Below are goodies that could not be included in the book due to space constraints. Enjoy.
After World War II the Bell System embarked on a quest to automate long-distance dialing. To do this, they needed a few key things: better long-distance transmission systems, improved switching and routing intelligence, workable long-distance signaling, and an automated billing system.
In Exploding The Phone I didn't have space to do justice to the automated billing piece of the puzzle. Since the blue box and the black box defeated this billing system outright, the details of how it worked weren't so terribly important to the book's main story. But it's too bad, because the automated billing system was way cool, in an audacious retro-technology kind of way. So, herewith, a paean to the miracle of automatic message accounting.
[Short attention span? tl; dr? Watch the video down the page.]
In the good old days before Skynet, when human beings still did things themselves — that is, up until the late 1930s — toll operators manually handled long-distance billing. The operators wrote down the details of the calls on pieces of paper called toll tickets. These tickets were later collated and processed by other humans in order to bill customers.
AT&T knew that this wasn't going to work forever. So starting around 1938 its engineers began developing something called "automatic ticketing." A small mechanical printer that could be attached to each outgoing long-distance truck line, each automatic ticketing unit printed up slips of paper containing the same information that the toll operator would have manually written on her ticket. It's just that it did it automatically, with no operator intervention.
For automated ticketing to work, the billing system had to learn a new trick: it had to be able to figure out — automatically and quickly — what number you were calling from. It sounds crazy today but, up until the advent of automated ticketing, the central office equipment didn't really know who you were or which line you were calling from. So, when you made a long-distance call the operator would ask you for the number you were calling, and then also ask you for your number, so she could bill the call to you! This was called operator number identification or ONI. In contrast, automated ticketing meant that Bell Labs had to come up with something called automatic number identification or ANI - basically a bunch of circuitry added to a central office to allow the equipment to determine who was calling whom.
A key thing about the automated ticketing system was that its output was designed to be read by a human being, just like an operator's toll ticket. That is, while it relieved the operator of the burden of writing call details down on a piece of paper, a human still had to process its output to bill the customer for the call. An improvement, but not one that scales very well if you expect to be handling huge numbers of long-distance calls in the future, like AT&T hoped. In addition, automatic ticketing required one ticketing unit per outgoing long-distance trunk, an expensive proposition.
In 1948 AT&T installed the first deployment automatic message accounting, or AMA, its new long-distance billing system. Unlike automatic ticketing, AMA used one billing recorder unit per 100 trunk lines. More importantly, the billing recorder unit didn't produce printed output designed for human consumption like the automated ticketing units did. Rather, the recorder — called a perforator — saved its output as perforations in a 3" wide roll of paper tape. Billing data were punched in the paper tape chronologically, with a typical phone call having three entries: (1) the initial entry, consisting of start time, calling number, called number, and trunk number; (2) the answering entry, showing the trunk number and time the call was answered; and (3) the hang-up entry, showing the trunk number and the time the call was terminated.
At this point a typical history of AMA would say something like this: "These paper tapes were delivered to an AMA accounting center where they were processed to generate your long-distance bill."
True. But that's a bit like saying, "The patient was delivered to a hospital where a heart and brain transplant were performed" — it may be technically correct but it covers up a lot of important and messy and stressful details.
Imagine yourself back in 1948 at one of these AMA accounting centers. Every day trucks show up carrying rolls of paper tapes from different central offices. Entries on these tapes are mixed up, in the sense that a given tape might track 100 outgoing trunk lines, and the billing entries get made as they happen on the tape, so first there might be an entry for trunk #1, and then one for #37, and then one for #83, and then one for #1 again, etc. You need to build a machine, or a set of machines, that will read these tapes in, sort them, collate them, and produce a telephone bill. It has to be accurate, it has to be reliable, and it has to be timely — after all, more truckloads of tapes are showing up tomorrow. And, much like the 4A crossbar switch, you need to do this miracle without the benefit of digital computers or transistors, neither of which have been invented yet.
As usual, the men and women of the Bell System were equal to the challenge. The video below gives an idea of the kind of shenanigans required to make all this happen, from the hole-punching to the final output.
One problem with the AMA system was that it was still pretty expensive for telephone central offices that didn't handle a lot of long-distance calls. That is, you don't want to go to the trouble and expense of deploying these fancy perforator units if they only get used a few times a day. Additionally, it wasn't economical to set up AMA for step-by-step or panel offices, which represented a large fraction of the COs out there. What to do?
The solution was something called centralized automatic message accounting or CAMA, deployed in 1953. (The original AMA became known as local automatic message accounting, or LAMA. There is a joke in here somewhere about "CAMA LAMA Ding Dong" but it eludes me.) With CAMA, the AMA equipment such as the perforator was located at a tandem central office, which would serve as a billing point for lots of other central offices. When a long-distance call was made under CAMA, the originating central office where the call was being made from would forward the call to the CAMA tandem office, using multifrequency (MF) pulsing to communicate the number being called, just like usual. In addition, it would also use MF to send the calling number to the CAMA office for billing purposes. The AMA equipment at the CAMA office would then punch this information on paper tape and track supervision of the call going forward. In the event that the originating central office couldn't determine the calling number, provision was made to have these unidentified calls routed to special operators whose job was to ask what number you were calling from.
As time and technology marched on, AMA evolved. In particular, magnetic tapes made several assaults on AMA's paper tapes, starting in 1966 and continuing on through the 1970s. The #1 ESS also used magnetic tape for recording AMA information. A computerized LAMA system, LAMA-C, using PDP-11/30 minicomputers, was deployed in 1975, and a computerized CAMA system, CAMA-C, using IBM System 7 computers, was deployed in 1974. As mentioned in Exploding The Phone, CAMA-C also had some toll-fraud detection features.
Today, though the paper tape is all gone, AMA lives on: the standard format for exchanging telephone billing information is called the "Billing Automatic Message Accounting Format."
Amos E. Joel, Jr., A History of Engineering and Science in the Bell System: Switching Technology (1925-1975), pp. 130-148 and p. 379.
Pacific Telephone, Survey of Telephone Switching, 1956.
The video above about AMA is an excerpt from "Nation At Your Fingertips," a 1951 AT&T movie, available at the Internet Archive courtesy of the Prelinger Archives. It's well worth watching.