Mail on the Internet is transferred around using something called the Simple Mail Transfer Protocol (SMTP). News is transferred using Net News Transfer Protocol (NNTP). This discussion pertains directly to SMTP and mail programs, but most of it also applies to NNTP and newsreaders.

SMTP (and NNTP) assumes that each of the characters in the mail message is defined by 7 bits, and is therefore limited to a total of 128 (2x2x2x2x2x2x2) characters. This character set us called US-ASCII (American Standard Coding for Information Interchange). Most of the terminal emulation programs and other tools available commonly available across a whole range of operating systems utilize the ASCII character set as well, but more on that later.

Now, 128 total characters is a bit limiting. It's only a little more than the keys on a standard keyboard (each of which has its corresponding code) and a few other things such as EOF and EOD (end of file and end of data) tags which you never see but are vitally important behind the scenes. But, this 128 character limit means that there are no codes for all the other symbols used internationally, such as an "o" with an umlaut, or the fancier things used in word processing such as complementary curved double-quotes. If SMTP encounters one of these characters, the only thing it can do is to convert it to one or more characters it knows about. For instance, an "e" with an accent grave is converted to the character string "=E9" by SMTP.

There are ways to get around this limitation, though. The most commonly implemented standard is a MIME (Multi-purpose Internet Mail Extension) compliant encoding defined by the International Standards Organization (ISO) in their specification #8859, Part 1. This is more commonly known as ISO-8859-1. This specification utilizes an eighth bit for character encoding, allowing a total of 256 characters, which enables both the inclusion of lots of special characters, like that umlauted "o", but also the direct transfer of binary data files (because that binary data is based on 8 bits, i.e., 8 bits = 1 byte) such as a spreadsheet or database file.

So, it is possible to transfer this type of thing across the Internet - probably happens millions of times a day - but it is required that both the sender and the receiver have MIME-compliant tools. Many mail readers include the ability to install MIME-compliant tools (Eudora, a common mail program, has these tools built in), however, some of these mail programs (examples would be ELM and PINE), if not fully configured by the system administrators, will display *nothing* when they encounter a MIME-encoded message. This is a worst-case scenario. Many times the mail program will open the message, but the special characters will be garbled.

And then there's a third factor to be considered. Even if the mail program has no problem with MIME encoding and will process the message just fine, the special character may *still* not display correctly. This has to do with the terminal emulation program used by those who are directly connected to computer networks and using a text-based interface. The most common terminal emulations across many, many operating systems and networks are VT-100 and VT-101 (named after simple VAX Terminals manufactured by Digital Equipment Corp for their mainframe systems back in the '70s). These emulations are based on US-ASCII. DEC came out with a more advanced terminal in the '80s, known as the VT-220 (and later the VT-320), which included the extended ISO character set, and this emulation is also widely available, but much more complex for the user to configure.

Out of any given set of subscribers to a mailing list or newsgroup, I wouldn't be surprised if far more than half aren't set up with just the right combination of mail/news reader, MIME-compliant tools, and terminal emulation or display system to properly view a umlauted "o". Integrated tools such as Netscape with its built-in news reader and mail reader, along with a suite of tools, is making major inroads in this area, but there is far to go yet. So, unless you want to confuse or outright exclude a large portion of your readers, you have to ensure that messages are viewable by those using plain mail readers with no MIME tools and VT-100 terminal emulation. Hence, no extended characters or special formatting in mailing lsits or newsgroups for the foreseeable future.