The relationals allow you to specify the relative ordering of characters. For example, the following string expresses that 'a' is less than 'b' which is less than 'c':
"< a < b < c"
For the time being, just accept that a string should start with '<'. That rule will be both relaxed and explained later.
Many languages (including English) consider 'a' < 'A', but only as a tertiary difference. And such minor differences are not considered significant unless more important differences are found to be equal. For example consider the strings:
Since 'a' < 'A', then "aa" < "Aa". But "Aa" < "ab" because the difference between the second characters 'a' and 'b' is more important the difference between the first characters 'A' and 'a'. This type of relationship can be expressed in the collation rule with:
"< a, A < b, B < c, C"
This says that 'a' is less 'A' by a tertiary difference, and then 'b' and 'B' are greater than 'a' and 'A' by a primary difference (similarly for 'c' and 'C').
Accents are usually considered secondary differences. For example, lower case e with an acute accent might be considered to be greater than lower case e, but only by a secondary difference. This can be represented with a semicolon like:
"... < e, E ; é, É < ..."
Note that characters can be entered in hexadecimal or universal format. They can also be quoted with single quotes (for example 'a'). If it is ambiguous whether a character is a command or a text argument, adding quotes specifies that it is a text argument.
Characters not present in a rule are implicitly ordered after all characters that do appear in a rule.