Another trick useful for linked lists is to create dummy head and tail elements of exactly the same type as the real elements. This way you completely avoid any special cases when adding or removing items from the ends of the list because you are inserting *between* two existing elements (one of which is the dummy).

This is supposed to be an introductory course for deer-in-the-headlight folks whose only prior experience is, you know, last year’s C++ course.

Anyhow, yeah, once I drew a little diagram and got the stars and ampersands in the right places, it worked like a charm. But still, it would depend who else was seeing the code. Me in six months would probably still get it. But would co-worker X get it?

BTW, if you’re even in Waterloo before September, Dan, you need to stop by for dinner so you can get USB mice working on my Gentoo…

I guess you engineers are probably working in C as opposed to C++? In C++, if you aren’t worried at all about performance, you can write some very tight binary tree code using classes and recursive calls (I seem to recall one that actually stored extra leaf nodes that contain no data, but simplified the cases, allowing very short methods.)

The only reason to be using C these days is if you’re really performance paranoid, in which case you ought to be pretty comfortable with strange pointer operations.