In Win98 or WinME the original setup for character coding is limited to ANSI values, and 8-bit characters. For the few characters not on your keyboard, as several have noted, you can hold down the Alt key and type the decimal character number on the Num Pad to insert the character. (Num Lock must be on.) It was not anticipated that character numbers past 0256 (= 2⁸) would be needed.

The Alt-NumPad method retrieves character codes from font "pages" loaded according to the "language" selected by the user, and "extended fonts" have been released that allow you to use some higher (decimal) character numbers, but full Unicode character sets aren't easily used. The characters available in fonts that have been released are "spotty" in coverage, and in some cases don't present themselves correctly when viewed in non-Windows systems.

An example of a "misfit" character is the "euro" sign, which Microsoft has added to font tables as decimal character 0128. US users, who won't have a "euro" key on their keyboards, can enter it in almost any Windows program as Alt-0128 = €. (The char value, decimal 0128 or Hex 0080 is unasigned in the Unicode standard.)

(This same "font page" usage is probably responsible for the few remaining "misconversions" when Mac users try to work with Windows users in critical typesetting applications, and is presumed to be the cause of all of the funny glyphs in Amos's cut-n-past posts in the "Random Traces" thread.)

Right-to-left character strings are available to Win98/WinME users only if they have a "regionalized OS" and an appropriate "regionalized keyboard" for a language that uses right-to-left characters. According to the latest Microsoft "bible" on Internationalizing, the right-to-left feature cannot be added to a Win98/WinME version that wasn't so equipped at release.

There are, of course, utilities that will invert a string of characters and print them left-to-right-in-reverse-order-right-justified to give the appearance of right to left char strings. While these give the "correct" appearance for most users, they generally crash when viewed by people equipped for the languages where this display is appropriate (although I've seen only limited comment on this).

Win2K and later (esp including WinXP) at least theoretically have full range Unicode character capability, if appropriate fonts that contain the characters you want are loaded. For some language features you do need to install "international character extensions," but the required additions have been available only as options in Server OS packages which makes them unavailable to most users who don't work for a big company that chooses to make them available.

Recent versions of most browsers, and recent Office programs, can look for any font installed on your computer that contains an isolated coded Unicode character (in the html code in the "&#xHHHH;" format) to display "foreign" characters. The html specification of course supports full range Unicode char values; but individual servers sometimes don't accomodate full range features; and your own computer has to have a font that knows how to display the particular char that's coded.

Microsoft Word versions since Word 2002 allow you to type the hex code number for any Unicode character and with the cursor immediately to the right of the code, Alt-x will convert it to the glyph if the char is in a font you have. Similarly, any "funny character" in a Word document (or that you can copy and paste into Word) can be "read" by placing the cursor immediately to its right, hitting the same Alt-x, and the Unicode hex value for the character will be displayed in place of the glyph.

John