> I think what he's talking about is whether or not it returns what was > just added to the document or the original element. Another example > would be whether or not .clone() return the cloned elements in the > jQuery set or the original elements. Those sorts of things aren't > listed.
Good point, and sometimes I have to pause when using a method I haven't used before. However, I think most jQuery methods fall into just three categories. (I use $ to mean "the elements currently selected by this jQuery object.") 1) Standard methods -- most not listed below Do not change $ and return $ for chaining. If the method creates nodes, they are not part of $. If the method removes nodes that are part of $, the references in $ remain. (That is, they have been removed from the document but are still in $ so you could put them back using a method like appendTo further down the chain.) 2) Filtering methods -- filter, add, not, parents, etc. Push $ on an internal stack and return the (new) filtered $ for chaining. The original $ is available by using .end() in the chain to pop the stack. 3) Set/Get methods -- height, width, css, etc. Called with N arguments (usually none), return the value of the property for the *first* node (element 0) in the jQuery object. Called with N+1 arguments, set the value of the property for each node in $ using the final argument. Return the unchanged $ for chaining. Can anyone think of exceptions? Perhaps the doc could define (Standard, Filtering, SetGet) and we can just document the types of return-value behavior using those terms. _______________________________________________ jQuery mailing list discuss@jquery.com http://jquery.com/discuss/