Should have a look at the way \'e are managed in Citezen

[Ducasse]

When dealing with bib citation

visitSectionTitle: aCZString
	outputStream
		nextPutAll: '<h4>';
		nextPutAll: 	(self treatString: aCZString expanded);
		nextPutAll: '</h4>'.
	self
		line;
		line

treatString: aString
	"There are two aspects that should be covered: 
		- handling tex encodings ie. \' or \& or \'{o} to be turned into the correct glyph
		- strange characters that are meta characters in different outputs."

	^ self treatSignsInString: (self treatTexEncoding: aString)

treatSignsInString: aString
	"By default do nothing because subclasses may apply way to handle strange characters such as _ \ and others. For example for latex _ is a special character for pillar too."

	^ aString


HTML >> treatTexEncoding: aString
	"we want the HTML encodings to tex encoded characters"

	^ CZAccentuateCharactersParser parse: aString