Left-To-Right Parsing During Wiki->HTML Conversion
Currently the system uses regular expression matching, which can have somewhat surprising results sometimes, such as if you mismatch '' with '''. It would make more sense if pages were parsed into tokens from left to right.