Caution about junk mail filtering by MSN Hotmail

I noticed today that MSN changed my Junk Mail filter settings to

"Enhanced". This was done without my approval and, as far as I'm

aware, without any formal notification. I figured this out because

earlier today, I requested the e-mail address of another PN member.

After a couple of hours, I noticed that I had not yet received an

e-mail from PN providing me with the requested address. It was then

that I checked my Junk Mail Folder, which contained the e-mail I was

waiting for, as well as my Junk Mail Filter settings, which were set

to "Enhanced".

 

As far as I can tell, there is no longer an option to have Junk Mail

filtering turned off, only the choice of 3 different filter

levels--Low, Enhanced, and Exclusive. If I'm wrong about this, I would

be grateful if someone would point out how to turn Junk Mail Filtering

off.

 

I apologize if this topic has already been covered. However, it was

news to me.

Can you tell what it was about the mail that caused the "Enhanced" filtering to consider it to be spam?

No, Brian, I cannot tell why it was marked as spam. I will forward you the e-mail in the hope that it might have tags that will allow you to identify the problem.

Brian, I just requested your email address. I can't believe hotmail judges that to be spam! My best guess is that it's because it contains references to spam itself. Try taking that out, see what happens...

 

That said, a bit of Googling shows many complaints about the stupidity of the Hotmail spam filter. Very surprising as Spamassassin proves that 97% filtering with no false positives is easy. (OK, except some who are just begging for it with 100% HTML messages, try to sell many things and unsubscribe link...)

 

Steve, if you want to get rid of Hotmail, I can send you a GMail.com invite. :)

Brian, on my system, running a somewhat dated Spam Assassin 2.55, photo.net emails always trigger the rules MSG_ID_ADDED_BY_MTA_2 ('Message-Id' was added by a relay (2)) and NO_REAL_NAME (From: does not include a real name).

I knew about the SpamAssassin rules. Somebody who relied on those two criteria to filter mail would classify as spam a lot of non-marketing mail from web sites they had joined since those mails are usually generated by software. Putting a personal name as the sender of those mails would be misleading, in fact. Moreover, these days spammers put personal names on their messages anyway; so a non-personal name is actually more likely to indicate that a message is not spam.

 

As for the MTA putting a message-id on the message, most MTA's do that automatically, and I don't see how that points to a message being spam. photo.net just uses qmail, which is one of the most common MTA's. Indeed, since spammers tend not to use the standard MTA's, and all the standard MTA's put on message ids, I would hazard the guess that a message that does not have a message id put on it by the MTA is *more* likely to be from a spammer than not.

I think the logic is that most email clients put an ID on the message themselves (becuase they store them localy as sent messages anyway) and most robots would not. In any case, those two score very low points and should not cause them to be filtered by any system.

 

But I think the "personal name" would be nice. AFAIK, this is just a from address like "Photo.net BBoards <bboard-alerts@photo.net>", as opposed to the current "bboard-alerts@photo.net". Personaly, I always send messages from websites or other software I create, it just looks better!

The photo.net "address requested" messages always end up in my Hotmail Junk folder too. Since I check my Junk folder before deletion and I don't really want to read them anyway, that's OK with me.

 

Odd thing is that "@photo.net" is in my safelist, which should mean I see all email from any address at photo.net.

 

I've pretty much given up on trusting Hotmail's spam filtering, so I use it but I check it all. In my experience it has about a 10% chance of letting spam into my inbox or putting real mail in my Junk mail folder. You just can't trust it.

I must be the lucky one. My longtime Hotmail address has gone through ups and downs in spam blocking but has generally been very effective for the past year without going overboard. Surprising because for the first two or three years I used that account most web forums didn't bother to disguise the members' e-mail addresses. It was wide open for spamming and, sure 'nuff, the account was flooded daily. Not now, tho'.

 

I recall reading on some website about the heuristics of developing spam blocking software (far too technically oriented to interest me for long) that the folks who run Hotmail do pay attention to this stuff and try to stay on top of it. But it's an ongoing process - no sooner are the spam blockers adjusted to compensate for the tricksters than the spammers change tactics.

 

Frankly, I don't know why a free e-mail service would try at all, let alone make much effort. That's the attitude of Netscape's freebie e-mail: there are absolutely no user adjustable provisions for blocking unwanted e-mail. But Hotmail uses their space heavily for advertising while Netscape mail does not. The main advantage to Netscape mail is it has capacity for 5MB compared with Hotmail's 2MB for nonsubscribers.

 

Oddly, the worst of my e-mail accounts at spam blocking have always been accounts that come with the ISP service. For years CompuServe was completely spam-proof. But within a year or so after being taken over by AOL our CompuServe e-mail accounts were flooded with spam, accounting for more than 90% of received mail. My current AT&T ISP mail account is equally useless. I quit checking it more than a year ago.

 

My more web savvy friends tell me various tricks for getting spam free e-mail but I'm just too lazy to listen. To me e-mail is like telephones in the house - a necessary evil. I'd bet that 99.99% of my phone line usage is for web access.

The problem with Hotmail's filtering isn't wether it blocks spam or not, it's how many false positives it produces. And from what I hear, there it really lets down. Personaly, I would rather have 10 spams every day than 1 false positive a month!

 

I have to say I don't know how many false positives SpamAssassin generates for me, I checked for a long time in the beginning. Never spotting any, I gave up on checking altogether; I get 150 spams directed at me every day, checking them all is too much work! That said I never had any friends and other sirprised I hadn't recieved their emails. So if there ar FPs, it must be just junk that I technicaly did subscribe to...

 

I am a geek and while running my own server costs more time than I'd like, I do it anyway as it's the only way to guarantee independence and quality but unfortunately it's not for everyone.

 

As you say, it's surprising how little ISPs do. If I ran an ISP, I would redirect any mail traffic going out of my own network through my own mail servers, blocking spam and virusses at the sending end. But few do...

Slight change of topic.

 

photo.net is a sender, recipient, and forwarder of email. We originate on the order of tens of thousands of messages per day. These are primarily email alerts for forum and photo posts. But there are also welcome messages for new members signing up, subscription expiration notices, and a few other smaller categories.

 

We receive about 40,000 messages per day, most of which are forwarded to the 2000+ subscribers who have @photo.net forwarding addresses. A fair number of these messages are bounces of our outgoing messages caused by people requesting alerts when they have an invalid mailbox setup on their photo.net account. A very high percentage of the 40,000 incoming messages are bounced back, and a fair number of the bounces bounce back.

However, while I don't have the exact numbers, I think we are probably forwarding a lot of spam to the @photo.net addresses that people have signed up for. We do not run any spam filters on the @photo.net forwarding addresses, although we do run them on our own administrative addresses.

 

Question: should we run spam filters on the @photo.net forwarding addresses that subscribers have set up. Would that be regarded as a feature or as a problem by subscribers?

 

Speaking for myself, if I were a subscriber, I think I WOULD NOT like this, since I prefer to do my own filtering, and be responsible for the false positives myself. But what do other people think?

I think if it was offered, maybe it should be offered as an option.

 

I'd rather get the RAW unfiltered email and do the SPAM rejection locally. That way I can review what's flagged as SPAM and retrieve the stuff that's falgged incorrectly.

 

My guess is it might cause more work for photo.net, dealing with complaints that the SPAM filters were rejecting legitimate email.

I don't use photo.net's mail service but I suppose I'd prefer to control spam filters at my account. I get a bounce now and then to the freebie account I use for photo.net purposes. Invariably these are from folks who've either got an invalid or out of date address or have their filters set too narrowly.

Maybe you could scan, then just flag in the headers. But not sure how usefull that is. I don't actualy use my PN address at all, though.

 

Have you ever thought of going all the way with email for PN, offer web based email? It could be a way to pull in more patrons, and get renewed subscriptions even if people don't use the site that often anymore, just so they can use their cool/permanent email address. In that case, you'd have to offer Spam and virus filtering, obviously.

 

I did some testing for another project recently and concluded that a modest Intel/Linux/Courier/SquirrelMail box could easily support several thousand subscribers. Authentication against the PN database would be a breeze as well.

 

Just a thought...

On a recent check of my Hotmail junk mail filters I found a message that was related to photo.net only in that it had "photo.net" as part of the subject line.

 

I haven't had time to test this, but I wonder if Hotmail really has it in for photo.net and deletes all mail with "photo.net" in the subject line, as well as email it thinks is spam on the basis of the sender's email address!