Logo Background

Multilingual spam

  • I regularly correspond using email with many people overseas, including several Japanese friends. Recently I have been receiving up to four unwanted Japanese language emails each day - some of these are adverts for Viagra and others look like inappropriate websites. Despite running the full gamut of Norton security software, most of these emails are not identified as spam. Prior to deleting each message I always block the address of the sender, but this has hardly affected the flow of new spam. Can anything else be done to stop the receipt of this spam, and is there any point in collectively advising my Japanese correspondents that one of them presumably has a worm in his system? Alternatively, could I prohibit all mail containing Japanese characters, as my correspondents address me in English?

    While getting these emails may be a nuisance, only receiving four such spam messages per day is quite a manageable amount compared with the amount of spam some people receive. As you have discovered, spam filtering software is not perfect. Based upon the features and elements of a message, spam filtering software uses probabilities to determine the likelihood of a message being spam. So occasionally some messages will be detected as false positives (legitimate messages which are tagged as spam) or false negatives (spam messages which are not tagged as spam). Perhaps you could try adjusting the Norton spam blocker settings to see whether you can set the filter to act more aggressively, and lower the threshold at which the software determines that a message is spam. Be aware, this could result in legitimate messages being tagged as spam so you should regularly check the spam folder in your email client to ensure no such legitimate messages have been accidentally filtered.

    While blocking the senders of such messages may help curb the inflow of spam, whenever a mass unsolicited email is distributed it will generally be sent using a fake ‘From’ address. This is called spoofing the header of the email message, meaning the header of the message (which contains details such as the message sender) is falsified. This is very easy to do and is the reason why spam messages generally come with strange and varying email addresses. Advising your Japanese friends about the possibility of a worm or virus on their systems would likely not greatly help since the spammers already have your email address. While there is an outside possibility that one of their computers has a virus, it is hard to say for certain. That said, it would certainly be worthwhile to suggest a scan just to make sure.

    There are a few other procedures we could try to curb the amount of spam you are receiving. Be aware, these procedures will not work if you are using Outlook Express, as the message rules in Outlook Express are not flexible enough to accommodate these more advanced rules. My recommended email client is Mozilla Thunderbird (www.mozilla.com/thunderbird) which is an excellent free email client.

    First, you could explore whether the email is actually addressed to your email address. Often, when spammers send out mass emails they are not specifically addressed to a particular email address. Open one of the spam messages and check the ‘To:’ field. If the ‘To:’ field does not list your email address, we could use this aspect to configure a mail filter which filters email, not specifically addressed to your email address, into another folder (e.g. a spam folder). In Thunderbird, go to the ‘Tools’ menu > ‘Message Filters’ and click ‘New’. In the window that appears, give the filter a name (e.g. ‘Non-addressed email’). Enable ‘Match all of the following’. In the filter criteria, select ‘To or cc’ then ‘doesn’t contain’ and enter your email address. We next need to add the relevant action. Select ‘Move Message to’ and then select the folder where you want to move such messages. Once this is done, click OK. Make sure the rule is ticked (enabled) and close the Message Filters window. Now, whenever the ‘To’ or ‘CC’ field on emails does not contain your specific email address, the message will be filtered (moved) to the nominated folder. Be aware, this could result in legitimate emails being filtered, especially if you subscribe to mailing lists, or receive group email messages, which do not specifically address messages to your email address. Should this be the case, you should be able to configure the mail filter further to allow receipt of such messages.

    Second, as suggested in your question, it may be possible to filter emails based upon the message body containing Japanese characters. There is no easy way which is guaranteed to always work, but there are a few approaches we can try. First, you could copy a few characters from one of the Japanese emails and create a mail filter with the criteria that if the email message body contains those characters, then move the message to a particular folder. In such cases, the best approach is copying as few characters as possible, to maximise the chance of catching that combination of characters in emails. Be aware, the characters could be detected in legitimate emails so you may get false positives.

    The second approach is creating a filter based upon the ‘content-type’ header of the email message. These headers denote the type of content in the message and, depending on the client used to send the messages, sometimes contain a ‘charset’ attribute defining the encoding of the email message. If the spam messages contain the ‘charset’ attribute, and it has a value representing Japanese, we can filter based on this attribute. In Thunderbird, select one of the spam messages and go to the ‘View’ menu > ‘Headers’ > ‘All’. This will display all the email headers. Look at the ‘Content-type’ header and check whether there is a ‘charset’ attribute (e.g. ‘charset=ISO-2022-JP’). If so, we are in luck. If not, we can’t proceed any further.

    Assuming the ‘charset’ attribute is present, have a look through several legitimate emails to check whether there is any charset attribute and, if so, whether the value is different (so the filter does not act on legitimate emails). Assuming the legitimate emails do not have this attribute, or the attribute has a different value, we can proceed to setup a filter based on the charset equalling the Japanese encoding. Create a new mail filter (as outlined previously) and for the filtering criteria select ‘Customize’. This will allow us to specify a new message header criterion. Enter ‘Content-type’ and click ‘Add’ and then OK. This will add ‘Content-type’ as a criteria option, so select ‘Content-type’ then ‘contains’ and type the charset attribute with the value (e.g. ‘charset=ISO-2022-JP’). Configure the action you want to happen (e.g. move the email to a particular folder) and click OK. Make sure the filter is enabled, and now all messages which contain that character encoding will be moved to the particular folder. As previously advised, make sure you regularly check the folder for any false positives!

Leave a Comment
Hi there. If this is the first time you are posting a comment it will not appear immediately, but needs to be approved. This is necessary in order to combat comment spam. However, once you have submitted a comment (which is subsequently approved) you do not need to go through this process again - the site remembers who you are and auto-approves your comments. Nifty eh? Anyway, sorry about the inconvenience that this may cause for your first comment post.