Hacker News new | ask | show | jobs
by achamayou 2978 days ago
An important feature of Gmail from the get go was how well it separated Spam from relevant email. I don't think you'll find too many users who prefer to do that manually.

It's a matter of degree of course, but scary seems too strong a word.

3 comments

Yes, up to the point of marking legitimate e-mail from personal domains as spam, even though such domains were set up correctly, including stuff like DKIM or SPF.

Also, the wording was "... and what can wait for later", which doesn't sound like spam was meant - I don't think people keep their spam messages and read through them later, e.g. over the weekend. :)

Maybe I'm an exception, but I do scan through my spam from time to time :) Partly to see if I've missed something, but also because it can be amusing.

I agree that spam detection is not an exact science, and that it is in fact quite subjective. It's also quite valuable!

Yes, I do that too, for certain class of spam that did not score high enough in my spam filter solution to warrant automatic removal. :) But that's about 5% of total spam I get.
Right, because hiding without letting people know they have something is spam is good. /s.

It's not. This is how and why people keep missing actually important email, because friend X changed their email address for reasons, but they are not in the contact list, so it must be spam/phishing.

As I mentioned elsewhere: spam can be filtered objectively, but importance would have to be identified subjectively.
Can it really though? If my ISP, who I've exchanged emails with before or sends me my bills starts to send me promotional material about some TV bundle, is that spam?

Conversely, if someone I've never exchanged emails with before, but who got my details through a third party wants to discuss a business opportunity, is that spam?

If you define spam narrowly as "unsolicited", then sure, it's somewhat objective (although even then, you have to work what was solicited in ways other than an initial email). But there's quite a bit of subjectivity involved in the more useful and generally accepted definition.

> "spam can be filtered objectively"

"Objective filters" (by which I assume you mean "rule based") can easily be worked around by the people trying to get through the filter. All you gotta do is figure out the rule and bam, you are back in again.

Sure it might work for your tiny little family domain name with your tiny 10 mailbox email server. No spammer gives a crap about you. But when you are on the scale of gmail, you will have a lot of people whose entire purpose in life is getting through your spam filters. "Objective filters" turn into a continual game of whack-a-mole. Put in a rule, wait half a day until somebody works around that rule, put in another rule.... repeat ad nauseam.

"Objective Filters" don't scale at all. They might filter out the riff-raff script kiddies but anybody whose livelihood is based on getting past your "objective filters" will easily overcome them.

How does this magical objective spam filter work?

How can that same technique not be used for marking important emails?

Because of the word "objective". Different things are important for different people.