Improved Usability
Logs throughout IMail now use a single message ID for a message making it easier to track a message as it passes through various tests and processes. Messages flagged as spam can now include this ID in the header. The anti-spam UI options make it easier to configure.
 

Increased Effectiveness
Added subject modification allows for messages flagged as spam to include a warning in the subject making it easier for some email clients to quarantine or delete messages flagged as spam.

Also added is the the ability to move messages flagged as spam to a subfolder and the ability to use the same phrase list to search subject, body, or both. What's more you can also have IMail Server normalize words so that it can remove non-alphabetic characters from a word and use that in content filtering.
 

Better Prevention of False Positives
Spam checks can now automatically be skipped for trusted relay IP addresses. Trusted IP addresses, white lists by domain, and white lists by email address are all in the same location and can apply to connection filtering and content filtering.
 

New Word Normalization Feature
The addition of word normalization can dramatically increase the effectiveness of anti-spam filters by stopping tricks such as putting periods between each letter in an attempt to fool spam filters (example: r.e.f.i.n.a.n.c.e). Word normalization removes any non-alphabetic characters within a word prior to applying Bayesian Statistical Filtering or Phrase Filtering. The same phrase list can now be applied to both the subject and body, decreasing administration time.

Real-time Blackhole Lists (RBL) lists third party services that include domain names and IP addresses of known spammers; Comes with 5 free RBL lists by default and can be configured with additional ones by the administrator like ORDB or MAPS.
 

Verify the MAIL FROM validates that the MAIL FROM address actually exists.
 

Perform Reverse DNS Lookup verifies the DNS record for the sending servers IP address.
 

Verify the EHLO/HELO domain validates that the sending mail server domain exists in DNS.
 

Kill Lists (block by domain or email address)
 

Trusted IP Addresses

Bayesian statistical filters use spam’s characteristics against itself. Words are analyzed based on how frequently they appear in good email versus bad email. Then the message is assigned a probability of being spam.
 

Phrase Filtering searches for phrases in the body of the messages.
 

Trusted IP Addresses & White Lists of trusted e-mail or IP addresses that will allow messages through, minimizing false positives.
 

Advanced HTML Filtering Detects nested tables, hyperlinks, images, scripts, invalid tags, mailto links, deceptive URLs, and embedded comments
 

Blacklists URLs that appear in spam messages (most spam emails contain a call to action to visit a web page and these URLs can now easily be blocked)
 

Removes the HTML tags prior to the statistical analysis of the words in the message (useful when dealing with embedded comments or invalid tags that are used to intentionally confuse statistical filters)

Scan the message Header (To, From, Sender, Subject) or the body text
 

"Contains" and "Does Not Contain" logic
 

Deliver mail to existing or new mailboxes, or delete it
 

Pattern matching with regular expressions for finding near matches or matching more than one string
 

Finding specific file types, such as the .vbs files associated with the "lovebug" viruses
 

Eliminate obscene language
 

Stop confidential information from going out and viruses from spreading
 

Provides the ability to limit e-mail usage

If FROM doesn't contain "@ipswitch.com" move message to "LowPriority" mailbox
 

If SUBJECT contains "get rich quick" delete message
 

if BODY contains "(wsftp|ws_ftp)" move message to "WS_FTP" mailbox