Q: How does the learning work?

A: Each e-mail is turned into a series of words which are then learned individually by the program. When you identify the e-mail as spam the words are learned in the spam file and for a legitimate e-mail the words are learned in the non-spam file. When a new e-mail comes in, the words in it are compared to the spam and the non-spam files and a decision is made whether to mark it as spam, possible spam or not spam.

Q: What does MishMash mean?

A: MishMash is a routine used by ESE Inbox Sweeper to break apart and combine words from the e-mail. Many spammers like to include spaces in words, for example "vi ag ra" or "h u m a n g r o w t h h o r m o n e". The MishMash routine is smart enough to recombine these letter sequences into normal English words. Another function of the MishMash routine is to break words apart that have been mashed together by the spammers. Running "healthproductsnow" through the routine outputs "health products now". In both cases the learning will work much better when using proper English words.

The Full version will add the recognition of number and other symbols substituted for letters. For example, "pr0blem s0lved" will become "problem solved" and "V1agra l@rger" will become "viagra larger".

Q: What is a domain?

A: A domain is the last portion of an Internet address, i.e. ibm.com or whitehouse.gov.

Q: I use Outlook Express, how do I find my POP3 server address and my POP3 username?

A: Start Outlook Express. Click Tools in the menu bar. Click the Accounts tab and select your mail account. Click the Properties button. Click on the Server tab. The Incoming(POP3): is your POP3 Server Address. The Account name: is your POP3 username.

Q: Why does it take so long to load?

A: The program is reading in a list of English words. This list is used by the MishMash routine to form proper English words.

It is also processing what needs to be deleted from the "Saved E-Mails" area.

The program also checks on each startup if there is an imbalance in the total number of spam values to the total number of non-spam values. If there is a difference of more than a few percent then one of the files is adjusted down. This may take a few seconds to do. Also, note that if you have cleared the learning files that this "balancing" of values will only occur after each file has learned at least 40,000 words (very roughly about 600 total or 300 emails for each) and also each file must have learned at least 5,000 unique words. Check the ESEInboxSweeper.log for the status of these values.

Q: Why is there no blacklist function?

A: Blacklists are not needed by ESE Inbox Sweeper because the senders e-mail address is part of the learned words. This also keeps friends and friendly domains from accidentally being added to a blacklist. Blacklist servers are not used because an e-mail you may want to receive is on the servers blacklist, but the server would be incorrectly telling the program that the e-mail is spam.

Q: I am getting a warning message about deleting a message that wasn't fully downloaded. What does this mean?

A: The Sweeper program retrieves only as much of the e-mail from the mail server to determine if the e-mail is spam or not. If the e-mail has been marked for deletion and the entire message was not downloaded, you will not be able to fully recover the e-mail in the "Saved E-Mails" area and you will get this warning.

You should decide to either unmark that e-mail for deletion or let the Sweeper delete it even though you will not be able to recover the full e-mail later.

You can turn this warning off under the Tools menu.

Q: When I recovered a "Saved E-Mail", the entire message wasn't there. What happened?

A: The Sweeper program retrieves only as many lines of text from the mail server to determine if the e-mail is spam or not. There is no option to retrieve the entire message for safe keeping to the "Saved E-Mails" area. If you need this capability, upgrade to the Full version.

Q: How does it handle the gibberish words in the e-mails the spammers send, and why do they send it?

A: ESE Inbox Sweeper turns each word from the e-mail into an integer value, this is called hashing, and sometimes different words hash to the same value. The spammers are hoping that one or more of their gibberish words hash to the same value as a word with a high non-spam value as this would increase their chance of having their spam e-mail appear to be legitimate.

If the e-mail contains more than a few gibberish words then the probability of being marked as spam will be greatly increased.

Q: What does it do with spams that are only graphics?

A: ESE Inbox Sweeper examines the senders name and address, the subject and the text of the e-mail. If there are very few words in the e-mail, i.e. fewer than fifteen words, and the words are not identified as spam words then the status will be set to possible spam.