Actually it was shipped. “Microsoft disabled the AARD code for the final release of Windows 3.1, but did not remove it, so that it could have become reactivated later by the change of a single byte in an installed system.” [1]
[1] Schulman, Andrew; Brown, Ralf D.; Maxey, David; Michels, Raymond J.; Kyle, Jim (1994) [November 1993]. Undocumented DOS: A programmer's guide to reserved MS-DOS functions and data structures - expanded to include MS-DOS 6, Novell DOS and Windows 3.1 (2 ed.). Addison Wesley. ISBN 0-201-63287-X.
Only if the application has code to change that byte. “It could be changed by a patch” doesn’t count as a condition under which it could be executed, because a patch can make any change it wants to. You wouldn’t say “there’s a condition under which Windows will wipe your hard drive and every visible network share” just because someone can write code to do that.
As documented in the excellent book "Undocumented DOS", whose red cover I can still picture. (Noted in the article, but it was a good book that I still recall fondly.)
No. It's on the bookshelf next to Petzold, and below the other Petzold book, that I suspect far fewer have read, on the shelf above. On the other side of the Lower Petzold is Undocumented Windows.
If you deduced from this that the shelves were divided into Windows and OS/2, the sad truth is that the lower shelf is slightly too high, and the books are divided by height differences that amount to millimetres in some cases. Lower Petzold fits; Upper Petzold does not. (-:
IIRC, the author of "Undocumented DOS" discovered and disclosed this feature during the beta, before release.
It seems at least plausible, if not probable, that Microsoft disabled the xor encrypted bogus error message generating code in the release version as a result of this disclosure.