I don't really use IRC anymore but at the time I remember really hating loggers. Chats were informal and I really didn't like them showing up in google results and elsewhere. Not everything needs to be saved.
I run a mastodon server on the fediverse (distributed social networking over activitypub). Mastodon and Pleroma support deletes, but say you post something and someone boosts it and one of their followers sees it on a server your server blocks.
You won't see their replies obviously (split thread) but if you delete the message, it gets deleted from your followers, but won't continue on to the server that your server blocks. So a copy will continue to exist.
I just treat the fediverse as a big chatroom/reddit thread that can never be deleted.
Yes, I agree. Lots(most?) of channels on freenode are explicitly not publicly logged anywhere, and the freenode staff are pretty helpful at helping chanops enforce this. None of the channels I help run have any public logs available.
Perhaps a naive question, but how do chanops enforce this? What's to stop someone lurking in a channel and publishing a log. You can't obviously connect an individual lurker to a log and busy channels have dozens of 24/7 clients connected.
> You can't obviously connect an individual lurker to a log
I don't think that's necessarily true.
While I don't know if the IRC Server software supports it, there's no reason you couldn't encode some kind of unique tag into the stream of IRC messages sent to each client. Unless you were running two clients and diffing their logs you wouldn't notice.
It could be as simple as seeing additional events, or modified whois/client information on join/leave, or even completely fake client join/leaves.
If you control the IRC server, there's nothing stopping you from modifying the events sent to every client in a way that lets you identify that client.
The difference here is that Freenode is not used so much for informal chat as with other IRC servers. Many free software projects use Freenode as the primary communication method for project coordination, and in this case it is essential to keep logs. Not everyone is online 100% of the time and logs enable everyone to catch up if they've been offline.
Sure, but those projects can also trivially set up logging with their own bots and in their own clients, with better control of what is logged, when, why, and how those logs are shared. This service appears to have logged fewer than 100 channels. This wasn't an aggregate logger and anyway aggregate logging probably isn't permissible anyway due to GDPR.
I would say chat history and asymmetrical scheduling for chat are expectations in this day and age, and the lack of them is one of the biggest reasons people find IRC irritating or difficult. (I ask a question, and if the person fit to answer it isn't logged in right now, it never gets even responded to.)
Every fairly modern chat solution, Skype, Discord, Slack, etc. allows you to see messages while you were offline. Compared to things from the older eras of messaging like AIM and YIM, when generally you couldn't even message someone unless they were online as well.
So it's not surprising to me for IRC loggers to be a relatively more modern element: They're filling in a gap IRC has with modern chat clients.
ZNC and earlier implementations have been around for 15+ years to provide queuing proxies. Pretty much all of the "unique" features found in current chat systems were also implemented in one fashion or another in the IRC ecosystem decades ago. Ease of use has definitely improved, in exchange for for-profit company control over communications in these systems. IRC's decentralized nature still argues for its relevance even now.
IRC is not really decentralised, is it? Or are you arguing that it’s decentralised in the sense that each user can implement and control some features like logging?
You probably did without realizing it. A persistently logged in user (through e.g. a bouncer, or just someone not rebooting too often) is essentially a logger. Those logs are private by their nature, but where I IRCed it was common for channels to each have their own logging bot.
I kinda doubt that Facebook actually deletes, as opposed to setting a DB flag... weren't they caught logging comments that were typed in but then modified / deleted / not actually posted?
The other party, i.e. the recipient still has an copy, so it could be a DB flag indeed. Although when both persons delete the message it might still lay around for who knows how long.
If you want a quick, casual conversation, I don't see why pseudonyms wouldn't be sufficient. You can also use random names if you're asking something really sensitive, which you would want to do in any case because it's a public space. For more in depth you can use your name if you want or a fixed pseudonym so you can recognize and be recognized.
You won't see their replies obviously (split thread) but if you delete the message, it gets deleted from your followers, but won't continue on to the server that your server blocks. So a copy will continue to exist.
I just treat the fediverse as a big chatroom/reddit thread that can never be deleted.