From an email message posted by Joe Buckley to Baen's Snerkers Only dated May 23, 2009:

Detection of upward hyper translations

    Actually, upward transits into hyperspace are detectable from the hyper-space end, but not from the normal-space end. I phrased myself poorly in The Short Victorious War when I described what the Manties saw when Alexander hypered out. The "sparkle" wasn't really a hyper hyper footprint; it was the effect of a hyper transit by a ship underway under impeller drive. So, yes, it was a hyper footprint, but it didn't occur simply because a ship crossed a hyper wall. It happened because the ship in question was underway under impeller drive when it crossed the wall. That's never been a factor in my need to describe what's going on before because everyone's been using impeller drive when they made their transits. In this case, they wouldn't be. Further, the only reason the Manties were able to pick up even what they picked up was because they'd already localized Alexander on their sensors. They knew exactly where she was, despite her best efforts of stealth, and (as you say) they were literally right on top of her. (This is how I've always visualized the process working, but I probably should have made it more explicit. It never struck me as being necessary. Or, to be more honest, it never even occurred to me to wonder about whether or not it might be necessary.) The hyper transits which haven't been being detected over the last several books have been made by ships under stealth, which hadn't previously been localized, and which were either not using their impellers at all or under absolutely minimum power when they [made] transit. Under those circumstances, hyper footprints from upward translations are effectively [indetectable], even by the kinds of arrays the Manties have in place to protect the home system.