I'm assuming you know what you're doing, but it's perhaps worth
pointing
out that when (per the instructions on http://www.wikipedia.com/wiki.cgi?Wiki_ASCII_codes_page ) I hold down alt key and then, using the number pad, type the number
of
the character in question, I do indeed get the character in question. I'm fairly sure it works the same way for many other people, too.
It gets complicated, because the Alt-Number trick gives different results whether you use 3 or 4 digits. Before Windows 1.0, MS-DOS used national code pages, so typing Alt-151, for example, would produce code 151, which in the standard code page of the time was u- grave. By the time Windows came around, ISO-8859-1 had become a much more prominent standard, and it was decided that WIndows fonts would use that encoding. But since people were used to Alt-XXX doing certain things, Windows grabs those and translates them into the ISO equivalent of what used to be at that position, so when you type Alt- 151, you really get character 249, u-grave (the ISO code). This is what's on the ASCII codes page, and why it works for most Windows/ISO machines, but it's misleading. Your chart shows "151" next to what is actually character 249 (look the file in hex if you doubt this).
Now, since ISO-8859 leaves characters 127-159 explicitly undefined and not usable for printable characters, Windows decided to put some stuff there that's not in the ISO code, such as curly quotes and real dashes. Character 151 is, in fact, an em dash. Type Alt-0151 and you'll see. But since it's not defined in ISO, it won't appear in X fonts, and so you won't see that dash on a Linux box. The Mac uses another set entirely, so 151 is a o-acute and 249 is breve accent.
To be in strict compliance with *ML standards, you can either specify which character set is used in you HTML DOCTYPE header, or you can just use plain old 7-bit ASCII and encode the specials with named character entity ("&") references. Numerical references are theoretically standard as well, but officially they reference into 16- bit Unicode codes (the first 256 of which are the same as ISO-8859- 1), and no browser I know will correctly display an em dash, for example, whose code is in the 2000s, but many will display the "mdash" named reference correctly, even displaying "--" if the local font doesn't have it.
--------------------------------------