Since this is practically my line of work, I thought I could be of help.
The character set, used by the webbrowser to determine which character-shape ("glyph") is used for which character-code, is currently left unspecified on story pages. This makes webbrowsers activate their "auto-detection" logic, which looks at the actual character-codes it encounters and guesses which character set is most likely to display the page as the author intended.
However, as you might've noticed it doesn't always guess right, and that's where explicitly setting the character set to be used comes in. The technically nicest way is to add this as a HTTP header, but that might not be possible for you. Another way is to add the following meta tag to the head section of the html document:
<meta http-equiv="Content-Type" content="text/html; charset=utf-8">
This indicates to the webbrowser that the universal UTF-8 character set, which allows for encoding of all existing character-shapes, is to be used for the document. If you have a document that uses another character set, figure out its name and fill it in in the place of "utf-8" in the above string. For example, for a japanese document encoded as Shift JIS you should specify:
<meta http-equiv="Content-Type" content="text/html; charset=Shift_JIS">
However, the easiest is probably to use UTF-8 on every document and have your your text editor save everything as UTF-8. Every self-respecting text editor should have functionality for that, since even Notepad can do that.
Oh yeah, and one more nitpick: the main page (
http://www.eunuch.org/main/)'s character set is declared to be iso-8859-1, which makes it mangle certain non-english story titles. Maybe set this to UTF-8 as well?
Anyway, hope this helps.