Wow, that's both sad and funny. A bit of a time capsule, though when the pressure is on is there ever a team that wouldn't have that kind of interaction?
In a couple of past projects I was involved with, email had been sent around among "team" members that by comparison makes those old MS messages look pretty tame. I plan not to leak potentially embarrassing emails, at least while some of the individual authors are still around, but keeping info safe requires an effort to make sure it really is secure.
The article does show the need to be careful when constructing things and make sure there are no leaks. Maybe that's more easily said than done.
Every time I read something about OS/2, I'm just waiting to see one of my old friends from IBM Boca Raton names pop in there. Man I loath of those days again. What a great bunch of people to work with. I loved it there and to this day LOVE OS/2.
Hell, old compilers? Even modern compilers don't do that. I once fixed a bug that occurred because someone was doing a memcmp() between two structures to check if they were the same, and it was returning false because some of the padding between variables differed between the two. This was in C++ in a modern version of Visual Studio.
That's not so unreasonable within the process' own heap. I was thinking more about padding inside .exe files on disk with the risk of redistributing that - either from nonzeroed drive slack or from random memory at build time (ms-dos/win95?)