I linked well-order and well-quasi-order. I do wonder whether there are constructions that assume an index set needs to be well-ordered that could use the weaker notion. I don’t know the consistency of asking that every set admits a well-quasi-order, but surely it’s weaker than full AC.

]]>At *well-order* I found it non-trivial to extract the usual classical definition. (Before I reached the second-but-last bullet item that states it, I was following a dozen links to other kinds of orderings and their properties).

For the sake of the reader, I have now given the usual definiton the first paragraph in the Definition section (I didn’t remove it from that bullet list).

But I think some expert might still make that entry a bit user-friendly.

]]>Good; I added redirects such as partial map.

]]>I made some edits at well-order. I am removing a query box, having duly extracted some punchlines. These edits also forced an edit to partial function, where I added the generalization to partial maps in any category with pullbacks.

]]>+– {: .query} This need not exist; in particular, $S_a$ may be empty. What do we really want to say here? (We could talk about the successor

ofa well-ordered set.) —Toby Mike: Yeah, or we could say that successor is a partial function. One definition of a limit ordinal is one on which successor is totally defined. =–