search for tweets : that one is missing
• Mentions : Mentions tab has it (https://twtxt.net/mentions)
• Tag : the search tag bar is there (https://twtxt.net/search?tag=mytag)
• User : you can see a profile (https://twtxt.net/user/brasshopper/)

@lyse@lyse.isobeef.org yep it’s the main component (how many time people talk about you). I can share my code, but not now, I will be really ashamed to share such horrible code. Let me 1-2 weeks to refactor, and I will.

Start account was a random one (https://niplav.github.io/twtxt.txt) then interconnections are found by following and mentions.

@darch@twtxt.net@lyse@lyse.isobeef.org @mckinley@twtxt.net Based on an idea by @eaplmx@twtxt.net I used a simple quickly and dirty made walker algorithm (bypassing the spider trap) to find the interconnected account (1289) I only found the interconnected ones (since 2016) (if you have accounts that you know and are not referenced I can send the walker algorithm across them) :

• Some don’t answer anymore, the one that does answer => accessible (721)
• Some still post in 2022 => active (111)

Then I computed a normalized ranking algorithm in all of those with the number of time a people talk about you for the influence (it sounds like I’m talking like a star trek characters).

Guys, I have a bad news, I went through the twtxt-osphere :

• I found 1289 twtxt account
• among those 721 are accessible ( 712 http / 9 gemini / 0 gopher)
• but only 111 account are still active in 2022 :S (107 http / 4 gemini / 0 gopher).

I have also rank all the accounts by influence. (just for fun) without surprise @prologic@twtxt.net is the 1st most influential twtxter, @lyse@lyse.isobeef.org you are 3rd, (I’m 32nd)

results here

@prologic@twtxt.net have you ever tried expect

@prologic@twtxt.net. The issue is that (I don’t know why) programmers (often backends) are full of protocol, and they forget the basis of clean code, to pursue their protocol.

Thanks @lyse@lyse.isobeef.org

@brasshopper@twtxt.net I HATE Standard Go Project Layout, I have always prefered to structure my program by bounded context.

@prologic@twtxt.net for yarn integration, I have the idea to create a Meta tag for PostHook (like : {{yarn}}) that will do behind the scene like yarnc login + yarnc tweet. so the twtxt local will only be a backup. On the second step I will imlement the same PsotHook solution for follow and unfollow.

@prologic@twtxt.net, twt Hash is already implemented the same way as yarnd. For the subject ?? I don;t know about that one. But I will implement if you want. The reply is already done twx tweet -r hash <text>, the fork not yet.

@prologic@twtxt.net, twt Hash is already implemented the same way as yarnd. For the subject ?? I don;t know about that one. But I will implement if you want. The reply is already done twx tweet -r hash <text>, the fork not yet.

@prologic@twtxt.net is on the pipe , I need to stabilize a beta and have a pretty timeline, then yarn integration.

@prologic@twtxt.net I will be interested to see that @lyse@lyse.isobeef.org code

@darch@neotxt.dk What is the orange code of that orange version ?

Thanks for that new idea you gave me @eaplmx@twtxt.net.

Yep you can’t, in yarn.social as it’s more centralized you can see it in Discover, but in local you only see what you follow. What can be done as well to extend it, will be to create a Discover like by downloading all the users followed by the one that you follow an so on. Here you can create a big map of who follows who, and have all their tweet as well.

Yep or even the thread you are just participating in.

They don’t announce, when you will look at your timeline, the client will connect to all your Followings and get their twtxt.txt, having all the twtxt.txt it will sort it by date and present it to you, then is up to you to read your timeline and understand who reply to who.

But you gave me a good idea to implement in mine, something to show the last replies to my tweets, or mentions of my name. Thanks

Please don’t test it for now BUT just to announce that the version 0.1.0 alpha (not yet tested) of my future twtxt client twx has been pushed.

It should have the basic for a client (follow/unfollow/tweet/timeline), the fancy part will soon come (to see on the readme file).

Using bitmaps to run interactive retention analyses over billions of events for less than $100/mo : https://vikramoberoi.com/using-bitmaps-to-run-interactive-retention-analyses-over-billions-of-events-for-less-than-100-mo/

You can also use yarnc (very good), if you want to interact with your twtxt.net account

@eaplmx@twtxt.net in order to install jdtron/twet :

GO111MODULE=off go get -u github.com/jdtron/twet 

After trying all of them I decided to write my own. I hope (If I’m not lazy) I will have finish for next week .

@prologic@twtxt.net, actually I was wrong, I had a misunderstanding of how the hash was supposed to work, I though that the hash was created at each tweet to be able to reply to it. But I noticed that actually a hash is created at the reply. (Am I right?) but it can only work if all the twtxt client share the same algorithm. Am I right ?

@darch Sorry, It was just a way to tell that I loved both as well. I didn’t want to give you more job :D

Hello @ghklhijsd@twtxt.net you have a very hacker name, like nobody will never remember you.

Yep I want to use the dark blue on the morning, and the dark yellow during the night (Don’t ask why)

Because I really enjoy working on Linux and the docker environment integration is better. BUT Apple has a better integration with others apps and with my phone.

If I have the choice between a Linux and an Apple, I really don’t know which one I would choose.

If you really think to do that, the more complete command line will be :

git --no-pager log --all --reverse --date=format:'%Y-%m-%dT%H:%M:%SZ' --format=format:'%cd (#%h) %s' > twtxt.txt  

Yep,as you can see :

prologic #oriex7q (57m ago)
(#b3xm5ua) @tkanos(tkanos) Yeah I was wondering that myself too.

as you see you post has 2 hash : the one in your message : #b3xm5ua, and the one generated on the fly #oriex7q.

So if I try to reply to #b3xm5ua, I will receive a message like :

Twt hash #b3xm5ua is not in your timeline. Proceed? (y/N):

Encouraging me to use the wrong one oriex7q.

Did a lot of manual tests and so far everything is Ok. But I see that Jdtron on timeline, thread and reply (instead of parsing the line) recreate the hash using the text that includes the hash generated on the writing of the tweet , so the hash generated is never the same.

Congrats, it’s a big achievement.