People often object that using a full-blown HTML/XML parser on "just a simple string" is overkill: it's "too much code". The reply to this is that a "simple string" all too often becomes complicated (*ML is, after all, a complicated spec), and then the overhead of maintaining a regex-based solution can explode. Do you know of a tutorial or discussion on this or any site along the lines of Dominus's Why it's stupid to `use a variable as a variable name' that addresses "Why It's Stupid to Parse HTML/XML With Regexes"?