The Firefox data source for Zeitgeist causes Firefox history data to be stored in two separate places: Firefox's history database, and the Zeitgeist log. This could cause a few issues:
1. Firefox history takes more disk space on a computer with the Zeitgeist extension
2. Firefox and Zeitgeist will likely have separate behaviors and rules for clearing internet history (deleting after a set number of days, "Forget About This Site", Private Browsing, Clear Recent History)
3. Two different user interfaces for working with history
4. History synced over Firefox Sync not imported when the user may wish it to be imported.
5. History generated before the user installs the Zeitgeist extension is never logged by Zeitgeist.
6. Issues #2, #4, and/or #5 could cause the data in Zeitgeist to be different from the data in Firefox. This could mean a user searches the Zeitgeist history for something that's only found in Firefox history or vice-versa. It could also mean the same query could have largely-different search results in Sezen and the Firefox "Awesomebar" even though the two have very similar functionality (Site A could be near the top on one but Site B almost completely hides Site A results on the other).
The data source could, instead, provide Zeitgeist with an interface to Firefox history, even telling Zeitgeist where the history database is found, and recording (in the Zeitgeist log) extra data specific to Zeitgeist that Firefox doesn't capture.
This probably affects the Chrome data source as well.
The data being duplicated in Zeitgeist's database is a design decision that is unlikely to change anytime soon.
Stuff like "Forget about this site", Private Browsing and possibly synced history should be handled by the extension (it's a bug if it doesn't) and 5. could be considered as a one-time option in the extension (actually an older version of the extension supported that).
Deleting after a certain number of names is not something planned for Zeitgeist, since the goal is to get as much data as possible (so that relevancy algorithms can be applied to the data). If Firefox removes old history, then that's one more reason to copy it into our database.