For example, I use four "equation" environments: starred and unstarred, and including a "\begin{split}...\end{split}" construction or not. Currently, to have all four available for auto-complete, their names in the XML file are "equa", "equat", "equation" and "equation*", and I just remember which word corresponds to which variant. It would be much easier to be able to assign arbitrary text for a completion/substitution keyword and use, say, "eqn", "eqn*", "eqs" and "eqs*".
I did try to insert some escaped backspaces () in the XML file, but the parser complained, and I don't think this is allowed in XML 1.0 anyway. (Newlines (
) and tabs ( ) work just fine, though.)
A less effective but still viable solution might be to allow individual "own text module" entries to have shortcut keys assigned to them. But this almost defeats the purpose of auto-completion
