First, SC reads characters in groups of four, so if your string is longer than that, you'll need multiple EUDs to detect the entirety of it. Take the string "Hello World" as an example. The first step is to split the text into groups of four:



The last group ends with only three characters, but this is fine. For this string, you'll need three EUD conditions to detect the exact phrase. When SC reads the text, it reads the group of four characters in reverse order. The next step is to reverse the order of the letters in each group:



Now, SC doesn't actually store the strings as values, so it represents each string as a number. To convert the text into a value, we must look up each letter in a character map (Windows computers should already have this) and use its hexadecimal value. When we've done that for all letters, we should end up with a hexadecimal number:



For the first group, l's value is 6C, e's value is 65, and H's value is 48. When put all together, it makes a large hexadecimal number (6C 6C 65 48 -> 6C6C6548). (It should be noted that the hex value of a space is 20, but it doesn't show up on some character maps). Finally, we need to convert these values to decimal, as that is what SC reads. Just use a calculator or a hex to decimal converter:



Now you finally have the values you need to make an EUD condition for detecting the text "Hello World" from some address. The trigger will end up looking a little like this:



Notice that for whichever starting address value you're using, the next group of four characters is at the next address value (+1). Above, "Hell" is read at 186879, "o Wo" is read at 186880, and "rld" is read at 186881.

Note that the above did not cover the null terminating character, player chat problems, nor the issue of values not resetting (they are just overridden as needed). Getting a functional system using chat is fairly complicated. Refer to the EUDDB entry for general information.