10

A function is needed to remove duplicates in your addressbook.

I understand that it should be just run by Contacts → Remove duplications or something like that.

But how do you define "duplication"? KAddressBook is capable of pretty much informations of users. Typically it would be name and e-mail address, though. Are contacts with the same e-mail addresses, but different names, "duplication"? Are contacts with the same names, but different e-mail addresses "duplication"? If two contacts have some informations at common, but some not, is this "duplication"? What is part of data is common, but other part simply differs (i.e. I have two contacts with same name and e-mail address, but different work phone numbers)? Should that function work for certain address book, or for all at once? Or maybe should it be configurable by user?
Perhaps list of questions could be continued.

I do not mean to bash around, I just think that your idea is too ambiguous right now. There should be at least proposal of how that function should work, if this idea is to be implemented by developers.

duplicates should be merged
1.
first there should be checked, if two contacts are exactly the same (for example cause you imported them from 2 other addressbooks)
then one can be deleted.
2.
check if one contact A has all data from another contact B, but the other has some extra data.
then the contact A can be deleted
3.
two contacts have the same name or email but slightly different other data,
then they have to be merged with user-interaction:
both contacts should be shown next to each other left and right with an option to select for each field, which side to keep.

For two contacts with the same name (first and last) but different e-mail addresses, they should be (per user confirmation) merged to the same contact with both e-mail addresses (a given person can have multiple e-mails, multiple phone numbers, multiple IM accounts, etc. We should strive for a 1-1 relationship between contacts and people (or organizations))
for two contacts with the same e-mail address but different name, they should be presented to the user to either leave as separate contacts (decision remembered so that it doesn't ask every time the merge process runs) or to merge them under the same name (with options to pick either of the names or fill in a new (more appropriate) name)

The user can have a menu to merge the contact with another. Then, he can be asked on the fly for avery data if he wants to keep or to delete of the source contact, which after the merge will be deleted...