Incremential Backup & renaming/moving files

Question

Hi!
I guess usually, when an incremental backup is done, a moved or renamed file will be treated exactly as two different files - one that has been deleted somewhere, and a new one somewhere else, resulting in an "additional copy" of the file and the loss of the traffic & disk-space?
If so: Is there a solution, that actually just links the moved files trough different backup stages instead instead of initiating another "transfer/copy" and keeping different copies of the same file? Or something like that?
Or in general: Which backup solution might be a good one to get into, if moving/renaming big files is common? What man-pages will I most likely want to read and what words to google to be able to make the right "decision"?
Thanks!

Answer 1

Mmmh... something like...
1) make list of "deleted on source" and "new on source" files just like I guess rsync etc. do too at some point
1a) optionally drop files with filesize smaller than X from list
2) look if a distinct "new" (source) file has same file-size as a "deleted" (old backup) one / match up
3) compare md5 for files that match (again source <=> old backup)
4) link finds from old backup to new increment backup instead of copying from source
Am I missing something? Maybe I just can't wrap my brain around that one to see the real problem...