df.world-site.xml:world_site mapping doubts

[PatrikLundell]

I've made an attempt to make sense of any of the anon_3 - anon_24 fields (that's what they're called in the version of the file downloaded an hour ago: may well be shifted at the time this is read [referencing anons is a pain as they change numbers over time]) at the end of the struct (I believe they were added recently).
What I see is occasional patterns of something that definitely isn't a pointer (value too large, typically or always with 4 in the lower half), followed by what looks like 5 valid pointer values, and then followed by another non pointer (value too small). These patters appear on different sites at different loadings of the same save (at least when DF is restarted in between), and the pattern shifts around among the anon fields. I've seen the non pointers at anon 6/12, 8/14, 10/16, 12/18, 14/20, 16/22, as well as a pattern starts at 18 and 20 (with anon_23 = 0, and anon24 = nil). To clarify: 6/12 means a non pointer at anon_6, 5 valid pointers, and another non pointer at anon12.

Furthermore, it seems the pointer values are all the same in all the instances, with number 3 and 4 being the same. The starting non pointer might always be the same (not sure), while the end one varies. I believe the pointer sets get different addresses on different DF invocations (which would be reasonable).

No non zero values have been seen in anon_3 - 5, but that may just be a matter of a sample set that's too small.

Edit: Encountered invalid pointers for anon_3 - anon_7 in a save. Won't bother to try to figure out if the anon numbering has changed since the original post.