Hrvoje> But can't you build a `weak-list-list' and Hrvoje> `set-weak-list-list' upon whatever interface you provide? Hrvoje> Why change the interface? ms> Because it exposes the internal representation: Since these ms> procedures are currently the *only* way to get at the list ms> elements, they need to be fast. This in turn pins down that ms> weak lists pretty much must be represented as a special header ms> around regular lists. And if they _aren't_ the only way to get at the list elements, they can be as slow as you like. "Unlikely" user code keeps working. Just remember to deprecate it so the byte-compiler will warn about it. ms> (Note that all of this work is presently happening on a ms> branch---we'd just like to discuss issues such as this one to ms> avoid having to revert them later.) Make reversion less likely, yes; avoid, no. I don't see why Hrvoje or Ben should be convinced without a live implementation to look at. I see no guarantee that they will be convinced when you put in in front of them. So you have to accept that risk. It seems to me that since the generic GC is at present in the blue sky/proof of concept stage, you really should make the implementation as clean and effective as possible. We can dirty it back up :-( for hysterical reasons later. -- University of Tsukuba Tennodai 1-1-1 Tsukuba 305-8573 JAPAN Institute of Policy and Planning Sciences Tel/fax: +81 (298) 53-5091 _________________ _________________ _________________ _________________ What are those straight lines for? "XEmacs rules."