Replyallcalypse is always fun. I experienced one a few years ago when working at a consulting company with 40k IT consultants. There were thousands of reply-all until the mail servers went down. It baffled me how so many IT consultants could see their mailbox fill up with endless "stop sending me mail"-mails and think "I better tell everyone to stop mailing me".
The most destrutive form of fun i had - now repeated in multiple companies and organisations - was to simply place a Flummi (a small hard bouncy ball) in an populated space: It is guaranteed that you will relatively fast hear the typical "thud-thud-thud" of someone throwing it against a wall... the sound will get louder and will inevitable lead to a "thud... Thud... THUD... crash...".
My conclusion from this social experiment is:
1: Humans are unable to leave a ball lying around
2: Throwing a hard rubber ball in an enclosed space will lead to destruction
3: It is irrelevant how old, dignified or high up people are in the hierarchy of an organisation to withstand the in 1 mentioned effect.
Many years ago we had one too. It was very hot in the first day or two, but like with a forest fire, you had new outbreaks when people returned from vacation and before having read any other replies, replied to all "unsubscribe" or "why am I getting this?". Which prompted again replies, not to reply to all and so on...
Took about two weeks to get completely stopped. It seems, after that, mail to all in the company was effectively eliminated except for a whitelist of accounts.
Reminds me of the time our CTO sent 35000 emails to a single customer. The poor guy had his email ddos’d for days. Even though our email api quickly stopped sending emails after we ran out of credit, the guy was still getting the same email for days. It was as if the emails were taking weird paths through the internet and queuing in certain places before turning up
I host a service tool to contact our suppliers in case of a defect. Every 20 minutes the tool checks if there are new processes and if so suppliers are contacted per mail with a link to a form that contains questions and relevant files.
Newbie mistake of me was to send the mail and set a database entry that the supplier had been contacted. In exact that order. You can guess which part failed at some point and who got the same mail every 20 minutes...
They did respond with above average speed though. Didn't stop the mails of course.
If the receiving e-mail server doesn't acknowledge that it has received the e-mail, the sending server will try again for 5 days. Maybe it timeouted due to the huge e-mail spam.
This might explain why he got the same mail for days.
Not sure what you mean. A bounce occurs when a receiving MTA refuses to accept delivery. The sending MTA might then compose a non-delivery message to the sender; but that's not under the receiving MTA's control.
That is, if the sending MTA sends a "bounce message" direct to the sender's address, that's not specified by any email RFC, and is arguably bad behaviour - similar to running a "spam-bouncer" plugin on your email client, the bounces it issues are bogus.
> all emails with a reply to of a certain msg id.
I've no idea what that means. The reply-to is an email address, not a message-id.
I'm surprised it hasn't happened before. When I was in the air force, it was common practice to send out embarrassing emails if someone left their computer unlocked. They'd send them to the whole squadron or wing if they could.
It happens at my workplace. If someone who has access to elevated privileges leaves their workstation unlocked and unattended, it's highly likely that a colleague will send an email to the team to embarrass them.
I've pointed out that this technically breaches company policy regarding use of someone else's user account/identity, but it falls on deaf ears.
Leaving your computer unlocked usually also breaks policy (except if the policy is very much useless). The non-embarassing way would be to report the user leaving the computer unlocked, leading to disciplinary action and eventual dismissal. I think something slightly embarassing is the kinder option here.
Once I received spam in the form of a giant group text. People started responding with "STOP", attempting to opt out without understanding the situation. Trying to be helpful I sent a message like, "hey, please don't, you're sending this to dozens of people & it won't do anything."
Immediately someone from the list started trying to call me, and sending me threats me to report me for spamming. I think everyone else got the message though, and the replies stopped shortly after.
It is very interesting that "reply all" is a forbidden action with email. Other mediums have the possibility, sometimes by default, to let messages and replies be visible by everyone not just in the conversation but in the whole world and it's working fine.
Why is it forbidden with emails ? The functionality is useful. It is forbidden not because it is useless, but because the UX of emails make every single new message the most important thing to read and process. It is there in the Inbox, it is at the top, you can't ignore it.
Other UX make this more manageable. Usenet is used to frequent back-and-forths, so it arranges them inside their respective threads. Zulip lets you focus on the topics you're interested in and come back to a specific thread when you want.
I keep thinking we don't have a problem with email, we have a problem with UX.
I think the reason is also that people receiving emails have not necessarily opted in to receive it.
Someone in a Whatsapp group has chosen to join it. They can leave whenever they want and turn off notifications as a middle ground. Someone emailing the entire office because of a massage class is just lazy or an attention seeker (and the people replying to all "please remove me from this list" are total idiots)
I blacklist the "massage class tomorrow at lunch time" ladies in our office upon the first email but the problem is I still get all the replies to all. Email lacks advanced management options.
> I think the reason is also that people receiving emails have not necessarily opted in to receive it.
Having an email address amounts to opting in to receiving email.
I think the reason is people not knowing how to use email. Most people make much more use of instant messaging services than email; with IM, generally you either reply or you don't.
Reply-all storms have always been a problem; I think it was made dramatically worse by Microsoft Outlook, which idosyncratically placed the cursor above the (quoted) message being replied to. That encouraged people to reply without scrolling down (because they've already seen what they're replying to); so they don't see how many quotes of quotes of quotes they could safely trim before clicking send. They just type a one-liner, and happily send off 16K of stuff that everyone else has already seen too.
Trim and contextualize. That is, keep those parts of what you're replying to directly, and remove all the dross.
It's too late to say "Please don't top-quote" - unfortunately that ship sailed two decades ago.
I use reply productively all the time, in asynchronous technical discussions that involve two to four people.
Most stories where things go wrong involve catch-all addresses that forward to thousands or tens of thousands addresses, which either shouldn't exist in the first place or should be limited access.
> It appears that an individual did inadvertently send an email to an AF-All email address. The individual was looking for help with a computer issue and mistakenly added the wrong address.
I know why people opt for this sort of linguistic puffery. I know not, though, whether they realize that it has the opposite effect on anyone who has the competence and mastery that they're trying to emulate. It seems like a "no"; otherwise, they wouldn't do something that makes them look so stupid.
I don't think the average user really knows the difference. They just think that Reply-All is the only behaviour and even if presented with a choice they would likely still choose to get their message to more people so that they can be removed from the list (surely that will work)
This happened at Oracle once, it shut down a large number of engineering’s mailboxes for a day since they got full with “please remove me from this thread” reply-alls.
I’ve long thought an easy solution to this would be to block sending to a mailing list over a certain size (say 1k, but may depend on context) to only admins of that list.
Way back in uni there was a distribution address with all the emails in it. To protect it, students weren't supposed to know it so they put it in the BCC field.
Except of course once, they forgot. Much hilarity ensued. This was also back when you could still send an email using any from address you wanted, though I can't remember if anybody used that combo.
40 comments
[ 3.6 ms ] story [ 84.9 ms ] threadMy conclusion from this social experiment is:
1: Humans are unable to leave a ball lying around
2: Throwing a hard rubber ball in an enclosed space will lead to destruction
3: It is irrelevant how old, dignified or high up people are in the hierarchy of an organisation to withstand the in 1 mentioned effect.
Newbie mistake of me was to send the mail and set a database entry that the supplier had been contacted. In exact that order. You can guess which part failed at some point and who got the same mail every 20 minutes...
They did respond with above average speed though. Didn't stop the mails of course.
This might explain why he got the same mail for days.
Not sure what you mean. A bounce occurs when a receiving MTA refuses to accept delivery. The sending MTA might then compose a non-delivery message to the sender; but that's not under the receiving MTA's control.
That is, if the sending MTA sends a "bounce message" direct to the sender's address, that's not specified by any email RFC, and is arguably bad behaviour - similar to running a "spam-bouncer" plugin on your email client, the bounces it issues are bogus.
> all emails with a reply to of a certain msg id.
I've no idea what that means. The reply-to is an email address, not a message-id.
/In-Reply-To:.*msgid-2948573294867@example.com/ REJECT reply-storm
However, blocking the whole thread that will originate is not as easy.
I've pointed out that this technically breaches company policy regarding use of someone else's user account/identity, but it falls on deaf ears.
Agreed. But at this point, only one person has breached policy and it's a forgetfulness/oversight kind of breach.
If someone actively takes control of the machine to send an email, that's a much more serious offence.
Some of the emails were embarrassing and very funny.
Immediately someone from the list started trying to call me, and sending me threats me to report me for spamming. I think everyone else got the message though, and the replies stopped shortly after.
Everyone chimes in with something either saying "STOP", "drop me off the list" or "hey everyone stop replying to everyone"
Why is it forbidden with emails ? The functionality is useful. It is forbidden not because it is useless, but because the UX of emails make every single new message the most important thing to read and process. It is there in the Inbox, it is at the top, you can't ignore it.
Other UX make this more manageable. Usenet is used to frequent back-and-forths, so it arranges them inside their respective threads. Zulip lets you focus on the topics you're interested in and come back to a specific thread when you want.
I keep thinking we don't have a problem with email, we have a problem with UX.
Someone in a Whatsapp group has chosen to join it. They can leave whenever they want and turn off notifications as a middle ground. Someone emailing the entire office because of a massage class is just lazy or an attention seeker (and the people replying to all "please remove me from this list" are total idiots)
I blacklist the "massage class tomorrow at lunch time" ladies in our office upon the first email but the problem is I still get all the replies to all. Email lacks advanced management options.
Having an email address amounts to opting in to receiving email.
I think the reason is people not knowing how to use email. Most people make much more use of instant messaging services than email; with IM, generally you either reply or you don't.
Reply-all storms have always been a problem; I think it was made dramatically worse by Microsoft Outlook, which idosyncratically placed the cursor above the (quoted) message being replied to. That encouraged people to reply without scrolling down (because they've already seen what they're replying to); so they don't see how many quotes of quotes of quotes they could safely trim before clicking send. They just type a one-liner, and happily send off 16K of stuff that everyone else has already seen too.
Trim and contextualize. That is, keep those parts of what you're replying to directly, and remove all the dross.
It's too late to say "Please don't top-quote" - unfortunately that ship sailed two decades ago.
Well yes but from everyone. There is no granularity which is the problem.
Most stories where things go wrong involve catch-all addresses that forward to thousands or tens of thousands addresses, which either shouldn't exist in the first place or should be limited access.
I know why people opt for this sort of linguistic puffery. I know not, though, whether they realize that it has the opposite effect on anyone who has the competence and mastery that they're trying to emulate. It seems like a "no"; otherwise, they wouldn't do something that makes them look so stupid.
In the US, bananas will not get people to move, now free Ice Cream ? There will be a stampede of so many people the building would shake :)
Or is that only the Marines?
In my outlook, I usually modify the screen so thge reply all button is far far away...
[0]: https://twitter.com/alxbrsn/status/1544707673282723840
Except of course once, they forgot. Much hilarity ensued. This was also back when you could still send an email using any from address you wanted, though I can't remember if anybody used that combo.