Likely slop

Likely slop

No.
You likely came to this conclusion because client tags and capabilities are all over the place in the code.
But appreciated.

Pretty sure I saw someone in IRC who was linking to boxlabs say they were working on a client they were filling with slop. I'm tired.

no need to get personal swears

I double checked my logs. There was a person who posted a couple of screenshots hosted on boxlabs and literally posted in IRC "I have a fork of Hexchat I've been filling full of AI slop."

But on closer inspection it looks like this wasn't you. I apologise for the error.

However, the initial response from 3 other people to this PR was that this looked like an LLM project. The website almost certainly is.

AI assisted projects or websites aren't banned outright, but we do try to keep low effort submissions out, especially if the submitter has no history of contribution or even discussion in the IRCv3 working group. Reputation makes a difference.

We won't download or install submissions usually, and we can't verify support levels claimed, but if there are strong signals that a submission has involved less effort than it takes to do even a basic smell test, then we'll likely close the PR.

I'll reopen this submission because I mistook you for someone else, and I will ask others to make a decision on merging, but please understand this from a maintainer's point of view when we receive multiple submissions that look suspicious.

However, the initial response from 3 other people to this PR was that this looked like an LLM project. The website almost certainly is.

You just merged ZoiteChat and their website looks terrible and very much like an AI threw it together. No offense to ZoiteChat, but let's be frank, if you think Mango's website looks like AI slop, and you think Lurker.Chat looks like AI slop, but you didn't think ZoiteChat looks like AI slop, then we really need some clarification on how you and others are classifying websites so that we can adhere better to the guidelines you've decided on.

The process should be clear, not up to the whims of single community member. At this point your whims have now left it up to a vague "other contributors" with no clear idea of who that is or why your decision was good enough for ZoiteChat but not the other apps.

AI assisted projects or websites aren't banned outright, but we do try to keep low effort submissions out, especially if the submitter has no history of contribution or even discussion in the IRCv3 working group. Reputation makes a difference.

I reject the idea that someone needs to be involved in the IRCv3 spec to make an IRC client that is recognized by the community. I think it's actually better if someone makes a client that isn't invested in the IRCv3 spec itself, so that they can implement features objectively and not for the sake of supporting the spec even if it does not make sense for the client or for the users.

Setting one of the parameters for inclusion as "must be friends with people in the #ircv3 channel" is textbook gatekeeping and favoritism and I hope it is not a literal part of the guidelines for inclusion. Reputation is a codeword for favoritism and gatekeeping.

We won't download or install submissions usually, and we can't verify support levels claimed, but if there are strong signals that a submission has involved less effort than it takes to do even a basic smell test, then we'll likely close the PR.

You do understand that, even with AI, making an entire IRC client is a lot of work? I've spent hundreds of hours on my IRC client. Just because you think the PR took less time than someone doing it manually does not "pass the smell test", which is an arbitrary standard in which you admit you will do nothing to actually vet the project being suggested for incision, doesn't make it okay.

People are enthusiastic and are trying to contribute to keeping IRC vibrant and alive, and your response so far has been textbook gatekeeping and dismissal. That's disappointing and you should be disappointed in yourself for reacting this way.

I hope someone in the community steps up and sets up clear guidelines for inclusion instead of relying on someone saying "slop" to any PR that crosses their desk on a late Sunday evening.

You just merged ZoiteChat and their website looks terrible and very much like an AI threw it together

That's true, but ZoiteChat publishes their code and development history, so we judged the quality and effort based on that instead of their website.

That's true, but ZoiteChat publishes their code and development history, so we judged the quality and effort based on that instead of their website.

Ah! Thank you for clarification.

Lurker does as well, it's fully open source and I use public GitHub for tracking the project. Starting May 6th up to now I have a devlog over here: https://discuss.bradroot.me/t/lurker-devlog/89

I can also point to the (kind of insane) amount of issues I have filed and closed while I have been developing lurker. I've been developing fully out in the open, which is why it was a shock to me that my PR for lurker was closed with no discussion originally. https://github.com/amiantos/lurker/issues?q=is%3Aissue%20state%3Aclosed

I'll link to this over on the lurker issue so it can contribute to discussion over there. Thank you again.

I can also point to the (kind of insane) amount of issues I have filed and closed while I have been developing lurker. I've been developing fully out in the open, which is why it was a shock to me that my PR for lurker was closed with no discussion originally. https://github.com/amiantos/lurker/issues?q=is%3Aissue%20state%3Aclosed

That Claude filed and closed *

Quite ridiculous.
I'm no longer attempting for hexdroid to be included, and I'll probably spin off ircv4 in the future. There's issues with the running of the ircv3 group, whether that's around the profitability of irccloud as more clients continue to include the specs, who knows.