[ Home ] [ Table of Contents ] [ About Lee Goeller ] [ Search ]

Voice Communication in Business Volume 2
Essays on telecommunications, 1981-2002

Of course, some people and most system designers do not know what a 1A2 key system is, or how it is used. Thus, as a follow-up to the small PBX article, I wrote this for the July-August, 1987, issue of Business Communications Review.

The exponential 2 in the title is a joke based on a series of ads being run at the time announcing the merge of Burroughs and Sperry into Unisys, exponentially more powerful than its component parts.

The Power Of 1A2
(Business Communications Review, 1987)

I must have gotten ten calls in the last two or three months asking for the minimum feature list a PBX must have to be competitive. Not one caller asked what features would be needed to blow the competition away...just what are the minimum features it takes to get by. Unfortunately, this is not an unreasonable attitude, considering the vastness of the feature lists of established PBXs, and the importance of initial cost to many who evaluate proposals. Under competition, you cannot afford to innovate and minimize cost at the same time, particularly if you're the new kid on the block.

But as the sophisticated know by now, lengthy feature lists tend to be paper tigers. Few people know how to use more than two or three features, and even fewer actually use them. So if we were to talk about the best of all possible worlds, how should we actually approach features to make a PBX valuable to the customer? Different kinds of features need different approaches, but the one I have worked out for station features seems to to be headed in the right direction. It is based on the following general idea:

If you have to train the station user,
the system is designed WRONG.

I can already hear the shrieks of dismay. Training is one of the most important aspects of putting in a new PBX. You HAVE to train the users to make full use of the incredible sophistication in their telephone system. Without training, anarchy will reign.

I have very little sympathy for such attitudes. To me, the purpose of a sophisticated system is to make sophisticated users unnecessary. To suggest what can be done, consider hunting. When my line is busy, an incoming call goes to my secretary automatically. Neither of us have to do anything...it just happens. And, at a more sophisticated level, we have automatic route selection: the user dials 9 plus the outside number. The PBX, if it is set up properly, decides whether or not that particular person can make the call at all, and if so, what facilities he or she is permitted to use. Then it places the call. The user need know nothing about networking, WATS lines or choosing the right access code for the proper facility to the desired destination. Just dial 9 and go.

Most of the station users working for my clients are NOT professional telephone callers. They are office workers, salesmen, administrators, accountants, engineers, managers, executives. Their purpose is to do whatever they are hired to do. The purpose of the telephone system is to help them do it. As unobtrusively as possible. They should not have to have instruction on how to use their phones, just as they no longer need detailed instructions on how to make long distance calls with ARS.

But how can a PBX with a thousand features possibly be easy to use? There is a simple approach. Just be sure its electronic telephone sets work like the 1A2 key sets of yesterday, and you are home free. The 1A2 key system evolved over 30 years to meet the needs of station users; study its design and operation and you'll see what users have voted for with their dollars for decades. Indeed, the amount of 1A2 still in use behind PBXs and, to an even greater extent, Centrex, shows that it has to be doing something right. In particular, 1A2 gives users access to several lines rather than one, and that is the secret of its power.

Station Features

Between 1975 and 1980, the makers of electronic PBXs, completely unable to compete with existing electromechanical systems on the basis of price, started developing features, mostly based on inexpensive memory in their control computers, to make their systems seem like more for the money. Then, as a result of this profusion of features, they offered the customer the opportunity to get rid of "all that expensive 1A2 key equipment," replacing multi-button sets with single-line (residential) telephones.

In the worst of such systems, the customer had to learn to differentiate between a great variety of call progress tones to figure out what the system was doing, and then had to remember dozens of "feature codes," keyed into the system after a "hook flash" to tell it what to do. In short, the user had to become a telegraph operator to make a phone call. And he or she couldn't do it without extensive training. And retraining. And retraining.

The logical way to correct this problem, now embraced by most PBX manufacturers, is to provide proprietary telephones which look sort of like 1A2 key sets, and SOMETIMES behave sort of like them. But all too often, they retain flash and feature code thinking. That is, they emphasize single button activation of single-line features rather than the multi-line access that came naturally with 1A2. Such phones have sometimes backfired on PBX manufacturers because the continuing decline of computer and memory costs has made it easy for independent manufacturers to add inexpensive repertory dialers to conventional phones. When these dialers are programmed to send a hook flash, detect dial tone, and then send a one, two or three digit feature code, they do just what the proprietary sets do and often at half the price. The repertory dial approach is not a bad one. In many instances, it is all that is needed. But there are a good many places where it works no better than the conventional single-line sets it was designed to replace.

Northern Telecom pioneered a suitable solution with the original SL-1 set in 1975 which could duplicate the functions of 1A2 while using more appropriate modern technology. AT&T followed shortly thereafter with electronic sets for the Dimension family, and Fujitsu, then American Telecom, made a line card to support an electronic key set from ITT. Most other electronic sets of that period, including the Rolm ETS 100 which could pick up three lines, put more emphasis on single button activation of single-line features. It was not until the 1980s that electronic sets designed to work with parent PBXs to emulate the multi-line approach of 1A2 became generally available; then, curiously enough, the ROLMphone series moved more toward multi-line pickup, while AT&T's electronic sets for its digital PBXs took a strong turn toward buttons primarily for feature activation on a single line.

A 1A2 Example

Let's consider an example of 1A2 operation. Suppose the boss is talking on line one and the secretary is on line two. A third call comes in on line three. The secretary puts line two on hold with the HOLD button, selects line three by depressing its button (or key, in telephone terminology), and greets the new caller. The call is for the boss. The secretary puts line three on hold, gets on the intercom, and signals the boss who hears a distinctive beep. The boss can now put line one on hold to answer the intercom, get the appropriate information from the secretary, and return to line one or take the new call on line three, simply by pushing the appropriate button. Just try to do that with a single-line 2500 type set! Or, for that matter, with some of the feature-activating proprietary sets presently available.

To make things easy for the station user, 1A2 buttons are illuminated, and the status of their lines is known instantly by just looking. If the line lamp is off, the line is idle. On steady, in use. Flashing one way, on hold, and flashing another way, ringing. All this sounds complicated when you read it, but most people can handle it, almost intuitively, without any training at all. To maintain that level of non-training is our basic objective with new station sets and features.

Two More Scenarios

The above scenario is usually referred to as "boss-secretary." Let’s look at a couple of others: principal and assistants, and what AT&T calls "car parts" but is actually much more general. In "car parts," we have one phone at the parts desk in an auto agency, and several other phones back in the parts room and perhaps out on the service floor. When somebody calls in for a part, the parts clerk answers and, after learning what part is needed, puts the calling line on hold. The line starts to blink at its appearance on the several phones. The parts clerk goes back to find out if the desired part is in its bin and, when the situation is clear, picks up on the nearest phone by simply depressing the held line's button and reports to the caller.

This approach can also be used in any small group where all the line numbers have their own buttons on each phone. Suppose, for instance, I am deeply into zip terms and practice MBWA (management by walking around). How does my secretary, or the receptionist, or whoever answers calls, find me? Maybe by paging. The answerer puts the call on hold and calls me over the paging system. "Mr. Goeller, pick up line 14." Then I, wherever I am, go to a phone, push the button for line 14 (which is blinking suitably as a clue), and greet the caller. Notice how much easier this is than having the call "parked" on a parking orbit by the original answerer, and then having me pick up a phone, get dial tone, dial the pick-up code (assuming I remember it), and then the parking orbit number. Parking orbits are often hard to identify when parking the fourth or fifth call.

With "principal and assistants," we have several assistants in a bull-pen area with a boss nearby in a private office. The assistants each have their own extension, and those extensions may or may not be in hunt, depending on circumstances. But all the assistants' extensions appear on all phones, as in "car parts" above, so that anybody can answer for anybody else, and know exactly which phone is ringing and then answered. The boss, of course, can pick up all the assistants' lines, but will probably have a private line, or a boss-secretary pair to which the assistants cannot connect by a simple button push. The idea here is that the assistants usually deal with the incoming calls but, when they get in over their heads, they use the intercom to call the boss who bridges on to help out. Note that "bridging on" in 1A2 simply involves pushing down the assistant's extension button and joining the conversation. This, of course, is a conference call, and was the most common way of making conference connections (we even do it at home, when the kids pick up on the bedroom phone while we use the one downstairs to talk to Grandma).

Of course, SOME of the above operations can, indeed, be done with single line phones, switch-hook flashes, and magic numbers, and the repertory dial phones that simplify such operations. For instance, my secretary can put the incoming call on hold or park, somehow identify the parking orbit, and call me, a call-waiting beep bursting in on my existing conversation; after flashing my switch hook to put my existing call on hold while getting information from my secretary, I may possibly be able to retrieve the new call with some variation of directed call pick-up, assuming the status of my original call permits. But this requires both me and my secretary to be trained to use the system, retrained when we forget, and trained again when we go to another office where a different system is used. Much harder than pushing down the blinking button. I had a friend who, during the course of a year, smashed three 2500 sets in frustration when his office's key system was replaced with a small PBX.

In 1A2 key systems, each line had a physical appearance on each phone that could pick it up. Thus you actually did the switching in the set when you picked up line four, and the set had a 25 pair cable (or larger) connecting it back to the key control equipment, to accommodate each line's talk and control paths. In new electronic systems, one has a single talk-path and a separate control path between the PBX line card and the proprietary set. All switching is done by the PBX switching matrix. This makes office wiring much easier (the number of pairs in the cable do not depend on the number of lines you want to pick up), and it offers new opportunities, but it also produces a whole new set of problems.

Problems Emulating 1A2

Let’s consider some of the problems first. With 1A2, control equipment for each line was called a KTU (for key telephone unit) and several KTUs would be located in a KSU (key service unit). In many instances, the KSU would be near its telephone sets; thus the multipair wiring would go from the KSU to the sets, but only a single pair would be needed from a KTU to the PBX or Central Office. Earlier electronic sets attempting to emulate 1A2 typically used 2, 3 or 4 pair cable rather than the older 25 pair cable or larger. This "skinny cable" has led IBM, AT&T, Northern Telecom, and others to offer standardized wiring plans, typically based on four pairs, although newer electronic sets seldom need more than one pair plus, perhaps, a pair for power. Four pairs seems like a lot less wire than is found in a 25 pair cable, and it is, from set to key closet where the KSU used to live. But from key closet to PBX, the same four pairs have to be extended instead of one, requiring much larger riser cables than before. You win some, you lose some.

Another problem concerns the number of matrix ports. With 1A2 (or, for that matter, 2500 type sets), one PBX or CO matrix port could support several "bridged" telephones. With 1A2, it was not unusual to have, say, three lines come in to the KSU, and five or more sets pick up the three lines. Because each set could access each line for both incoming and outgoing calls, and the lines could be put in hunt, the users were often better served than they would have been with five lines going to five single line sets. With the new electronic phones, it is not possible to have two sets bridged onto one PBX matrix port. There has to be a one-to-one match between set and port. Thus an inventory to prepare for an upgrade to a new system has to count sets, not ports on the existing matrix. Typically, the new system will have to have 20% more matrix ports for exactly the same service. A lot of people are still getting bopped with this one.

The numbering plan and inventory control also need attention. With 1A2, there was no problem in having more sets than lines (as suggested above), or more lines than sets (as when someone has two or more lines but only one phone). It was usual for each set to be identified with one specific extension number until the numbers were used up, and additional sets to be called "A Stations," and listed with one of the extensions followed by a letter designation. Thus three phones might be named 236, 237 and 238 after their extension numbers, while two additional sets, picking up the same lines, might be called 236A and 237A. In this way, specific phones could be matched against equipment listings.

With new electronic sets, the numbering plan becomes trickier. Although there has to be a pairing between set and port, there is no necessary relation between port and extension number, or set button and extension number. Because stored program systems can (and usually do) make a line-number to equipment-number translation, any extension number can (usually) be assigned to any port. But suppose the same extension is picked up on several sets. Then what? Our line/equipment translation is no longer one to one. How do we relate button, set, matrix port and extension number? All this has to be done in software, but different manufacturers do it differently. The most common approach is to define a "prime line" on each set, preserving the one to one match. Handling numbers without sets of their own is sometimes difficult; in earlier versions of the NEAX 2400, it was not possible to have an extension number without giving it a physical port on the matrix.

Opportunities With Electronic Sets

Because electronic sets usually have a single talk path and a signaling path, you do not mean quite what you used to when you say you are answering a call on line 243. What you mean is that the switching matrix will connect a call that thinks it wants extension 243 to your talk path when you pick up the phone; the control path is used to blink the lamp associated with the button labeled 243 on your set (and maybe on some other sets via their control paths), and ringing, either sent down your talk path or triggered from your set by another message on the signaling path, is announcing the presence of the incoming call. How it works, of course is less important than the fact that you answer the call just as with 1A2.

On an outgoing call, you use the signaling path (usually) to tell the PBX who you are calling, and the PBX connects your talk path through the matrix appropriately. Your matrix port is all the system needs to know; it really does not care which "extension number" you use. You may care if you want the call BILLED to your prime line, but that is accounting, not switching. If your secretary gets the other party on the line for you (a thoroughly rude practice) and then you pick up, simpler billing may be assured if the appearance of your prime line on your secretary's set was used to initiate the call. It is also easier for you to take over the call by pushing the button on your set rather than having the call transferred to you from the secretarial line. In either case, the switching matrix connection will be changed from the secretarial phone to yours, but the 1A2 approach works without training.

With all this flexibility, made possible by computer (stored program) control, other things can happen that are not like 1A2 but are equally easy to use. With Northern Telecom's Meridian SL-1, for instance, you can have one line "appear" on several sets, but not return busy tone until all sets are busy with different calls to that number. That is, all sets receive an incoming call indication on the button associated with the particular extension number, but as soon as one answers and the call is switched to its talk path, the lamps on the remaining sets go out and these sets are available for a new call. This contrasts sharply with the traditional approach that has the lamp lit at all appearances of a line to show that the line is busy. But the "line" isn't busy; we just have to understand what we mean by "line" in this case. We have a new opportunity here that did not exist with older systems: in effect, we have a miniature call distributor. This approach does not work well for call screening by a secretary, but it does what it is supposed to do just fine.

We can also have several appearances of the prime line on a given set, letting the user handle several calls at once on a single "extension number." If you have three different buttons labeled 243, you can use the button positions, along with the lamping at each button, to identify specific calls. This eliminates the need for hunting; it is a variation of call waiting that gives you visual cues to help understand what is going on. This approach, pioneered by Plessey and AT&T's Horizon, is also used on AT&T's current digital systems. Note that System 25 does not offer hunting except for conventional single-line phones.

But now we have a new problem. How does your secretary answer the second call coming in when you are tied up on the first? Here we have as many answers as we have system designers. The hard line approach says you answer your own calls. Period. Or maybe use call forward on no answer to some other line. However, the possibility of EITHER answering your own calls OR sending them to "coverage" is also possible, a variation of call forwarding all calls, but you have to push the "send all calls" button when you do not plan to answer. In any case, you do not have the convenience of call screening typical of 1A2. In some systems, the call is gone from your phone and is not necessarily identified as yours to the secretary.

System 75 originally permitted only a prime line (with several appearances) on its electronic sets. To add the "bridging" feature demanded by customers, AT&T chose to give all the principal's prime line appearances buttons on the secretary's set so that a call coming in on extension 243, middle button, could be identified at both. A somewhat different approach, used by Hitachi, has four buttons used as "talk lines" associated with each set. Single buttons are associated with the set's prime line and the prime lines of other sets to be covered. Normally your own prime line is selected, letting you handle four simultaneous calls. However, if another line is seen, by its lamping, to need attention, its prime line button is pushed to replace your prime line's appearances on your talk lines.

As we get further and further from basic 1A2, we can explore new ideas. Buttons on other sets can be programmed as "call forwarding targets" and/or as indicators for directed call pick-up. In either case, such buttons are identified with the prime line on somebody else's set, and lamping gives you a clue as to what is happening. With call forwarding, the other phone has to send calls to you, while with directed pick-up, you push the button to select the ringing "line." You can't tell the difference between this operation and depressing the button on 1A2...until you try to announce the call to the called party and get that person to pick up. Once the call is on your phone, in a number of systems it is no longer on the phone originally called.

Repertory dialing, where a button on your phone can be used to signal the system to ring some other phone and then set up the connection, looks like a good possibility for the intercom operation. But the far end needs a distinctive ring, and some means of responding to an intercom call vs. regular call, after putting an existing call on hold.

Repertory dialing is also useful for selecting particular trunk groups; you might, for instance, have specific buttons for the FX groups to New York and Chicago, and Local. Lamping would tell you if at least one trunk in the group is free, and pushing the button would give you a trunk in that group. Similarly, group pick-up for a given group of extensions can appear on a particular button; lamping can tell if any station in the group is ringing unanswered, and pushing the button can cause the switch to connect it to your talk-path. Although not quite 1A2, these techniques need little training. However, the problem is not answering a call for somebody else, but screening the call and announcing it privately to the called party, or putting it on hold and picking it up yourself at another phone. These are the things 1A2 did best, but are also the main thing that more "modern" approaches do not do well at all.

Because anything made of plastic has to be cast in molds that get more and more expensive as the number of parts increases, it is often desirable, for economic reasons, to minimize the number of buttons on a set. One way to do this is to use a liquid-crystal display capable of showing 40 or more alphanumeric characters. Such displays can tell you the name and extension number of the person calling you on an inside call, identifying your secretary, for instance, who may be screening an outside call. Or, with group pick-up, you will see the name and number of the phone you are answering. With such a display, the system can also tell you that the number you just called is forwarding to another number, or has "do not disturb" in effect. Such displays are helpful, and can often let you get away with fewer buttons on your phone. But they also do two other things: they let you handle messaging, and they provide feature prompts.

A number of systems now have a variety of "canned messages" such as "do not disturb," "out to lunch," "back at 3 pm," "please call me," etc., which can easily be displayed to others. It is even possible to make up your own messages, but entering them with a regular push-button pad tends to be tedious. Feature prompts, on the other hand, eliminate the need for much training, and make it possible to use some of the features normally ignored. The Mitel Superset 4 is particularly good at this; using software-defined buttons, which are labeled in the display area by the system program, a variety of helps are made available in almost all circumstances. The Telenova set not only offers helps for its PBX but also can be arranged to do the same thing for connected systems such as voice mail.

Testing For 1A2 Capabilities

It is hard to get a straight answer to the question: Can your electronic phones emulate 1A2 key operation? Most vendors say they can, and many of them actually believe it. But the chances are quite good that they do not know what you mean, and you may not find out what they mean until after cutover. So how can you test?

The first thing is to be sure one set can pick up several different lines, each with its own button, and a line can appear on several different sets, any of which can originate or terminate calls using that particular line identity. It should be possible to originate or answer a call on one set and pick it up or bridge (conference) to it at another without going through the transfer procedure. An intercom function should be available to permit separate signaling and then conversation in connection with call screening. Both personal (where a single button selects the intercom channel and alerts the other party) and dial intercom (where the channel is selected and the signaling pad used to designate the party to alert) are appropriate in differing situations; programming an electronic set to do either should not be too hard. A HOLD button should be available for use with any call on any line appearance. Lamping should make clear the status of each call on each button. Try out the phones in a working system, emphasizing, in particular, the boss-secretary pattern, and you can probably tell if the system works like 1A2 or has been improved and modernized in such a way that user training is essential before calls can be made or answered.

Of course, there is no reason to limit a modern telephone to 1A2 emulation. Directed and group pickup and repertory dialing, along with messaging, calling party, directory, and other displays are all easily added. But these are features that should be available in addition to rather than instead of what the customer is already used to.

Remember, the whole idea is to eliminate the need for training. Not just initial training at cutover of your new system, but refresher courses for long-time employees and special sessions for each new hire. Even if the vendor offers to do the training for free, the time the trainees have to take away from their regular jobs is expensive. Station users should not have to have a PhD (Phone Dialer) diploma hanging on the wall before they can be trusted to make a business call. Organizing systems for this objective should be a major strategy of those other PhDs, the Phone Designers, looking for a minimum feature set.

[ Top ] [ Next ] [ Table of Contents ]


Copyright 2006 Lee Goeller. All Rights Reserved.