English

SPAM-Filter for K-9 Mail?

Is there a possibility to install a SPAM-Filter for K-9 Mail?

1 Like

I believe a spam filter is the job of the email provider and not the email app.

I use k-9 with riseup.net and never receive spam.

Well, this depends. Thunderbird brings its own spam filter.
A email provider’s spam filter necessarily requires that the provider is analyzing your emails.
Besides that my local spam filter (in Thunderbird) is working significantly more precise (after some training) than my provider’s filter.

2 Likes

According to what I’ve found via web search, K9 has a build-in spam filter, but it only works with POP3 and not with IMAP.

1 Like

I’m using POP3 and it does not work.

If you access your mails from multiple devices, a SPAM filter on a single device won’t do.
I’ve disabled the filters in all my clients and set one up for my account. (Luckily, my mail provider allows detailed configuration.)

@YtvwlD and @paulakreuzer :
This are no answers to my question. My provider has a spam filter, which filters most of the spam. Some spam mails come through and I want to filter them on client side.

And I do not want to use intelligence services like Google (gmail), Microsoft (hotmail) or the like.

As far as I red now zhere is no working spam filter for K-9 mail, right?

Well those are maybe not the answers you were looking for but we provided you with alternatives. I share your dislike of google and microsoft, because I don’t want giant companies profiting from spying on me. If that is your concern too than my and probably @YtvwlD’s suggestions are quite the opposite of that.

I can only speak for riseup. It’s an open source, privacy aware, gratis and donation-dependent provider. They work tirelessly to make sure that the NSA and hackers can’t access your data. They were threatened by American intelligence agencies, but never gave up their principles.

What you write about riseup sounds good. Nevertheless I don‘nt want to change my hosting provider who also provides my mailserver. Again my question: Is it right that there is no working SPAM-filter for K-9?

(Don‘t know why I am not notified of new posts anymore.)

The information I have found in German here says that the spam filter only works with IMAP.

@Subhash It is the duty of the service provider to eliminate spam. You should send them a list of all the spamming email addresses so that they can blacklist them.

Basically the modified K-9 just provides the option to create an email blacklist.
Since nowadays (at least as per my experience) almist all spam comes from varying generated sender addresses, tgis does not really help.

Imo the same is true for sending email addresses to the provider. Since the next spam wave will use other mail addresses the provider will either need to blacklist the spamming IPs or feed the whole email into a self-learning statistical filter.

1 Like

That’s what they will do, won’t they? They will block the whole IP where the mail came from.

As far as I know, most spam filters work a bit more complex than just blocking IPs or mail addresses, which would be endless work. Instead there are some algorithms in the background, that analyze the structure of the mail.
For example, spamassasin, which I have installed on my mail server (and which seems to work very well as I am hardly ever bothered by spam - even though my mail address is publicly posted in several places) is described as a

spam filter with numerous features including automatic white-listing, RBL testing, Bayesian analysis, header and body text analysis.

So there’s more behind it than just blocking addresses/domains/IPs.

I think spam filters on clients only sort out mails that have been indexed by server side filters as spam, but don’t analyze the mail again. E.g., my Thunderbird junk filter setting only checks if mails are marked by spamassasin as spam, and then sorts them out to the junk mail folder.
So I would agree that a decent spam filter needs to be implemented on your server, not your client!

Though I am not an expert on this topic, this is just from my little understanding on how that works.

2 Likes

Well, I’m not sure about that, but I agree that filtering is most useful when done on the mail server.

But blacklisting email addresses is not really feasible. They can be easily changed, because when there’s no protection such as DKIM anyone can enter any address as sender.
Blacklisting may be useful if done by your provider and if they block IP addresses that are currently sending many spam mails.

That would be the right thing to do.™

But there are problems with RBLs, too. (There are many cases where they block providers who have done nothing wrong and refuse to correct or even admit their mistakes.)

I’m running DSPAM which does a Bayesian analysis and my provider uses a few blacklists (if I’m right).

2 Likes

Afaik it’s usually not that easy anymore. Of course if a single server is sending out spam all the time the URL is blocked. But since spam mostly is distributed by botnets (from devices / IP addresses that also send regular mails) you need to do a deeper analysis including sender, mail text, number of mails sent, number of receivers and probably more. Then a statistical filter can calculate a confidence value that a single mail is spam or not.

This leads to the fact that usually the first mails of a new spam wave get through the provider’s filters (plus false positives btw).

Bottom line: if your provider can’t provide a spam filter (or don’t want him to) and don’t want to sort it out yourself manually, K-9 as email client won’t do atm.
Don’t know if there are alternative Android mail clients that implement a spam filter.

Edit: looks like Freibadschwimmer was quicker.

Don’t have the docukentarion at hand atm but Thunderbird ca do more than that. Iirc it brings (optionally?) a Bayesian filter like @YtvwlD mentioned) which I had to train in the beginning with > 100 spam mails and that learns with every additional mail I manually classify as spam. In the meantime it does significantly better than my provider’s filter. So for me it’s a perfect add-on to the provider based filter.

This topic was automatically closed 182 days after the last reply. New replies are no longer allowed.