Common Junk Classification as a Cloud Servie

[mmeinesz]

Junk filtering has become one of the most important features of email frontends.

Webmailers like GMail do this automatically: If a user rates a message junk, it will be rated as possibly junk for all users. The more active users, the better the junk filter. If I am the only one who classifies my email (like with Thunderbird) I will never get a good junk filtering.

Thunderbird has no central instance (as far as I know of). A common Junk classification would need a cloud backend. I don't have a problem with that. I am even happy to pay for a service like that.

I might not be the first with this idea. If this is a duplicate, I apologize. Otherwise I am happy for upvotes 😉

[Jon]

Thanks for submitting an idea to the Mozilla Connect community! Your idea is now open to votes (aka kudos) and comments.

[MattAuSupport]

Personally, I would be quite opposed to such a system. 

There was one US based ISP about 10 or 15 years ago that used a system exactly like you have suggested to crowdsource how junk on their server was treated.  They quickly learned how many folks used the junk buttons to delete mail, and how many used the junk button as an alternative to unsubscribing from mailing lists.  Next thing, they were inundated by complaints from customers about non-delivery or delayed delivery of their chosen newsletters and coupon emails. One person's junk is another's treasure.  The idiom applies equally to car boot sales and email.

SPAM definitions are a very personal thing, and for this reason Thunderbird learning algorithm appeals.

There are some obvious indicators, of course, like the email address and the server the email originated on not being associated at all.  But that is something that is managed using server things like DKIM, DMARC and SPF

Perhaps the most important thing about SPAM is the sender is fundamentally irrelevant in the process as SPAMMERS do not tend to use their own email addresses in the from of email spam.  They use those "borrowed" from someone else.  A check of just about any email support site will see instances of folk complaining abut the backscatter from spammers using their email address, and they get all the nondelivery notifications.

Despite the desirability of thinking that using the collective wisdom of folk to determine what is spam.  The reality is, what I consider spam will not very often match what you define it as spam. 

For instance, in my country if I get unsolicited email from a business touting their business or products.  I do not flag it as SPAM.  I forward it to the federal government, and they prosecute the business for sending out.  They fine them large amounts of money,  in the order or tens of thousands of dollars for each email sent.  So legitimate businesses in my country are very sure to get consent for their mailing lists.  They are also very responsive to requests to be removed from those lists as not doing so get an even bigger fine.

In Europe,  they have the GDPR which also imposes huge obligations on anyone with your email address as to what they can do with it and when.  Again unsolicited commercial advertising is not such an issue because there are effective laws in place.

In the USA by contrast,  it is considered the right of the business owner to harass and flood whatever contact means they choose with unsolicited garbage.  Be that email, SMS, phone calls etc.  So in that part of the world we see individuals trying to use technology to fix a social issue.  This is something that very rarely works.

I mention those three, because they are three I know a little about. Not because they are particularly unique. The point is that how spam is handled and managed really varies country to country.  Some have basically nonexistent laws about it, others have very strict ones. It does reflect in the amount of spam that is sent and received.

Many folk are reluctant to use unsubscribe links, there is this story about how doing so confirms the validity of the email address and so should be avoided.  There are two issues with that;

1.  The mail server accepted the mail for delivery to you.  That is something that would not have happened in the majority of cases if the email address did not exist.  The server would have rejected the email.

2. Spammers don't care if a significant proportion of their lists are to nonexistent addresses. They are really only expecting a response rate that can be measure in the points of one percent of mail sent anyway, and rarely make any effort to clean their lists of dud email addresses.

So I would prefer a spam tool in my personal email client that respects my decisions overtime, learning from them based on each decision I make. I will leave the broad brush checks to my mail provider who on the whole does a pretty good job on Nigerian scams and the like.