Also this statement is quite awkward:

IRC messages are short because they are simple ASCII text strings in a specific format.

First of all, encoding issues are a major pain point in IRC. Luckily most people use UTF-8 now, but the transition from legacy 8-bit encodings was not easy. Considering the history of IRC, I don't believe that IRC ever used plain ASCII.

Secondly, what is XML but "text strings in a specific format"? That's incredibly vague statement to begin with.