Do filters only work with webmail?

[chrism]

I've set up a filter rule on the spamcop webmail site to try to sort out the current problem of vast amounts of German spam. It appears to work OK from the webmail site - mail is being deleted by the filter, as it should be.

I generally, however, download mail from spamcop via POP3, and mail downloaded this way doesn't seem to be getting deleted by the filter. Forgive the silly question, but is the filter only applied if I read my mail via the webmail site? I had assumed - perhaps wrongly! - that filters were applied to mail as it arrived in my mail account, and hence would also apply to POP3.

Thanks in advance for any assistance,

Chris

[StevenUnderwood]

You are corect that the filters only work when you access via the webmail application. The webmail application is only an IMAP mail client, like your desktop client you use to read mail there. It is not a part of the server where the mail is held.

You could setup a filter on your local client to do the same thing. Or as I do, access via webmail throughout the day and download only the good messages to the local client.

[cliffski]

both of which kind of defeat the object of a paid spam filterer

[chrism]

The situation does seem to be slowly improving. My 800 overnight messages on Friday was down to 370 this morning, so it seems as though the machines which are sending it are slowly being added to Spamcop's "blacklist". Hopefully it'll die out in another few days.

Regards,

Chris

[StevenUnderwood]

both of which kind of defeat the object of a paid spam filterer

The spamcop spam filters (the blocklists) and the override filters (personal black and white lists) both work without logging into webmail. These are the spam filter functions.

The only filters that are webmail specific are the ones that can move any message (not just spam) to any other folder. There is nothing spam specific about them. Think of it as a free bonus feature for using webmail.

Another way to look at it:

The parts of the system that AUTOMATICALLY determine whether a message is probably spam work without your intervention (webmail login).

The parts of the system that you are programming, require you to login for access.

[Ivan]

I'd like to add my voice to the users who feel that SpamCop filters

should work regardless of how one chooses to check their email.

This is really a huge oversight and puts a large dent in the usefulness

of SpamCop as an email service.

It also ticks me off that I had to come to this forum to discover why

my filters were not working. The filter page should have some sort

of disclaimer about this incredibly lame limitation.

Or better yet, SpamCop should simply run the filters every few minutes

which is what most users assume is what's supposed to happen.

This really needs to be corrected.

[StevenUnderwood]

I do not see how it "puts a large dent in the usefulness of SpamCop as an email service." Spamcop email service is there to remove the spam it detects (per your settings) from your inbox and holds it in the Held Mail folder. It does that whether you forward your mail to another inbox or access it from another client (desktop or webmail).

Any additional filtering is done at the client level (desktop or webmail). The current release of Horde IMP does not support that type of access and I doubt securitywise it would be a very good idea (giving one machine access to delete messages on another machine without the owner interaction). I know of no webmail based client that can run the filters on the server without being logged in. If you do please let us know and maybe JT will implement it instead of the current Horde IMP product being used (doubtful).

I can see it now, "every message sent to spamcop is missing...give me back my messages" only to find out that person had setup a filter to delete all messages with "From" in it (or something else dumb like that).

Does your desktop client access your email without you starting it first? Same concept. The webmail client is only an human interface to the messages stored on the server.

Again, the blocklist filters and the whitelist and blacklist filters are functions of the server. The other filters are a function of the client. Both are administered by the client.

Your idea of better documentation is a common thread throughout all of spamcop. Because I understand how servers and clients interact, I understood how the filters would work and had no problems with them. In fact, if they were automatic, I would have to disable them as they do things (like discard certain messages rather than move them to trash) after I have viewed them.

[Ivan]

With all due respect Steven, all of your criticisms are completely nullified by the simple fact

that SpamCop does due all of the things you are concerned about, if you log on via Webmail.

>>I can see it now, "every message sent to spamcop is missing...give me back my messages" only to find out that person had setup a filter to delete all messages with "From" in it (or something else dumb like that).

What is to stop this from occuring with a SpamCop user that logs in via Webmail?

>>Does your desktop client access your email without you starting it first?

YES!!! IT DOES!!! I have my Eudora set to check mail every 10 minutes.

Many many email users do likewise.

>>Again, the blocklist filters and the whitelist and blacklist filters are functions of the server. The other filters are a function of the client. Both are administered by the client.

I spent very little time manually activating the Spamhaus, CBL, or other filters I use.

SpamCop checks each piece of mail, as it is received against these filters.

What a concept!

>>I know of no webmail based client that can run the filters on the server without being logged in.

This confuses the whole issue.

SpamCop's purpose is not to be a webmail provider.

The purpose of SpamCop _is_ to filter email.

If there were no spam to filter, then there would be no reason for SpamCop to exist.

So the question is not about who's the best webmail provider, it's about providing filtered email.

Right now, my domain runs through my hosting provider, which allows me to custom

filter key words, or return addresses, I think are spam-indicative.

IT DOES THIS AUTOMATICALLY. I never log in.

If it trashes a message, I'll never ever see it.

And then forwards my email to my SpamCop account.

However email addressed directly to ivan[at]spamcop.net receives no such protection.

I've set up filters in SpamCop, but they have no effect because I check mail via POP.

Why on Earth should SpamCop not be able to do what nearly every hosting company can do,

offer _automatic_ user customized filtering to its customers?

There should be little to no trouble in programming the filters to run automatically.

All you'd have to do is take the already existing code that runs the filters when a user logs into webmail and set that code to run every few minutes.

Bottom line is that SpamCop is failing at its core mission which is to provide me with filtered email. If SpamCop's response to this is going to be that I should just filter things on my side, then why should I continue to be a SpamCop customer. Again, filtering the email IS the service I'm paying for and expect to receive.

As I continue to receive virus emails with "SUBJECT: Hi, Nick. In this archive you can find all those things, you asked me." that I should be able to easily filter out, I stand by my statement, there is a large dent in the usefulness of SpamCop as an email service

[StevenUnderwood]

It seems you want additional, personalized filters built into the server, which is not a bad idea and not something I am arguing here, but not the job of a webmail client, which is where the filters you are complaining about are handled.

I don't know if you are confusing the 2 different functions being provided (message processing and webmail access to those messages). You seem to really want the server modified to allow custom filtering.

With all due respect Steven, all of your criticisms are completely nullified by the simple fact that SpamCop does due all of the things you are concerned about, if you log on via Webmail.

Correct, with interactive control of those things. I get to do them if and when I want, not all the time.

What is to stop this from occuring with a SpamCop user that logs in via Webmail?

For one, the log messages that show up at the top of the screen for every transaction as they happen.

YES!!! IT DOES!!! I have my Eudora set to check mail every 10 minutes.

Many many email users do likewise.

I do the same thing. But we have started the Eudora client manually and left it running on the desktop. It will not run when we don't want it to (we shut down the client). Not the same thing as having the client start itself and process the messages with no control over the process.

This confuses the whole issue.

That IS the issue. You are asking a webmail client to do something it is not designed to do.

The purpose of SpamCop _is_ to filter email.

No, the purpose of the spamcop email service is to filter messages sent from known sources of spam. And that it does without you even touching webmail except for you to activate it. You want it to be something which it is not.

The webmail application is basically an out of the box applcation to allow you access to your filtered messages. There are some minor modifications made to it so you can configure the server functions (setting up the dnsbl's) and easily report the spam to spamcop.