Your best friend for file transfer.Fetch
Many computers, including the Macintosh, can display special characters that are important for communicating in Roman languages besides English (some examples: é ö å ç), as well as characters for non-Roman languages, such as Japanese or Hebrew. Unfortunately, different computers use different codes to refer to these characters. So, unless special translation is performed, a filename that looks fine on one computer may be a garbled mess on another.
The Preferred encoding preference in the Miscellaneous preferences pane addresses this problem. This preference tells Fetch which character set it should assume servers are using, so that it can display filenames and file contents correctly on your Macintosh. If servers you regularly connect to show the filenames as gibberish or not in the characters you expect, you should try a different encoding.
For example, if you connect to a server that has filenames in Japanese, and your preferred encoding is Western (ISO Latin 1), the filenames will be displayed in Roman characters because the Western (ISO Latin 1) encoding tells Fetch to interpret text as Roman characters instead of Japanese. This won't be too useful if you're trying to read them in Japanese. If you change your preferred encoding to Japanese (ISO-2022-JP), you'll see the Japanese characters. However, the Japanese encoding is probably not the best for viewing a server with filenames all in Roman characters. This is an extreme example; for certain servers a different Western encoding may be better for displaying accented characters than the default Western encoding.
The Preferred encoding preference affects the display of item names in transfer window file lists, and the names of files after they've been downloaded to your Macintosh (or names of files you upload to the server). However, it does not affect the contents of files at all.
If a file list contains characters that cannot be displayed using your preferred encoding, you'll receive an error and Fetch will stop trying to display the file list and disconnect from the server. To solve this problem, change your preferred encoding.
In addition to the character encoding of a file's name, file contents are also encoded when they contain non-English characters. Some files contain information about the character encoding used for their contents (such as in a Unicode BOM or an HTML meta element). This encoding becomes important when you are editing a text file in an external editor, such as BBEdit or TextMate.
Most editors automatically detect files that contain information about their own encoding, and handle them correctly. When a file does not contain information about its own encoding, the editor will sometimes use the wrong character encoding and the file will appear as gibberish.
If you are using BBEdit (9.2 or later) or TextWrangler (3.0 or later) to edit files from Fetch, then Fetch will pass along your preferred encoding when you edit a text file in the external editor. If the file lacks encoding information, rather than displaying it in a random (and probably wrong) encoding, BBEdit and TextWrangler will display it using your preferred encoding from Fetch. Since file names and file contents are often encoded the same way, this makes it more likely that you will see your text instead of gibberish.
If all that fails and you still see gibberish, you can use BBEdit's or TextWrangler's Reopen Using Encoding command (in the File menu) to display the file in the correct encoding.
Preferred encoding is an overall preference and cannot be specified for individual servers. If you have a particular server that requires a different encoding from your usual setting, you can change the preferred encoding temporarily to work with that server, and then set it back once you are done with that connection.
Western (ISO Latin 1) is the default preferred encoding because it is the most common and compatible character set. While filenames may appear garbled, they will at least be displayed. (ISO Latin 1 is the ISO 8859-1 character set.)
Other encodings of note are:
Unicode (UTF-8) - Use this encoding when connecting to Mac OS X, Java, or newer Windows servers.
Western (Mac OS Roman) - Use this encoding when connecting to servers based on Mac OS 9 or earlier.
Western (Windows Latin 1) - Use this encoding for connecting to older Windows servers.
Fetch may do additional translation of filenames depending on the setting of the Encode and decode file and folder names preference. When that preference is checked, in addition to translating to/from the preferred encoding, Fetch will encode and decode item names according to the AppleSingle 7-Bit ASCII standard: slash (/), percent (%), null, and all 8-bit extended ASCII characters will be encoded as a three-character sequence of the form %xx, where xx is the character's hexadecimal value. This is useful for preserving special characters in names on servers that don't support any sort of special characters.