And is also an indictment of the people who wrote the book.
I've profited greatly from information from books. But it's no substitute for real world experience.
In this case, not really, but I know what you mean.
The "book" in these cases was the ITU specifications, which really don't aim to document all of the real world odd-ball stuff that carriers and equipment manufacturers actually do in signaling, they just define the base signaling.
On one of the products, the interaction between in-band signaling and sending all of the same bits in all timeslots ( conference calling ) was the problem.
The short version is: If you would pick up a microphone and whistle into a conference bridge using in-band alarm signaling on its T1 circuits, you'd hit and hold a tone that would hit and hold the second most significant bit across all 24 channels of the T1 for long enough for the far-end to register that as a Yellow Alarm. Response to Yellow Alarm (at that time) was undefined in the ITU spec, but most carriers would instantly disconnect all DS0s and reset the span.
So basically... You couldn't use a non-ESF span for conference calls unless you could guarantee you wouldn't put 24 channels of the same conference call on the same span simultaneously, or you'd randomly lose 24 people from the conference when someone's voice triggered the right bits.
Actually you could lose multiple spans simultaneously.
Goldman Sachs and other large investment houses didn't like it too much when people were falling off "important" conference calls in multiples of 24. Haha.
Temporary solution: Take one of the channels off-hook on every span and disable using it (waste of a channel, but cheaper than ticking off financial customers) so you'd never get all 24 second most significant bits set to 1 from sounds in the conference call.
Long term solutions: Only use out of band signaling on T1s used for conference call bridges. You couldn't guarantee you still wouldn't end up with 24 inbound callers for the same call on a span, so algorithms to avoid sharing the span would only work on out-dialing bridges only.
I was the 20 year old kid who had to prove this was happening... We really ****ed off Sprint that day, triggering hundreds of alarms on their Cheyenne, WY switch... Haha. They called our NOC and asked WTF we were doing! LOL. I'd found that whistling would trigger it way faster than singing or talking... Significant bits being what they are in a u-Law digitization scheme...
High pitch and loud... Whistling worked great. It also explained why our largest complaint stack came from a conference call who had a squeaky high pitched girl for the chairperson...