For example Discord has a splendid unofficial client(Ripcord), but because they don't care about their users they choose to ban people who use it and force them to use the original lag-o-matic mammoth of a client(Like, seriously, an IM that takes up a gig of ram on its own).
Which is intentionally garbage to use. I'm with the GP, pop it into an unprivileged LXD container and stop dealing with their corp snooping nonsense.
This brings me to a different pet peeve of mine. Discord is really opaque in a way that it handles activity status, it doesn't show you what status it thinks you are (i.e it can show you as Online but Away or Mobile to others). Which is very annoying when working with others.
For times when I work off my iPad, I use a vnc and xdotool to press a button every few seconds on my PC (not in Discord though, don't want to get banned). That is the only way I found to make sure it shows me as Online and not breaks their T&C.
Like I get that this is HN and the hacker ethos is “anything I can do as a human I should be able to automate” but that makes the already intractable spam problem that much harder.
Preventing spam is more a task for server-side detection and filtering. Gmail's been doing it very well for decades, and i can't see why similar methods won't work for messaging. And adding a rate-limit server-side is another defence. The least useful option of banning modded clients for chats is a bad trade-off for most users whilst providing very little actual protection.
Gmail actually does rely on various forms of "anti cheat" to control spam, including obfuscated code that detects attempts to automate the browser. That doesn't help for inbound mail but it helps a lot to stop people sending spam from Gmail.
Not by an order of magnitude, esp, for chat applications that are limited by http application protocols.
> ..that detects attempts to automate the browser.
Nobody sends spam from the gmail web browser client, but suppose google successfully bans (somehow?) all other email clients, the enterprising spam sending people would still easily automate the browser, esp. with things like autohotkeys (which doesn't require modifying the browser itself, nor run code in the browser).
The point is, banning a client side mod for any service is irrational if the justification is to prevent spammers.
"the enterprising spam sending people would still easily automate the browser, esp. with things like autohotkeys"
It's not at all easy for them, and people who try approaches as simple as that don't make any money. Or didn't when I worked on it at least.
"The point is, banning a client side mod for any service is irrational if the justification is to prevent spammers."
That belief is widespread, but I built a part of my career on proving it wrong. We got unexpectedly excellent results from cracking down on alternative clients that were pretending to be browsers.
Users still get banned for that, even when they are well within reasonable use, and then they are unbanned with words “This serves you right, now use our spyware* of a client”.
* Discord used to scan your entire HDD “for games” when starting up.