Comment 6 for bug 1277731

I used ALIAS FOR because if you change parameter names then you have to DROP (rather than CREATE OR REPLACE) the function. I try to avoid that when possible because we /can/ call functions with parameter names, and if (some time in the future) Postgres gets dependency tracking features for the guts of functions, the function signature will surely be the first thing it checks -- meaning a "Can't replace. HINT: drop cascade" type message.

The second is hand-wavy, and perhaps it's unwarranted? If you or others think so, I would be perfectly fine with an implementation that just changes the parameter name.