Hmmmm, nice FAQ and thanks for posting but I don't see anything particularly new in there.
However, quoted from the most recent service.status report:
What this is means for now is that less email is being accepted onto our platform and email meeting the following criteria is being rejected.
- Email that is detected as spam by our 'ClamSpam' filter (one of the detection solutions we use)
- Originating IP address of the sender is blacklisted on an RBL (list of known spammers). For more information on this method of blocking spam, see here:
http://en.wikipedia.org/wiki/DNSBL This particular method of spam blocking has always been implemented on our incoming mail platform.
I've been meaning to make enquiries about the recent changes to mxlast regarding spam detection but haven't quite got around to it ! I realise that now is probably not the best of times to ask but as it's going to be implemented across the board, I'm going to ask anyway.
I have a major dislike of any spam filtering done 'behind my back' where I do not have any control over what's happening and may not even know that it's happening. So many other ISPs and webmail providers frequently remove what they
think is spam but in reality, chopped ham and pork hasn't been within several hundred yards of the messages !
A classic recent example concerns hotmail and others who decided in their 'wisdom' that the Booking Confirmation e-mails for Glastonbury Festival tickets purchased from a very reputable Ticket Agency were automatically classified as spam and silently deleted on receipt REGARDLESS of the users specific account settings and obviously without their knowledge. It caused way more than a bit of chaos to say the least as many 1000's if not tens of 1000's of people failed to get confirmations and other very important communications from the Agency !
For almost 10 years up until now, I had been completely spam-free on all my used accounts. I had no need for spam filtering and therefore have none implemented despite easily having the ability to do so. I have Mailwasher Pro on all my machines which I often use to have a quicky look at what's on the PN (and various other) servers as well as using Agent news/mail reader on all machines to DL messages from all over the place as/when required. Both of these applications have way more than enough facilities to filter spam by various methods including Bayesian techniques should it be necessary although these facilities have always been disabled to date. Having a 'catch all' is very important to me and I do not generally use mboxes but DL everything and filter/sort/distribute locally.
Whilst the spam issue is a right PITA to say the least, it hasn't as yet reached the level where my harvested addresses are totally swamped or Random_Chars@ or Good_Guess@ prefixes are being used. It's only a matter of time before that changes of course IMHO. Although I appreciate that PN are trying to reduce the impact on customers (and the servers of course) by detecting spam on receipt, like I say, I strongly dislike detection/deletion going on behind my back and completely outside of my control.
I don't even have the optional spam tagging switched on and don't intend to do so anytime soon either. If spam detection becomes absolutely necessary in the short term then it is something that I will implement myself so that it is totally under my control. If/when the spam problem gets completely out of hand then it's going to be a very appropriate time to move on - simple as that. If I'm forced into changing e-mail addresses and/or usernames that have been in constant and regular (personal and business) use for ~10 years because of this security breach then I might just as well change them to something that doesn't have plus.com at the end.
So, what exactly is this ClamSpam filtering, what does it do and how does it do it ? And more to the point, how can I be 100% certain that I am not EVER going to have e-mail deleted on receipt or delivery refused just because someone else thinks that it
might be spam when in reality it is a genuine and wanted message ? I get the impression that both of these systems blackhole rather than 'tag' anything considered as spam which I'm most certainly not liking the sound of. I note that the volume of general spam received on my original F9 A/C addresses has decreased quite significantly over the last few days and I therefore suspect that this is due to these changes and perhaps also because the thresholds have been tweaked to reject more potential spam. Deleting spam at source is a very good thing of course but only providing that the system is 100% guaranteed not to get it even slightly wrong ... and thereby hangs my problem unfortunately. I'm very sorry but to be brutally honest, there is no way that I really trust PN to decide what e-mail is and is not delivered to me.
