Printed vs. actual run time differences

I've been noticing an increasing number of database notations lately indicating as little as :01 and :02 differences between actual and printed run times. It seems until very recently we had loosely held a "three second rule" when it came to noting such run time differences. Is this no longer in effect? Should all printed run times that are not identical the actual run times be noted in the database? Personally, I favor the "three second rule" but I'd be interested to hear other thoughts.

I also agree, Todd and Gordon, and that has always been my personal "reporting threshold". But in the case of "Do Ya Think I'm Sexy?", for example, I had a promo 45 where the "long" 45 side had a :04 listed/actual time difference (5:21/5:25). The "short" 45 side only had a one second time difference, though (4:16/4:15). I would NOT have posted the :01 "short" side difference on its own. But since the :04 "long" side timing difference WAS long enough to report, I simply chose to also include the :01 "short" side difference as well, as a courtesy, simply FYI, since I already had to post info about the other side of it, anyway. It just doesn't feel right for me to only report about "one half" of a 45! Also, I've been burned before, by having to "re-pull" the 45, due to a later follow-up question. I'd just rather "put the song to bed", once and for all, by including ALL of its specific details at the same time.