Weird. What is the point of have sequential serial numbers if they don't do them by date? Do the ones that don't pass QC sit in a queue to be fixed and then get the date stamp when they are repaired? (But if that was the case, the date stamp would be even later...)
I cannot come up with a reason why the serial numbers would be out of synch the way this one appears to be.
While we don't have an official reason for this, it is prevelent among many manufacturers of that era. Take a look at the 5200 serial numbers, different plants, some with six digits some with seven. With the 1200XL there were two plants making them, and its very likely that there are multiple assembly lines per plant and each had their own labels and inked serial number ranges. Just consider this scenario (completely made up of course): You have four assembly lines, the only way to manually keep track of where a serial number is is by assigning each line a range of numbers for a particular shift/week/run. So, say 1000-1999 for line 1, 2000-2999 for line 2 etc.. Eventually line one will complete and need 5000-5999 before the other lines complete. Say some broke down, or needed to halt the line to work on something, while the others kept going.
The part I always wondered about was why there were differences in the number of digits from the same plant.
83S DA 005333 123 neopeius USA 005333 1983 12 21-Mar-83 27-Mar-83
83S DA 48017 123 rockdoc2010 USA 48017 1983 12 21-Mar-83 27-Mar-83
Same week. 5 vs 6 digits.
Also, I'm not exactly sure the serials were stamped when the stickers were applied. The date stamp certainly was but the serial numbers almost look like Atari ordered up a bunch of spools from a printer. Sometimes they got various versions.