# Oral-History:Amos Joel (1993)

Amos Joel

Amos Joel was born 12 March 1918 in Philadelphia, but he soon moved to Atlantic City and then New York City. As a child, Joel became interested in mechanical and electrical devices, including electric trains and radio. It was Joel’s curiosity about the dial telephone system and circuits, however, which would shape his later career in telecommunications. As a young man he became fascinated with reading patents, particularly in telecommunication switches, and in high school he enjoyed his mathematics and science courses. He even invented his own switching system in the early 1930s. Joel was accepted into MIT and received his bachelors (1940) and masters (1942) from there in electrical engineering. In 1940, Joel went to work at Bell Labs, and over his long career there he worked in many areas and projects, such as the accounting center, operator services, transistors, electronic switching system at Whippany, Stored Program Control, No. 1 ESS and Traffic Service Positions System (TSPS) to name a few, as well as becoming a director. He received many patents himself, and also many awards including the Kyoto Prize, and at the time of the interview Joel was looking forward to receiving the IEEE Medal of Honor and Patent Recognition Award from AT&T. Joel was also very involved in IEEE, including being a part of the formation of the Switching Committee (1947) which he eventually became a chairman of, head of the Nomenclature Committee at Bell Labs, chairman of the New York section, and president of the Communications Society. Joel also has numerous publications to his name, including books under the IEEE Press Book Series imprint. Joel’s retirement from Bell Labs in March 1983 did not stop his involvement in the engineering world, as he became a consultant for Bell and other companies, as well as continuing to write, teach and lecture about switching systems.

This oral history covers the international history of telephone switching from the late nineteenth century through 1993. It examines manufacturing, markets, the effects of Bell divestiture, and technological advancements. From a global perspective, Joel assesses the history of technical publication on switching. He considers social, technological, and institutional influences on switching technology, as well as the social implications of telephone technologies. The discussion of compatibility requirements for technological advances provides insight on consumer demands and on the telephone industry in the 1990s.

See Amos Joel Oral History (1992) for Joel's detailed discussion of his own education, career, and achievements. The 1992 interview also provides insight on the work environment at Bell Labs, on the professionalization of switching through groups like the Switching Committee of the IEEE Communications Society, and on Joel's leadership in the IEEE and the International Switching Symposium.

AMOS JOEL: An Interview Conducted by Rik Nebeker, Center for the History of Electrical Engineering, 17 June 1993

Interview #163 for the Center for the History of Electrical Engineering, The Institute of Electrical and Electronics Engineers, Inc.

This manuscript is being made available for research purposes only. All literary rights in the manuscript, including the right to publish, are reserved to the IEEE History Center. No part of the manuscript may be quoted for publication without the written permission of the Director of IEEE History Center.

Request for permission to quote for publication should be addressed to the IEEE History Center Oral History Program, IEEE History Center, 445 Hoes Lane, Piscataway, NJ 08854 USA or ieee-history@ieee.org. It should include identification of the specific passages to be quoted, anticipated use of the passages, and identification of the user.

It is recommended that this oral history be cited as follows:Amos Joel, an oral history conducted in 1993 by Rik Nebeker, IEEE History Center, Piscataway, NJ, USA.

## Interview

INTERVIEW: Amos Joel

INTERVIEWED BY: Rik Nebeker

PLACE: South Orange, New Jersey

DATE: June 17, 1993

### History of telephone switching equipment manufacturing

Nebeker:

What I was hoping was that you could give me the broadest outlines of the field of telephone switching in several ways. And I'll ask you about them one by one. One thing is simply the main manufacturers of switching equipment over the last fifty or more years. If you could just list the companies that have been the biggest producers of switching equipment.

Joel:

When I think of telephone switching, I go back over to the invention of the telephone. Is that going back too far?

Nebeker:

No, that's fine.

Joel:

All right. So at the time the telephone was invented, or shortly thereafter, the Western Electric Company became the primary company to manufacture switching equipment.

Nebeker:

Was that always owned by AT&T?

Joel:

Well, it became a part of AT&T very early. It used to be called Gray & Barton--that was the name of a company who were making burglar alarms and telegraph equipment and things of that kind, plus consumer products. But they were soon bought up by AT&T. I think it was around 1881 or something. It became called Western Electric at that time. Then of course with the making of the switchboard -- I shouldn't say standardizing it--but as the form of the manual switchboard became obvious, you know, what you needed to do the job, many competing companies got into that business. Principally companies like Kellogg, and Stromberg-Carlson. I can't think of all of them right now, but there was a whole lot of them.

Nebeker:

Now were they selling to Bell as well, or to the independents?

Joel:

Well, the independents because the independents were big business back in those days. Bell was small, and Bell was pushing the technology to get a nationwide network, an inter-city network, at that time. So they were the long-distance company, but in the local market there was competition at that time, and there could be two or three telephone companies, each serving the same community. There were a lot of companies in the business. The company that eventually became Automatic Electric started in this business. It was first called the Strowger Automatic Telephone Company. I think that company was started by fellows named McGray and Meyers in 1889, just about the time the Strowger patent was issued. Because the Strowger patent was applied for about that time, they went into business as the Strowger Automatic Telephone Company, and it eventually became the Automatic Electric Company. That change didn't occur, though, until somewhere in the teens.

Nebeker:

And what has happened to it since?

Joel:

It eventually became absorbed into GTE in the 1950s.

Nebeker:

So in this country you had in the early years of the telephone, a lot of manufacturers?

Joel:

A lot of manufacturers of manual switchboards. Of course abroad, too. The Erickson Company got started with their own brand of switchboards that were sold domestically and also internationally. They started very early in selling switchboards out of Sweden.

Nebeker:

What are the main companies today in the United States producing switching equipment?

Joel:

Well, today we have AT&T, and we have Northern Telecom.

Nebeker:

Is that a U.S. company?

Joel:

Yes, Northern Telecom Limited is a subsidiary of Bell Enterprises up in Canada. But they have a U.S. subsidiary, and most of the equipment is produced here in the United States. In fact, that's true for all the other companies besides AT&T, that they all have foreign origins of some kind but are now producing in the United States. In addition to Northern Telecom Incorporated which is a U.S. company and produces equipment in Raleigh and in Richardson, Texas, we have companies like Erickson, who are producing equipment in Texas. And we have Siemens, who bought out Stromberg-Carlson, so it's Siemens and Stromberg-Carlson who produce equipment down in Florida. They're the major ones in the United States today. Now if you're going back over the history, a strong competitor in the United States was IT&T. They made equipment here at one time. Kellogg was bought up by IT&T. There's another company that was bought partially by Erickson, which is called North Electric, who produced automatic switching equipment back in the early days. So there were a lot of companies. But North Electric is a very big one, as were Stromberg and Western Electric back in the early days.

Nebeker:

Joel:

No, IT&T wasn't formed until 1922 or 1923. AT&T sold its international businesses and became IT&T.

Nebeker:

Did they produce switching equipment before they bought Kellogg?

Joel:

In this country?

Nebeker:

Or elsewhere.

Joel:

Well, elsewhere they did because these companies that were AT&T/Western Electric International were making equipment abroad and designing equipment abroad that was different than was sold in the U.S. market.

### International switching equipment manufacturing

Joel:

By the way, this history is going to cover international?

Nebeker:

Right, international as well.

Joel:

Oh, so then you do have a lot to cover here.

Nebeker:

Yes. [Chuckling]

Joel:

Boy! See, before IT&T was formed, AT&T had Western Electric International, which was located in pre-Soviet Russia. They were also in Belgium, in England, in Spain, in France, and in Japan. They were in many places.

Nebeker:

So they must have been advanced in their technology to move in so well.

Joel:

Well, they were making different technology than here, and some of it had its genesis here. If you go back and look at work that was going on in 1903, 1905, in that era, they were all being designed here at Western Electric Company, at a department which later became Bell Laboratories. They were designing what was a generic class of switching systems, known as power-driven switches, where you have motors that are continuously running, and you clutch on and off of the drive chain to cause the switches to move either up and down or around according to whatever kind of switches they used. That was the technology that Western Electric was really exploiting at the time.

Nebeker:

So they were designing things for different foreign markets?

Joel:

Well, the ideas started in the U.S., and then, for example, a great decision was made that we needed a large switch in the United States. Therefore, we were going to try to get a switch with which you could select one of 500 terminals. Whereas one of the other ideas that was coming out of the Labs was the rotary switch where you could select only one out of 300. So they decided, well, that would be a better switch for them to try to design in Europe, and so they sent that design over to Europe to begin.

Nebeker:

Was it the case that the technology was more advanced in this country generally?

Joel:

No, but don't forget that the idea of the automatic telephone, automatic switching to replace operators started here. So we had a head start in many respects.

Nebeker:

It sounds like Western Electric International was able to do a business in so many different countries. They must have had something going for them.

Joel:

Well, they had good manual switchboards. They had very good and quite advanced. Of course there was competition even in those days that produced a variety of automatic switchboards that were different.

Nebeker:

Did Siemens get in the business early?

Joel:

No, no. Siemens did get into the automatic switch business at the beginning of the century, but for the most part under license from Automatic Electric Company. Later they started making their own variety. By the way, at that time it was called Siemens & Haskell. It was not the same Siemens that produced generating systems and all the rest of it.

Nebeker:

In other words there were different branches of Siemens.

Joel:

Yes, it was a branch of Siemens. But they were an international company even then, and they produced systems in England and in Austria and Germany and, I think, many other places in the world.

Nebeker:

So the situation in the U.S. today is that you have AT&T, these others you named, Northern Telecom and Siemens--

Joel:

Siemens and Stromberg.

Nebeker:

Is it marketed as Siemens or Stromberg-Carlson?

Joel:

They only bought Stromberg-Carlson about three years ago, four years ago. So they're still marketing the Stromberg-designed systems under Stromberg, and the Siemens-designed system, which is imported from Germany, under the Siemens name. But it's really becoming one company. They've tried--this is something that's never worked out in switching--to merge the two designs so that they could have only one product. But it hasn't worked out that way, and it never has for several cases in the history when they tried. Not just that company, but other companies have tried to do that. For example, ALCATEL which is a very large international company, not in the U.S. but elsewhere in the world today--it's a French company. Well, the headquarters is now in Belgium, and it's got a Belgian name, but it's a French company, originally.

Nebeker:

And they're a big name in research?

Joel:

Well, they bought up all of the rest of the IT&T interests around the world, and they had their own in France, which had also had international markets. So they're very big worldwide, but not in the U.S. They have no toehold here yet. They've tried several times prior to their taking over IT&T. IT&T tried to start up here several times and also never succeeded. So ALCATEL is still looking over the situation as far as switching is concerned. Now they do market transmission here, but they haven't marketed any switching equipment, and they're not making any at this time.

Nebeker:

What about other large European manufacturers after Erickson and Siemens and ALCATEL?

Joel:

That's about it. Well, things have changed a lot. There's a company that was Standard Telephone & Cable [STC] that made switching equipment in England, the U.K. That was bought up or an interest in it by Northern Telecom. Another company is GEC, which is not General Electric in the U.S. but General Electric Company in England--it's a different company entirely. A big interest in them has been bought by Siemens. So there's a lot of mixtures going on nowadays.

Nebeker:

There are also, I understand, these niche companies, like Redcom, that produce a particular product.

Joel:

Yes, even in the U.S. Well, that's a special product, the Redcom, but we do have a company in the U.S., for example, DSC. It used to be called Digital Switch Corporation, but now it's DSC Corporation. That's a small company that makes switching equipment here for the inter-exchange carrier market. For example, they sell to MCI and Sprint, people like that, and private networks. So there are a number of niche companies, yes.

Nebeker:

What about Japanese companies? Are they big in international switching?

Joel:

Up until now, they've been rather unsuccessful. Nippon Electric, NEC, which has had facilities in this country making switching equipment for quite a while, since the late seventies, and sold a little bit here and a little bit there, but they've never been really successful. In fact, I believe they recently decided to pull out their current product line. They haven't been doing it for very long. On the other hand, we have Fujitsu who are coming along, trying in a big way to enter the market here with the next generation of switching stuff. They're working very hard to try to build up the American market. None of the other companies that make switching equipment in Japan have tried the U.S. market. There's Oki and Hitachi, but they're not doing much here.

Nebeker:

But they do manufacture switching equipment?

Joel:

Yes, they manufacture switching equipment for NT&T, the Nippon Tel & Tel.

### Market for switching equipment after Bell divestiture

Nebeker:

What's the market for switching equipment now, with all the baby Bells and all the long-distance carriers?

Joel:

And the independents.

Nebeker:

There are still a lot of them?

Joel:

Yes, there are still a lot of independent companies, but their market is not a very large market, and it's dropped quite a bit in the last few years. You see, a new technology came along in the late seventies which was the digital time division local office. AT&T had pioneered in big switching systems time-division digital for the tolls.

Nebeker:

Right. I know about that general system.

Joel:

Number 4-ESS. You've probably heard of Number 4-ESS.

Nebeker:

Yes.

Joel:

And that's still the backbone of the AT&T network. But they started with Number 5-ESS, and of course before that Northern Telecom had DMS-10 and DMS-100. Stromberg-Carlson had their DCO. North Electric had its own system too.

Nebeker:

This is for small branching networks?

Joel:

For local offices. The end office, so-called. The ones to which your telephones are connected. So there was a big push in the early eighties, after those systems were designed and successful, to convert all electromechanical systems into electronic systems. Those products were the ones that were sold for the most part which were time-division digital. Now the Bell Companies had already invested a lot of money in the so-called analog. I don't like the term "analog." It's actually space-division electronic switching system. Analog is a term used in the trade press which bothers me because it's not technically accurate. But nevertheless, they're called space-division systems. Like No. 1-ESS, 1A-ESS, and so forth.

Nebeker:

Space division means you're using different bars?

Joel:

Yes. Different paths through the switch. And in the case of the ones that were used in the Bell System, they used sealed contacts. So today about 40 percent of the switching offices in the Bell Operating Companies are still those systems. They were put in during the seventies, and they're not fully depreciated yet. They're still doing a good job for the most part. Except for some of the new services, things like ISDN, you can't do with those systems.

Nebeker:

I'm just trying to understand the market structure. You have obviously large telephone companies, either for local or long-distance service, and they need switching equipment. You also have, I'm sure, businesses that want their own PBX.

Joel:

Well, that's different. That's PBX systems.

Nebeker:

You're not talking about that in any of this?

Joel:

No, I haven't been talking about the PBX. That's a whole different situation. That's had its ups and downs, too, but in a different way. But the big push was in the central office area, or the so-called end offices, which are the ones to which your telephones are connected. They're the ones in the telephone building downtown center of the various cities, that either take care of your local calls or launch into the toll network through the country. Of course, when divestiture came along all the electronic systems had to be arranged so that you could have a choice of your long-distance carrier and so forth. But that was only possible if you had electronic switching.

### International comparison of market structures for switching equipment

Nebeker:

So we have this market for the fairly large switching equipment, but in most countries, isn't it pretty much a government-owned PTT that provides the telephone?

Joel:

Well, yes. We originally had a monopoly situation in this country with AT&T, but that's a private company. In other countries, the post office and the telecommunications were usually one and the same ministry, and there was a state-owned monopoly in most places.

Nebeker:

So they're buying all the switching equipment essentially for their own sake?

Joel:

For their own country. See, they had some internal competition in most of these countries. For example, in Germany they had Siemens, and they had an IT&T subsidiary known as SEL, Standard Electric Lorain, and that subsidiary made a different system than the Siemens system at one time.

Nebeker:

But the whole market in Germany, the PTT bought?

Joel:

Nebeker:

So it's really a very special market structure?

Joel:

Oh, yes.

Nebeker:

But, this country is quite open...

Joel:

Well, that's a point. Our country is so open that these companies all moved in when they saw the opportunity to sell to the Bell Operating Companies because the monopoly had been broken with the divestiture. They all wanted to try to get a piece of that action, and they have succeeded to some extent. Maybe not as much as they had hoped. But they at least now have their toehold here, and of course a lot of the production is going on here. So it's hard to call them a foreign importing company.

Nebeker:

Are there other countries that are open?

Joel:

Most countries don't allow this. In most, you have to have production in the country before you can sell in that country. See, these people came over here and brought their stuff in, and then after they sold enough, they set up their own production operations here. But initially they didn't have any U.S.-based production.

Nebeker:

What about any other countries that have independent telephone companies? Do they exist?

Joel:

They don't, no. Well, there's always exceptions to the rule, but generally speaking, no. There's a city in England that's independent of the regular--what was the regular--monopoly. In fact, recently in England, the monopoly's been broken, too. The post office ran the telephone company for many, many years--more than a hundred years. Back in the mid-eighties, the British Telecom became a private company and took over the post office facilities. And then the market opened up. They put in new long-distance carriers, just like they did here with MCI. They put in a long-distance carrier called Mercury, which was bought up by Cable & Wireless, so the monopoly has been broken there now. It hasn't been broken in France, and it hasn't been broken yet in Germany, but the handwriting's on the wall. Sooner or later competition of one kind or another has to be brought into those markets.

### PBX switching systems history

Nebeker:

Okay. So that's the structure of the market for large switching systems. Then you have PBX.

Joel:

Now, PBX is a different situation. PBX's are smaller, and they're made for rather special capabilities. They're much simpler in design than the central offices, particularly the maintenance and reliability aspects. Also, although only in the United States so far--there's some signs that it's trying to expand out of the United States--there's competition for PBX's. You have the alternative known as Centrex. What Centrex is, is the ability for the central office that we were talking about earlier to act like a PBX. So that instead, for example, of Rutgers having a separate little PBX switching systems for all the thousands of lines that are in Rutgers, they have all their lines going into a central office in downtown New Brunswick, and that is a regular central office that's serving ordinary customers elsewhere in New Brunswick as well. In there a group of numbers is set aside for inter-communicating within the university.

Nebeker:

So you can still dial an extension if you wanted to?

Joel:

Right. You can still do it as if it were a private thing. But it's actually the same central office that's serving other customers in the area.

Nebeker:

When was that introduced?

Joel:

Well, the concept started in 1958 or something with electromechanical switching, but it was taken over with electronic switching. That was a very interesting idea. The problem was getting rate structures that would support it and would make it attractive in service. That's had its ups and downs, but the main thing that happened was when divestiture came along. Prior to divestiture, there were lots and lots of companies in the business of making PBX's and selling them. You see, the first thing that broke open in this country was not just the divestiture, but the first thing was interconnection. You were allowed all of a sudden to buy your own telephones and connect your own stuff to the telephone lines. So businesses went out and bought PBX's and connected them to the telephone lines, which they couldn't do prior to interconnection.

Nebeker:

When was that?

Joel:

That started fairly early in the seventies, after the so-called Carter Phone Case, which broke open the ability for people to tie things onto the telephone network. AT&T and other people said, “You can't do this. It's going to degrade service.” So forth and so on. Then finally the government said, “You have to allow interconnection.” First they had interface devices you had to buy, and then later they got rid of that. So after that broke open, a lot of people started selling PBX's because that was a good market. Instead of leasing the PBX from the telephone company, which is what was going on before, you just bought PBX's.

Nebeker:

So the PBX's before would be produced by Western Electric?

Joel:

Yes.

Nebeker:

Pretty much all of them?

Joel:

Yes, with Automatic Electric for the independents. In fact, a lot of companies made PBX's for the independents, but it was the same lease idea. In other words, you'd go to your local telephone company, and you'd say, “Look, I'm building a building here. We're going to have an office or a factory or something, and we want a PBX,” and they'd put it in. You'd lease it from them. And you'd pay so much a month.

Nebeker:

Was it a matter of new manufacturers getting into the game, or the ones who were there moving into PBX's?

Joel:

Audio File
MP3 Audio
(163 - joel - clip 1.mp3)

Well, both. Certainly the ones that were there, like AT&T, wanted to continue to sell PBX's. So they tried to sell directly to the consumer, but then a lot of other start-up companies tried to get into the market. There was a time when the technology was changing to this time-division digital technology, and so a lot of people saw the opportunity to get into that without having to invest a lot in new technology. They could use the computer technology and the transmission technology that existed. They didn't have to design special switching gadgets and things of that kind. Remember I mentioned several of the switches had sealed contacts? Well, they didn't have to bother with that. They could go out and buy ordinary chips and do the job of building a PBX. But then that started taking business away from Centrex. Not only that, but Centrex was hidebound by tariffs that had to be set by the public utility commission, whereas the PBX had no tariffs associated with it. So Centrex started losing out.

Well, now divestiture came along around 1983-1984. And when it came along the telephone companies were faced with something different. Now the only thing that they could make some money on was to sell Centrex service, because they now had all these central offices, and they wanted to keep the customers they had. They came up with new tariff ideas that the PUC's approved. Centrex has been growing quite rapidly nowadays, and many people who had PBX's are back to using Centrex, because Centrex is much better in many respects from a customer's point of view because they don't have to have anything on their premises, they don't have to maintain it, there's the guaranteed twenty-four-hour service and maintenance service by the telephone company, and as new improvements come along, like ISDN, you get that, as soon as the telephone company has it. It used to be you had to pay mileage charges for every line that had to go from Rutgers into the central office, however many miles that was. In some cases it was a lot of miles out in the country, and it was just too expensive, but now they have tariffs that recognize that you have to make it more viable. So that's the PBX's. Abroad, the PBX's are still a hidebound situation. You can't buy your own PBX's; you can only buy them from the telephone company, or a type approved by the telephone company. The telephone companies in some places, like Germany, will approve, let's say, a certain type of PBX. It can only be approved by the government after they make all their own tests on it.

### Data switching; switching equipment for Local Area Networks

Nebeker:

So are there any other parts of the market for switching equipment that we haven't covered?

Joel:

Well, private networks, but that's the same as PBX's for the most part. And also there's of course data switching--packet switching and that sort of thing.

Nebeker:

So there's switching equipment that's manufactured specifically for it?

Joel:

Oh, sure.

Nebeker:

And where is that used?

Joel:

Well, a lot of companies have a great deal of data communications, and it starts out at the desk where you have your personal computer. Within the premises you have what's known as the Local Area Network. As far as I'm concerned, each one of those things is a switch because you can send a message to any one of these other stations on this Local Area Network. Now, it's also possible to use modems, you know, over the telephone network, so that's another way you can form data. The last one is that there are companies that offer data services where you take a message or a file and it's broken up into packets of 512 bytes or 1,000 bytes or whatever, and you send these packets, each with a header saying where it's going to. This goes over a private network or a separate network for data, instead of using the telephone network like the modems do.

Nebeker:

I see. I just hadn't thought of these Local Area Networks as being switch-serviced.

Joel:

In my definition of a switcher, the thing that makes a switch is anything where you have selection in a telecommunication network. It doesn’t matter whether the selection is done at the customer's premise or wherever it's done. If you have to be able to make a selection, then you have to have switching.

Nebeker:

And there's switching equipment manufactured by AT&T or Northern Telecom for these local area networks?

Joel:

Oh, many more companies than that. In fact those don't make very much in that area, although AT&T's got a thing called Datakit. Northern doesn't have anything. But there are many others--hundreds of companies--making software and hardware for this.

Nebeker:

But, I mean, is it really the same type of equipment?

Joel:

No, nothing at all like it. In the case of Local Area Networks, you'll hardly recognize the switch because it's distributed across every station where you can make the selection. “Here comes a message. Is it a message intended for this station or the next station along the line?” And so the message just goes on its way down this thing, and you pull it off when it reaches the right station. The switching is at the station.

Nebeker:

You don't have some PBX sort of thing?

Joel:

You don't have the central switch.

Nebeker:

But isn’t it based in the software?

Joel:

Well, yes, but there's also hardware there. The minute you enter the terminal through hardware, there's something that's in there that says: "We're going to take this message," or "We're going to pass it on to the next station," so it has to read the address on the message, on the packet. Now there are people in the business of packet switching because there's a big business in getting messages out of these Local Area Networks and into a value-added network. For example, CompuServe and all those people depend upon packet networks to get access--through Prodigy and so forth--to their data. What you do there is you dial up a number, and you use your modem, then you go over the telephone network to their network, there it's converted into packets, then it goes over the telephone network as regular stuff directly out of your computer. That's the other place where switching is big.

### Switching and broadcast television networks

Joel:

Now, of course there are many other special situations. At one time there was a lot of switching done to switch broadcast networks. You know, we had broadcast networks all over the country--NBC, ABC, and so forth. Those would use the telephone company's facilities, and we had very special switches to switch that. I like to talk about the fact that on a Saturday afternoon when there were all kinds of regional football games going on, the network looked one way, and then at various times in the afternoon they got switched around different ways. That was all done with switching equipment.

Nebeker:

And that was special switching equipment?

Joel:

Yes, special switches. But nowadays it's different because each network has its own facilities, and they do their switching on their own premises, but there are still switches somewhere. First of all, they're switching right at the production room from one camera to another. But today, you know, you pull in these different locations by satellite.

Nebeker:

They switch to this game they receive?

Joel:

Yes. So it's a different pattern. It's not the same business as it was when the telephone company ran the networks and provided the networks with the facilities.

### History of technical publication on telephony

Nebeker:

I think you've done a good job of telling me the users and the manufacturers in broad-brush terms. What about journals for the technical contributions to the field? What are the main journals?

Joel:

Well, if you go way back to the beginning of telephony, there wasn't much. In the late 1880's and 1890's, there were only one or two journals. For example, people looked at that time to the journal called Electrical World, which was mostly power, I guess. Everything was just starting then anyway...power and telephones, and so on. So Electrical World was the trade publication at one time back then. There was another one, a publication that's still in business today, called Telephony. That got its start as another name. It was called Acoustics? There was some other name to it, then it started to take on mostly stuff about what was going on in the telephone industry. And became, at one time, the backbone publication for the trade in the independent telephone industry, the non-Bell. In fact, this is one of the things that's bothered me over the years, that there were not a lot of technical publications in the telephone business.

Nebeker:

What about the Bell System technical journals?

Joel:

They used physics journals for some of the original stuff--Berliner and Bell and those people. There was also the AIEE Transactions. If you look at the AIEE Transactions--I probably have them downstairs--I'll bet you there's no more than twenty articles from 1900 to 1920 to 1925 about telephony.

Nebeker:

In all aspects?

Joel:

All aspects. You know, repeaters and switching equipment. Very, very few articles. Of course there were theoretical articles by Campbell and people like that who were working on the theory of transmission, but they weren't in those journals. They were in more learned journals.

Nebeker:

When did the Bell System Technical Journal start?

Joel:

That started when the Bell Labs started, about 1920 to 1925.

Nebeker:

What about overseas in the early years? Do you know of any journals?

Joel:

Yes. That's what I was just trying to think about. There were journals overseas. First of all, the companies we've mentioned--not all, but the big companies--Western Electric, Ericsson, and Automatic Electric and some of these others had house journals.

Nebeker:

They would present their work?

Joel:

They would present stories on their products. They were somewhat technical, not very, but somewhat. There wasn't much you could get out of those journals, and, the AIEE articles in Transactions weren't that great either. But from time to time somebody would give a paper. Of course, in those days, if you look at the technical transactions, they were truly transactions, because in those days, if somebody gave a paper, they would record the questions that were asked and the answers.

Nebeker:

Yes. I like that.

Joel:

Yes. That was great, wasn't it? I wish we could do it today. So that was good because it was interesting to read the questions. Overseas, by the way, I know in Great Britain as well as in Sweden, the post office being the body that ran the telephone company, had a journal, the Post Office and Electrical Engineering Journal. That was very popular, and went way back at least to the teens, and maybe even earlier than that at the turn of the century. They had some good articles.

Nebeker:

So most of the PTTs had a technical journal?

Joel:

Yes. Most of the PTTs, I'm sure. The Germans had some great textbooks back in the teens about telephony.

Nebeker:

I also wanted to hear what happened since the Bell System Technical Journal.

Joel:

Nebeker:

What about the Bell System Technical Journal? Did it start recording a lot of the advances?

Joel:

Then, no. Later on it did, when electronics came along. But, no. There's very little in the BSTJ. The Bell Laboratories Record, which is another publication that started in 1925 also, tended to have a lot more articles. They did publish simple articles, two or three pages, about improvements in the various parts of the Bell System.

Nebeker:

I see. So that's something one could look to?

Joel:

Yes. You could go in and look at the panel. You could look in the Bell Laboratories Record and find articles about the panel system improvements as they came along. They didn't even refer to the AIEE paper or anything like that. This was progress by Bell Laboratories. Same thing with switchboards--there were a lot of articles in there about manual switchboards and improvements being made in those simultaneously. And of course one of the major switching systems, known as the step-by-step system, was not a Bell System invention, so the Bell System didn't embrace that until very late. They got an agreement in 1918 with Automatic Electric, who were the main producers and innovators of that system, and they came along and installed it. They embraced that for smaller cities in the Bell System. Part of that agreement was that what Automatic Electric couldn't supply, Western Electric would build themselves. Of course that meant Western Electric went in there in a big way and started producing step-by-step equipment. Except for the Depression years when Automatic Electric had plenty of capacity, and therefore Western Electric wasn't allowed to make step-by-step equipment and you had to buy it from Automatic Electric. In the meantime, you read articles in the Bell System Technical Journal starting in 1925 about improvements that Western Electric/Bell Laboratories made in the step-by-step system on their own. But the technology didn't really change until after World War II.

Nebeker:

And what's happened since World War II?

Joel:

Well, since then, things have opened up. I couldn't be more pleased.

Nebeker:

And what are the main journals?

Joel:

If I were a student today and wanted to learn about switching, there's certainly a lot of places I could go to read technical stuff. It might not completely satisfy me even then, but at least you'd know something about what's going on in switching.

Nebeker:

But what are the main journals now for advances in switching? Where do these papers appear?

Joel:

Well, in the United States, first of all, you'd want to go to the conferences, or at least get the conference records of three things. The ICC in June, the International Conference on Communications. The GlobeCom in November or December. They’re both also held by the Communications Society. There's also the International Switching Symposium, which is held every two or three years someplace else in the world.

Nebeker:

When did that start?

Joel:

Well, we look back and say the genesis was in 1958 when we had the first such international meeting on switching at Bell Laboratories. It took its current name in 1972, but everybody that's in the industry now looks back at the 1958 one and says that was the first one. We just had one last year in Yokohama, Japan. It was the 13th ISS. I'm called the "Father of ISS", among other things. [Chuckling] Anyway, we started those meetings back there when we first wanted to tell the world about our progress in electronic switching.

Nebeker:

So the proceedings, or records, of those would be available?

Joel:

At least those now. The other thing is, just like when Bell Laboratories Record came along, BSTJ started publishing articles on switching. There are at least, two dozen issues of BSTJ that have a lot of stuff on switching in them, and individual articles over the years, too.

Nebeker:

I would presume the Transactions of the Communications Society is a good source?

Joel:

The Transactions of the Communications Society has had lots of articles on switching. That's a whole section of the Transactions. There’s also the house organs of other companies around the world now, not just Bell Labs Record.

Nebeker:

Are those things actually useful to the switching engineers?

Joel:

Oh, yes. Definitely, the current stuff.

Nebeker:

I see. So you could trace the history of a lot of the companies?

Joel:

Yes. Just look at the house organs of Erickson, which I said go back quite a way, and you could trace the history of what they did over the years. They were much better in many respects than AT&T.

Nebeker:

Are there good libraries that keep these publications?

Joel:

No. That's why I'm worried about what to do with all this stuff I've got.

Nebeker:

Maybe AT&T has a good collection.

Joel:

Well, they have, but they're not interested in what other people are doing nowadays. Maybe back in my day at AT&T they might have been interested. In fact at AT&T I could go down to the library and look at the Erickson and all these other trade in-house organs. Nowadays AT&T in many cases doesn't want to spend the money on some of these things, and they don't trade them like they used to trade things of that kind. There’s a lot more politics involved in this sort of thing than there used to be. In fact I'm not so sure just where I could go to get the Erickson Review today. But there are many such journals. In France there was a very good series, in French, on switching papers, a switching periodical that was put out. It's now been combined with a transmission one called Commutation et Transmission, which is "switching and transmission," and it comes out four times a year. IT&T's journal, once they got started in the twenties, had lots of good articles on switching and so forth.

Nebeker:

What's the motivation for these companies to reveal their developments if they have something they want to sell?

Joel:

Well, primarily to increase sales, especially, in the case of IT&T, who were competing worldwide with all these other manufacturers. In the case of the Bell System, Bell System Technical Journal and Bell Labs Record, it was just to keep the Bell System companies informed about things that they might not otherwise hear about through regular letters and things published by AT&T for the benefit of the Bell System.

### Patent literature on switching

Nebeker:

From what you say, I gather before World War II, people working on switching at Bell Labs weren't encouraged to write papers.

Joel:

No, no. There were papers at Bell Labs, but not very many. No, it was discouraging. I guess you know that the way I learned a lot about switching was to read patents, and there, of course, you get a completely different view of switching, because that tells you about things, ideas, that people have. It doesn't tell you what became commercial. You just take, for example, just the AT&T patents alone, which cover the field tremendously, and of which maybe ten percent are things that were actually used. Even in Automatic Electric, which was a big company in the switching field, a very small percentage of what they patented was actually used.

Nebeker:

How important would you say is the patent literature? Are people spending a lot of time reading patents?

Joel:

I would say, no. People don't spend a lot of time reading patents, and it probably is not too useful for the reason I mentioned, that perhaps ten percent of the patent literature is useful from a commercial point of view, learning about a specific product. I get involved as a consultant in patent interference cases and things. Redcom, for example, is suing another company on patent interference. Many times the patents are so broad that they don't even tell you what commercial product is. They don't want you to know. However, prior to my getting interested in this sort of thing, and looking at other parts of the telecommunications industry and seeing how poor we were in switching, I found what was missing. There were principles here, ways of classifying this knowledge, and people weren't doing that. Nobody was looking at all these different inventions and saying, “Now how would you classify all this?” I tried to do that. That was one of the things I felt was my contribution. Trying to say, “What are the principles of switching? Are there principles?" Of course, we did eventually do that, and we came up with a course that we first taught at Bell Laboratories on switching systems and circuits.

### Foundational texts on switching and networks

Nebeker:

It raises the question of texts--the texts that define a field.

Joel:

Yes, right. Now over in Germany, they had wonderful texts.

Nebeker:

On switching?

Joel:

Yes. Switching and networks.

Nebeker:

Beginning when?

Joel:

I would say beginning in the twenties, right after World War I. They had terrific texts, and they taught switching there. You could go to the technical high school and you could learn about it. Same with Austria, too. All of those countries in that part of Europe. You could learn it in Poland. In England if you were a technician, not an engineer, you could learn about switching because they had open civil service exams for people who maintained switching systems, and for those there were texts that were put out so that you could learn how. So you could take the test. And those were great things to--they were terrific, because in those places people would ask, “How does this such-and-such type of selector work, and what are its features?” And you had to know all that stuff. Now in this country we didn't do anything like that. The union people and the craft people who worked for the telephone companies didn't have to take exams. They did have, however, in- house courses. Those were not available to just anybody. Some of these hour courses were taught at Bell Laboratories at the time. But I couldn't get my hands on them. On the other hand, I did manage at one time to get a nice thin brochure on the panel dial system, and it had more darned information in it than anything I had ever had before. I met the guy who wrote the thing at New York Telephone Company.

Nebeker:

But what about texts in this country? When did the first real text on switching appear?

Joel:

Well, the watershed text in this country came out of the course that we taught at Bell Laboratories. It was not on switching systems, however. It was on switching circuits. It was called The Design of Switching Circuits, by Keister, Ritchie and Washburn. It became a very famous text, part of the Bell Labs text series. Bell Labs had a whole series of texts, and this was the only one on switching. In fact Ritchie went up to MIT and taught it as a visiting professor there one year. Then some people up at MIT got the idea that there was something here, and they started putting their own course together on logic circuits. Logic design became an "in" thing at schools. After all, that came out of switching circuitry, or logic circuitry. As for switching systems, in this country, I guess there were no good books on them, just descriptions of individual systems. And then there were transactions and meetings. I don't know much about textbooks.

Nebeker:

In some fields there's a tradition for such handbooks, the Antenna Engineers' Handbook and so on. Has there ever been such a thing in telephone switching?

Joel:

No, not in telephone switching. There are some individual chapters in communication handbooks. Nowadays there are some pretty good places where you can look up things that people have written about the crossbar system and even about panel systems, but nothing detailed. But with electronic switching things have changed quite a bit because people started writing articles about their latest ideas on switching. Everybody had different ideas about switching, and they all wrote them up and gave papers on them. So it changed, but as far as texts go in this country, there was nothing really to point to except the Keister, Ritchie, Washburn book. Of course, there are a lot of texts now, books on telephone systems and so on..

### Training and education in switching

Nebeker:

On the related issue of the formal education or training in the field, you'd mentioned the German-speaking countries in the twenties--

Joel:

Yes, they had formal people. It was just ideal. They could go and study telecommunications, and they could study and learn about switching as part of it, or in separate courses in fact. Nowadays, there's even more taught in universities over there, but not here. Even as of this time, there are no places that I know of in this country where you can go and get a really good course and learn a lot about switching. There's a course at Louisiana State University, there's a course in North Carolina someplace, and a few other places. You know, when you get around to looking at the curriculum you see courses in telephones that include switching, but they're telling you more about the politics, the competitive aspects, rather than about how the systems work. What the technology is and that sort of thing is missing.

Nebeker:

So there's never been much formal training?

Joel:

Not in this country. We've been very, very backward here. The main reason I guess is that we just haven't got any professors, any teachers, that are interested in that sort of thing. And we don't have very many coming out of industry and going into teaching. There's a few, including a fellow who left Bell Laboratories recently who's out at the University of Pittsburgh, but he's not actually teaching switching. He does have a great deal of interest in it.

Nebeker:

I think when I was being shown around the Rochester Institute of Technology, they had a lab that had a lot of switching equipment in it.

Joel:

Well, that's probably stuff left by Stromberg because they were there, and they left town. Maybe when they left town, they left stuff there, but I don't know of any teacher there.

Nebeker:

In European countries they are still teaching today?

Joel:

Today, yes. In fact, in the USSR before the breakup, there was--and there must still be--institutes where these things are taught. I remember one institute where a professor had 300 students just learning about the design of controls of electronic switching systems. He had notes that covered all the switching systems of the world and details about them and so forth. Boy, that was great! I sure wish I could have taken courses like that when I was in school.

Nebeker:

So in a way Erickson and Siemens have the advantage that they can hire people who've really got some training in it.

Joel:

Oh, yes. Yes. It works two ways, too. They also send people out to various underdeveloped countries, where they need good technical help. They send these people out to teach other people in these countries. Of course they promote their own products at the same time. [Laughter]

### Switching in power engineering and communications systems

Nebeker:

There's also switching in control engineering and power engineering of various kinds. What about these relations?

Joel:

Well, there are two fields that I know of that involve the communication kind of switching associated with these industries. Now of course these industries have their own switching besides. For instance, load switching for power companies is a whole separate field.

Nebeker:

And that has no relationship at all?

Joel:

Has no relationship at all. It's completely separate. On the other hand, there are communications systems associated with these switches. In these, you have usually centralized points where you then can push buttons on panels and operate switches that are sometimes hundreds of miles away. So that's remote control of switches. Companies who used to be in the telephone switching business, like North Electric, which was an independent manufacturer of switching equipment, made so-called supervisory systems, for power companies. Another breakout of this industry was totalizer boards for stock exchanges so the numbers change for the stock, and for racetracks.

Nebeker:

Is that right?

Joel:

For the totalizing of racetracks. They all use the same technology and switching systems. The last important area which also uses the same kind of technology and the same ideas in the design is railway signaling. For the airlines it's quite different in that it is all radar-based. But the railroad signaling is all based on relays and a lot of these circuits.

Nebeker:

So there are not only the ideas but the actual switches in common.

Joel:

Yes. Of course they're not quite the same because most of it's remote. Both the supervisory, for power, and the railway signaling, for transport, and telemetering in general, are more or less for remote control. Because after all, switching involves signaling. In other words, when you pick up your telephone and have to dial a number, you're signaling the central office what number you want. And so that's sending signals back and forth. All that started with telephone switching. Most of these systems use some form of signaling to control things remotely. After all, that's all an automatic switching system is. A telephone switching system is something you're controlling remotely.

Nebeker:

And North Electric has moved a lot of that?

Joel:

They did a lot of the supervisory systems for power companies.

Nebeker:

It's that kind of technology transfer that's interesting, the way fields are interconnected.

Joel:

Yes, well, there’s those three things: the totalizer boards and stock exchange stuff, the supervisory systems, and the railway signaling. And then of course in those days the computers--when they had only the cards, the IBM and Remington Rand cards, most of the stuff was mechanical. The cards ran through these machines and were sorted, and there were logic circuits associated with these that were switching-type circuits. But it's very trivial types of things. Nothing very complicated.

Nebeker:

Yes. There was not an intimate relationship between the two.

Joel:

No. It was not until the computer came along, post-World War II. You know, the computer history started with electromechanical devices, too. Now I'm not talking about way, way back with Boole or somebody there. But later on with the electromechanicals, Bell Labs and Aiken and the Mark I and II, sequential calculator, and those. But they rapidly, right after the war and even toward the end of World War II, converted to electronics.

### Professional societies, institutions, and government in telephone switching

Nebeker:

What about an overview of the institutions involved over time? We've touched on this in connection with some other things. There are the producers of equipment, the users in the sense of the telephone services, the PTT's and private companies in this country and so on. But other kinds of institutions. There are the professional societies, the AIEE in this country.

Joel:

We started, I think, in 1946 as Switching Committee of the AIEE. That was the first time that we tried to get any kind of industry interest to exchange viewpoints, and try to get papers.

Nebeker:

Do you know if there was such a thing in other countries before then? Where telephone engineers were getting together?

Joel:

I don't know of anything, but there may well have been. I've been in correspondence over the years with guys in Japan that worked on this prior to the war and who referred me to papers that appeared in journals there. In fact, there were people in Japan who were pushing this logic circuit business that we finally came up with the textbook on. They had written articles about logic circuits, the kind of thing that Shannon did, applying Boolean algebra to circuits. There were some people working on that in Japan at the time, too, around the late thirties. So there probably were professional societies at work in other places, but I don't know too much about it.

Nebeker:

What's the history in this country? You had this committee that was organized, then what's happened since then?

Joel:

Well, since then the Switching Committee's been a major item in the Communications Society.

Nebeker:

It's still called that?

Joel:

Yes. Well, it started when it was AIEE, and then when IEEE took over it became the Switching Committee in the Communications Society. So it's been that way ever since.

Nebeker:

What other professional organizations exist?

Joel:

Well, of course we have this International Switching Symposium, which is held every two or three years around the world.

Nebeker:

Is there a committee that runs that?

Joel:

Yes. There's an ad hoc committee of members. We try to keep the manufacturers out of this, so we try to have the telephone administrations in the various countries have members on this ad hoc board. They decide where the meetings are going to be held in the future, and they also make sure that in their particular country there is a strong committee set up to be the host of the meeting. For example, we have now planned out the ISS for next year--or maybe in 1996--in Berlin. In fact, we have it planned all the way up to the year 2000.

Nebeker:

Every other year?

Joel:

Well, it's supposed to be every three years. But whenever the Germans and the Japanese get together, it's every two years. [Laughter] Don't ask me why. That just seems to be the pattern.

Nebeker:

But it's just an ad hoc committee each time? Is there an international organization of some kind?

Joel:

No. Have you ever heard of ISLS? The International Subscriber Loops & Services? They have a more formal international group. But the ISS international group is just a committee that we formed back in 1972, and it's been running along pretty well ever since. We keep changing the members. I'm ex officio member for life or something. They gave me this since I'm the father of the thing.  [Shows something to Nebeker.]

Nebeker:

"The Father of ISS." I see.

Joel:

But at any rate it has percolated along pretty well up to now. I worry a little about the future, but at the present it's not very formal. It's an organization where everybody recognizes it's a good idea to have these meetings every three years.

Nebeker:

I suppose that within other professional societies, the IEE in England or the Japanese, they have specialist groups?

Joel:

Yes, they have their own. The same kind of thing we have. And then there are a lot of other meetings that have sessions on switching. For instance, there's the Zurich meeting--Zurich Digital? It's a prestigious meeting that's held every two or three years in Zurich that has sessions on switching. Certainly all these other professional societies around the world have switching groups or something, and usually they're the ones that bid to host ISS meetings. They decide that they would like to have it. They think it's time. We look at it. We say, “What've you got that's new? Why are we going to have it, say, in Italy? What are you going to be able to show us there?” And so forth. When it is decided, their ACT, or whatever the professional society is called there, would then take it over. They have their own committees and whatnot. We let the country who runs it keep any money that they get from it, so that it's attractive to them in most cases.

Nebeker:

Okay. So that's sort of the situation as far as the engineers getting together. What about other sorts of institutions? I've heard of the International Telecommunications Union. Tell me about this.

Joel:

Now, that's different. Now, of course, they have publications of their own. They have standards publications which are published after much deliberation and the consensus among all the people in the field.

Nebeker:

What is the nature of that organization?

Joel:

Well, it's supported by the United Nations. It's got a new name now. By the way, I was the second and the last one to receive their award. They haven't given one out since. That was in 1983. The problem they face is to find somebody that they'll all agree on. There's over a hundred and eighty nations with their own members. Believe it or not, the year I was awarded this thing, [Chuckling] I was not the one that was suggested by the United States. Some other country suggested me, and the United States had somebody else's name in. I guess every big country with communications had their own favorite son in there but somebody else proposed me, and everybody went along with it, so that's how I got it. But they haven't been able to agree since then. [Laughter] But their interests are primarily in fostering communications and in making sure that communications among nations is properly compatible and continues. They deal with tariffs, the exchange of money between countries, and they deal with the technical things which have to do with standards for signaling.

Nebeker:

Do they set standards for switching as well?

Joel:

No. They put out a few publications for developing countries, to give them recommendations for what they should be looking for in a switching system. Maybe something about the current technology at the time it was written. But those are just advisories. Their main job is to come up with standards for signaling for the interfacing of communications between countries.

Nebeker:

I wondered if the ITU in Genoa, or whatever it's now called, had either some archives there or if there's some senior person there I might talk to. Do you know?

Joel:

I know that they have archives from way back, from before the United Nations, when it was the League of Nations.

Nebeker:

Yes. That was based in Geneva.

Joel:

Ever since countries have communicated with one another, they've recognized the need for somebody to do this, and they've been doing it for a long time. A great person to talk to if you're over there would be my co-author, Chapuis. He knows his way around ITU because he was the senior counselor there for twenty years or so. He knows people there.

Nebeker:

I'm glad to hear they have records that go way back.

Joel:

I don't know about the early telegraph days, but I'm sure they go back a ways. It certainly goes back to the League of Nations. How much before the League of Nations, I don't know. But they must have had something. You know even in the very early days they still had telegraph between countries.

Nebeker:

Okay. What other institutions that have to do with telephone switching can you think of? And there's government regulations.

Joel:

We're running into all kinds of things now. We're running into forums of vendors who would like to make our equipment and are trying to set their own standards. We have a real hodgepodge. We also have standards within countries. The European Community have their own standards, and in North American we have the T-1 Committee that's having a name change soon, too. We have people like Bellcore that are out setting standards. There's a lot of people with their oars in this thing, who show their interest primarily by trying to work out interfaces with equipment and definitions of services and things of that kind.

Nebeker:

The old style of the history of technology applied to telephone switching would be that you'd look at, the manual systems, the semiautomatic, and you'd just look. That's a very important part of history, but just one part. Then there's the institutional history, the people who look at the institutions having to do with a given type of technology. So what I'm asking here is, if we were doing that kind of history, have we named all the key players?

Joel:

No, we left out one big one. The military. The military, at least in the switching area and telecommunications in general, has always tried to lead the parade. As soon as some new technology is even thought about, they want to get their oar in. At least in this country it's been that way--spend the money and be out on the leading edge, at least in experiments and possible products. In other words, if you're looking into the history of something in communications nowadays, it would pay always to say, “What were the military doing at this time? What contracts had they given out? What were their own labs doing?”

Nebeker:

And are they that important in switching?

Joel:No, they haven't made any great contributions. But let's put it this way. There have been things that they've supported that were early switching systems of a kind that later perhaps became important. For example, we're working hard nowadays to combine voice and data. Back in the 1960's they had a contract that they gave out to Bell Laboratories and RCA and GTE, and all of them contributed to this thing, to make a switching system that did voice and data. It didn't do it anything like it does today because they didn't have digital voice at that time. But they were always giving out contracts to look at, to explore some new ideas when they come along. GTE came out with a thing they called "burst switching," so right away they got a big contract from the government to go out and play around with it. I think those days are over now, but that's the way it was in the past. The military people had a chance, whenever something new came along in switching, of at least trying it out.

Nebeker:

Okay. In Europe clearly the major institution in most countries has been the PTT that's run things.

Joel:

Run things and pretty much dominated the vendors in those countries, and set all the rules for them, and allocated markets, and so on. Even financed them. There's all kinds of that thing going on. That's why it's so hard to get rid of them--because it's so entrenched in the country.

### Quantification and theory in switching

Nebeker:

Another topic that's in a way peripheral to this straight line of hardware history that I'm referring to is the advances in understanding and the application of Boolean algebra and Shannon theory, or the development of traffic theory, this kind of statistical study. I wonder if we could just discuss get the main advances in this.

Joel:

We almost alluded to it before, but the difference between transmission and switching is that, as I say, there are some underpinnings in transmission. First of all, the reason there's underpinnings is that you can measure things in transmission. There's nothing you can measure except traffic perhaps in switching. You can measure, you know, how many calls went through the central office today or in the busiest hour or whatever it is, and you can come up with theoretical approximations and so forth, but except for that aspect of it, you can't measure any particular lash-up of the functional boxes of a switching system like you can the functional boxes that you put into a transmission system. For that reason, switching has always been an art rather than a science, as I see it. It's something that bothers me, but up to now I've seen no breakthroughs that are going to change this. Even people who look at computer architectures, who have the same objective in mind, and I don't think they've made any progress either.

Nebeker:

Yes, that's very interesting that you tie it closely to quantification, being able to quantify the things that you're dealing with.

Joel:

I think so. The only quantification today in a switching system is the dollars, and that doesn't mean much because there are loss-leaders and all that kind of stuff. You can't really tell that just because this central office costs $20 million and here's one next to it cost$15 million, that one is better or anything. You can't measure them that way.

Nebeker:

But to take a few minutes, if we could, on the one area where it is measurable, traffic theory. I know that there were people like M.C. Riorty at AT&T early on who applied probability theory, but I was very interested to read something about a Danish engineer, A.K. Ehrling.

Joel:

Ehrling was very important. Also in England there was a fellow by the name of Dell. There were a number of them in England, too, and a whole bunch of people in the United States.

Nebeker:

And what they're doing is proposing probabilistic models of telephone traffic, and saying we assume a Poisson distribution of incoming calls, and what you do with the length of time of each call.

Joel:

And how you deal with it. And what is the distribution of holding times? And what happens to calls that don't go through? They die or come back, and all that.

Nebeker:

Right. Now, what has the importance of that type of theory been in telephone switching?

Joel:

The applications have been great because people have used this theory to develop models and also engineering tables for the quantities of equipment needed in a switching system.

Nebeker:

What do you mean by engineering tables?

Joel:

To quantify what you have to buy. In other words, I'm going to build a switching system that serves South Orange. How many of the various pieces of the switch do I need? And the quantities that I need of all the different parts are based on engineering tables that come out of theory calculations.

Nebeker:

I see. So they could say that if you put in so many trunk lines, you're going to get this level of functionality?

Joel:

Yes, this quality of service.

Nebeker:

I see. So that part was quantified earlier?

Joel:

Oh, yes. Early on. Of course, there's still lots of argument as to what quality of service you want to shoot for. One story I always tell is that the people in the government who are on the Federal Telephone System, which is the lease service of the government for their telephone communications, came to me and said, "You people are supplying this service that's awful." I said, "It's not our service that's awful. It's the level of service that the government wants to pay for that's bad. Not our service that's bad." That goes on all over the country. You go around the world in various places, like the U.K. and so forth, you'll find the level of service isn't as good as it is here. They engineer it to a lower standard, the quantities they want to provide. So that determines what the cost of some of these things are.

Nebeker:

But has this traffic theory been generally reliable?

Joel:

There are many different hypotheses that go into these, and people test them all against the real world. The real problem has been in telephone switching, as compared with other applications of tele-traffic theory. You can apply tele-traffic theory to railroads, to highway traffic, and to barbershops or anything you want, where there's queues. The thing that's been difficult in the tele-traffic theory, why it's important, is because tele-traffic involves large numbers. A telephone office, for example, typically serves 10,000 to 30,000 customers. So you're dealing with big numbers, and it's hard to get data. The data that many people use today in describing the tele-traffic theory go back to the twenties when people could make some studies with simple electromechanical devices. Today people don't make these studies anymore. What are the distributions of current-holding times? We don't really know very well.

Nebeker:

That's surprising. You'd think it would be easier today to gather such data.

Joel:

Yes. And we're not doing it. But we can't separate out what is a telephone call and what is a data call. If some guy has a modem, and he sets up a connection and he's up all night, you can't tell that from a telephone call.

Nebeker:

But the theory is reasonably good?

Joel:

Well the queuing theories and the tele-traffic theories and so forth have stood up very well. One of the things I'm concerned about right now is that we're entering a new era with some new types of switches, which are known as ATM, and some new ideas about shared bandwidth switches and things of that kind, mixing video and imaging and voice and data all on one switch. People have devoted a lot of time to writing papers about these mixes and what they mean from a tele-traffic theory point of view. To some extent I think they're reinventing the wheel, but I'm not sure. In other words, they're going over and around it; it's been well-traveled in the past. In any case, there's always a need to apply this theory.

Nebeker:

But it has been important in these decisions about how big a system you need?

Joel:

Yes. How big and what quality service you want. The other aspect has been forecasting. That's been very poor too, because the political decisions so much changed the forecasting that it doesn't mean much anymore.

Nebeker:

I can imagine several different ways one could do this traffic analysis. One could do it entirely on theoretical assumptions, based on some data in the past that...

Joel:

Yes, that's what they do.

Nebeker:

One could also do it with what's very popular in some areas of science, these Monte-Carlo simulations where you don't try to predict, but you try to just put the assumptions into a computer model.

Joel:

Yes. We've done an awful lot of that in the switching business. In fact, this same crowd, this Keister, Ritchie, and Washburn--and now there's another guy by the name of Frost--did a very extensive Monte-Carlo experiment of the details of a switching system.

Nebeker:

When did they do that?

Joel:

They did that around 1953. It was all written up in the Bell System Technical Journal.

Nebeker:

They must have had a computer.

Joel:

They had their own makeshift computer. It was not a general-purpose computer. It was computer model of what the actual central office was like. All the detailed decisions made within the central office were made in this Monte-Carlo thing. We have a very poor record when it comes to looking at Monte-Carlo stuff as applied to software problems. We've never done very well in software. If we try to predict, for example, the call-carrying capacity of an electronic switching system with stored-program control, it has not been very successful.

### Traffic generators and telephone traffic measurement

Nebeker:

I was interested when I was talking with Klaus Gueldenpfennig that one of their successes has been manufacturing these traffic generators.

Joel:

Yes. He's trying to find out on a practical system how much traffic.

Nebeker:

And that's another way to do it.

Joel:

Well, that's all right for his system because they're small systems.

Nebeker:

But you can't do that kind of thing for large system?

Joel:

It's hard to do in a big system, especially one with a large variety of services. That's something that really hangs it up. If in every switch system the only services were plain old telephone services called "pots", that would be fine, but what's happened nowadays is that so many varieties of things are going on inside of a switch, that you can no longer just make claims for what happens to the "pots" calls. You've got to know what the mix of calls are. And whatever you select as a mix will come up with one answer; a different mix will come up with a completely different answer.

Nebeker:

And the difficulty is knowing in advance what kind of a mix to use?

Joel:

Yes.

Nebeker:

But you could conceivably construct a traffic generator that gives you the mix you want?

Joel:

You could, but it's too hard to do. There are too many mixes. Too many variants.

Nebeker:

We have there several different approaches to coming up with the size and characteristics of a switching system you want. It can be theoretical entirely, or Monte Carlo, or it can be that you construct something and hook it up to some traffic generator.

Joel:

And by the way, in some of the switching systems today, they actually build in generators within the software.

Nebeker:

Is that right?

Joel:

To test them, yes. But it's still not satisfactory. For example, a vendor would claim, “I've got a switching system here that can serve a hundred thousand lines.” And so you say, “Well yes, but how many calls can it handle with a hundred thousand lines?” Well, he makes claims that it'll serve 1.2 million busy-hour calls or something, but nobody knows whether it will really do that or not. It depends upon the mix. And in fact telephone companies have really gotten stuck. No question about it. Companies have bought systems from vendors who make claims, and then they go out and buy and install millions of dollars' worth of stuff, and it don't work--it doesn't carry the required load, so they have to de-load it. They have to go back and cut over to a new central office. They have to uncut it.

Nebeker:

Well, I can see that it's not easy to measure telephone traffic. But has there been a big effort to reach agreement on this?

Joel:

Yes. Bellcore has tried to define better what some telephone companies should be looking for in these switching systems, and what criteria the vendor are supposed to use in assuring the companies that it does meet these standards, but it's still not that accurate.

Nebeker:

I know that computer vendors argue about computational speed, they've made real efforts to specify the type of problem that you test on.

Joel:

Right. Exactly. And that's exactly the same problem you have in switching.

Nebeker:

But it's still difficult to know what a vendor means when he claims a certain thing.

Joel:

Oh, yes. Absolutely.

### Shannon's information theory

Nebeker:

Maybe there were some other things on these advances in understanding. Was Shannon's information theory important? I know its important in some transmission problems.

Joel:

Oh, very important.

Nebeker:

Has it been important at all in switching?

Joel:

No. Shannon went off, after he came up with the information theory, and took a look at memory requirements in a switching system and tried to make some implications of that from his information theory. But, no, I would say nothing ever important came out of information theory as applied to switching. You know, there's nothing like entropy in switching that really involves that sort of thing. See, switching is a lot of logical complexity. We're just getting to the point where people are starting to write about complexity and chaos. Some of that is going to apply more, I think, to some of these things in the future.

Nebeker:

I came across this phrase, "switching algebra." Do you know what that is?

Joel:

Well, that's the same thing that really Shannon invented. His master's thesis was based on Boolean algebra applied to switching and circuit design, and that's the same thing. As I mentioned earlier, the book that was put out by Bell Laboratories on design of switching circuits used his Boolean algebra and extensions of it. Later on, other authors took on the same thing for logic circuit design in general, and that became switching algebra they talked about instead of Boolean algebra. One derives from the other.

Nebeker:

I see. I think your books do a very good job of telling the story of the technological advances.

Joel:

Oh, yes. In the Bell System Journal at least. And the other one was the one we wrote about the electronic switching history. Then Chapuis wrote the earlier one about the electromechanical switching history, but he just talked about the major developments in electromechanical switching.

### Technologies that influenced switching systems

Nebeker:

With any technical area, a lot of the advance is kind of internally driven--one thing leads to another--so advances are made, then you improve on things. Occasionally a line of technological development will get buffeted from the outside by some external developments such as deregulation...or as you mentioned, the interconnection requirement in the early seventies. Another kind of external buffeting can be when some new technology, such as the integrated circuit which wasn't developed within this line of technology, finds application and pushes the old one aside. Now, this is a tall order, but I'm wondering if we can enumerate the most important of these kind of external influences on telephone switching.

Joel:

First of all, you're talking about technical, right? How about social influences?

Nebeker:

Well, some of these changes are reflected in this technological development itself.

Joel:

In the technical end, something fairly recent, which we talked about earlier, was this push to time-division digital switching that occurred in the local plants in the early eighties, and which is still going on, although all it did was eliminate the electromechanical switching that had been there for years and years. We now have either time-division digital switching or we have the space division--so-called analog or digital in the trade press. That's one kind of push that's been going on in the technical side.

Nebeker:

But that's more internal, isn't it?

Joel:

Internal to switching? Don't forget the drive there is to get a better interface, more synergy, with digital transmission. And the synergy, of course, is to eventually get digital right from the customer's premise to the other guy's premise. That's the drive today. Now, we haven't got that last mile in there yet. Transmission's always driving it, always driving the switching people. Transmission comes first, they come up with fiber. What can the switching people do about it? The switching people are out now trying to design switches with optics, photonic switching rather than electronic switching.

Nebeker:

There's a prospect of actually doing that?

Joel:

Oh, yes. A very good prospect. But it couldn't come first. It couldn't have come before the transmission.

Nebeker:

I see. So that's a very important generalization then, that transmission changes have stimulated and caused changes in switches.

### Social impact of telephone switching; public demands influenced technology

Joel:

Yes. And the other thing is that the drive to automatic switching from manual switching came about because people wanted to have telephone service twenty-four hours a day, rather than just when the switchboards could be operated.

Nebeker:

Right. Also it must have had to do with simply the volume of the telephone traffic.

Joel:

Audio File
MP3 Audio
(163 - joel - clip 2.mp3)

You know, people back in those days tried to project how much of the labor force of women, if they were women, would be doing operating. It has been women almost since the beginning. In the very beginning were boys, but they got rid of those because they claimed they had profane language. [Laughter] I can't believe that, but that's the apocryphal tale. Nevertheless, the demand for operators, as you projected future traffic, started to eat up more and more of the female population. More than you would have, even if they were all on the telephone! [Laughter] Those studies have never been very much agreed to by sociologists.

Nebeker:

Do you think it was more the giving of twenty-four-hour service that was important?

Joel:

I think in the very early days that was it. Of course you've probably heard the other story, which I don't believe either, the apocryphal story about Strowger and the fact that he was an undertaker, and that he claimed that somebody at the switchboard was giving his calls to his competitor because she was the niece or the daughter of a competitor. So he was losing the calls. Therefore he was going to get rid of the operators... [Laughter] That's the other apocryphal story. But in any case, the impact has been social all along. The switching systems were influenced greatly in the very beginning by the fact that we didn't think people could remember numbers and dial numbers accurately, so they had to put letters on the dial. And people had to be able to say MAINE, and dial M-A-I-N-E, or whatever it was. When we started giving that up, people were objecting. In fact, we're right now trying to make changes in the dial system, the numbering plan of the United States, North America, starting in 1995, because we've used up all the area codes. In 1995 the first area code's going to be given out that doesn't have a 0 or 1 in the second digit. All of a sudden now in the woodwork we've got a lot of people saying, “Now, to do this, and do it easily, you have to say that you always dial 1 or a 0 ahead of a ten-digit number, to make sure that you know it's going to be an area code rather than an office code. “Well, we've got people coming out of the woodwork now saying, we want to have the ability to dial 1 to tell us it's a toll call, in which case for example here in northern New Jersey, you can dial all over the 201 area or where you are in the 908 area, you can dial all over, and a lot of those calls are toll calls. In previous technologies you could dial 1--in fact you did have to dial 1--to get out of the local and into the toll, even though it was a seven-digit number. Now, we don't want to do that in the future because we want to reserve the 1 only for the ten digits, and yet you've got political people trying to change this.

Nebeker:

I remember these discussions when New York City had to have more than one area code.

Joel:

### Quality and service in switching history

Nebeker:

That makes me think of another way of looking at this whole evolution. You used the phrase "plain old telephone service." If we think about somebody in 1905, say, who was lucky enough to have a phone then--not everyone did by any means--and then someone today, I suppose the quality of the transmission is better today. Is that right?

Joel:

Oh, it's much, much, much better.

Nebeker:

Just the fidelity and lack of static and so on.

Joel:

Sure. Much of it's due to switching, of course, because in those days when we had switches, we had base-metal contacts that would rub against one another and put grind noises in the connection. That's not the only thing, because we had to improve the transmission so that the streetcar lines and so on wouldn't introduce signals into the telephone lines. But the transmission people learned how to do that, too, how to get around that.

Nebeker:

Probably the most striking thing that switching has meant to people is that you can pretty much reach anybody in the world, anybody in the rich, industrialized world, with a dozen or so numbers.

Joel:

Yes. In other words, what you're really saying is this: That plain old telephone service means different things to different generations of people.

Nebeker:

Right. But I was going to get to more than that, and say that those two things are very important, and especially, I think, the second one: I mean it's really remarkable that you can think, say, “What about my friend in Barcelona?” A minute later you can be talking to him. That's a remarkable thing, but then there have been, it seems to me, especially in the last decade or so, all these special services that have come to the telephone. Is that the way it looks to somebody in the industry? That it's suddenly changed?

Joel:

Nebeker:

So the U.S. has been a leader in all these services as well?

Joel:

Well, I don't know if you want to call it leader, but we've been different. We have politicians that go abroad, and they say, “Why can't our system be as simple as that? Why do we have this?" Some of them think that that's better. But I know here, even with all our differences in service, we have a hard time getting it to the public, just like you do with anything I guess. We have a situation now in Florida where we're trying to introduce measured local service. Not just Florida, but around the country in general. In fact, the politicians want us to do this, so that every service is measured, not just long-distance service. In many places there are tariffs for that, but we haven't done it. However we've always built into our switching systems the ability to do this.

Nebeker:

Oh, that's clever!

Joel:

Yes. Ever since the panel systems. We also had to deal with coin calls, and we dealt with coin calls in a different way here than anywhere else. We enabled you to dial 9-1-1 and things like that without putting a coin in, and eventually get 800 service without a coin. All these things, by the way, include switches, and we've expanded. But when I go thinking back into the twenties, into the early panel system, they had all these party-lines, and had all the coin lines. You had to be open to PBX things built into the system. And they even had things for calling the telephone company without getting charged, things that today maybe we don't think too much of, but that list has kept growing and growing and growing. It just keeps growing. Every year another anywhere from ten to 100, depending on how you count them, new things keep getting added as services, and they all go into the switch. It doesn't affect transmission; it's just the switch that has these things.

Nebeker:

I can imagine this has really been a challenge to switching.

Joel:

Well, we're getting indigestion. [Laughter] We're getting indigestion in the software, that's the problem. But nevertheless services keep increasing. We talked about Centrex. Centrex service is not just one simple thing. Centrex service is all the things you do in a PBX, but done in the central office, and there's a lot more than that. We talk about services; I'm sure that you know the term "telemarketing." Behind telemarketing is what we in the telephone business call "automatic call distributors." And so we have automatic call distributing go out and buy one and serve a big business, that's in the telemarketing business. Or you can use the central office and let the central office do your automatic call distributing, but it takes an awful lot of doing to do that.

Nebeker:

Everybody knows that people use the phone more than they used to, and that phone is doing a lot more different things.

Joel:

And there are lots of specialized switching systems; we didn't talk about most of them. We talked about PBX's, but if you go down into the Wall Street financial area and my gosh, the lash-ups that they have there are unbelievable. You push one button, and you get a gang of the particular instantaneous connections for a bunch of people who bid on certain kinds of bonds. You push a different button, and you get a different clique of people who bid on different kinds of bonds. All are being simultaneously run, being simultaneously talked to, and conferenced. They've got some lash-ups that are out of this world. Some of them are stored-program controlled so you can change instantaneously how you want to do them.

Nebeker:

So when a person looks at the almost 120 years of telephone service, they see a tremendous increase in number of calls made, time using the phone, and different services, an increase in services.

Joel:

Oh, yes. The variety of services.

Nebeker:

I think one of the most remarkable things for this type of engineering is that you have to have compatibility with every change compatible with the previous one.

Joel:

Yes. It's hard to get rid of these things. I mean, party lines still exist, although there aren't very many party lines left.

Nebeker:

There are some left?

Joel:

Oh, yes. In my day when we talked about party lines, it was possible to have twenty parties on a line, often a farmers' thing, with a go-code ringing. We don't have that anymore, but there are still party lines around. Mostly two-party lines with selected ringing so you don't know the other party's being rung. It's hard to get rid of. So we have these odd services in our telephone systems, that customers still can subscribe to, and then we have other things that are built into the switch technically that perpetuate themselves because that's how they're become ingrained in the system, like some form of signaling that we don't need anymore, but it's in there because they were built in originally and are carrying more modern services.

Nebeker:

If you could start from scratch with a new switching system with the new technology...?

Joel:

Well, yes. Unfortunately, that doesn't get rid of all these things. You start with the new technologies today, but we've still going to do the party line because we haven't got rid of it. But we can get rid of some. For example, today most of the modern thinking is that the signaling should be out of the normal talking path that you use between the two parties, so that when you generate a signal, the connection is set up independently by a separate network of signaling nodes and signaling messages that are sent all over. That's the way that we're doing it today in the toll networks, in the long-distance networks. Now, we're trying to introduce that gradually into all the local networks. It's very expensive, and there's no compensation coming in to pay for it, unlike the 800 numbers and the credit card. We have the Caller I.D., that you can have by your telephone, which shows who's calling you and that kind of stuff. It's gradually coming in, but nevertheless, despite the fact that everything's going to this so-called common-channel signaling network, we still have plenty of places in the system where we're sending information over the telephone lines to address the calls. We're still doing it in many places. It's open to fraud and we encounter all kinds of other problems. You can't do anything about getting rid of it that easily.

Nebeker:

Yes. It's a matter of rebuilding a ship at sea plank by plank. You've got to stay afloat.

Joel:

I never heard that. That's a very good analogy.

Nebeker:

A computer designer can say, "Okay, I'm going to do a new computer." But if you've got this, as you call it, distributed computer, and it's got to keep running all the time, then you can't change it all at once.

Joel:

That's right. You've got to change pieces at a time, and we're not allowed to do that. First of all the regulatory situation in this country won't let you change it all out at once if you could, even if you had the capital, and you couldn't get the capital. But secondly, if you could do it within a year, let's say, you wouldn't because the depreciation rates on the old equipment have been set by regulatory degree, and if it's not fully depreciated, you can't get rid of it because that means higher rates to the customer. So, we're tied up in that to a lot more than just technical things in the switching business. I think that's the message I'm trying to say here, that in the transmission business you get the idea that you have better connections today than you had back forty years ago, twenty years ago. The quality's better. You're talking to somebody in San Francisco, and it sounds like he's right in the same room with you. So things have really improved there, and you recognize it, and the customer hasn't had to do anything to get it. He got it for nothing, so to speak. The switching thing is different. The switching thing is influenced by what these customers are willing to buy and pay for as new services. You know that if he wants Touch-Tone, he wants push-buttons, he pays for it, and he gets push-buttons. Same thing with Call Waiting or Caller I.D. or whatever he wants. The other thing is, you don't know about what's going on behind the scenes, as the services improve, the number of people we have back there running these offices. We've got an office down the street here in South Orange, that's serving 50,000 lines. I don't think there's anybody in there today, and there's probably someone who comes in once a day to do something. Change a tape or something.

Nebeker:

That would be an interesting thing to look at because I know that a similar story is true of the electric power. There used to be people at the local distribution centers. That was replaced by automatic switching and fuses and circuit breakers and so on.

Joel:

We have the same thing here. So the reliability has gone up, and the number of people we need to run the show keeps going down, but the complexity is still there, and it goes on.

### Joel's education in switching

Nebeker:

Well, as I say, I tried to sketch out from your book mainly, the Bell Labs one, these stages in the techniques of switching. And I think it's probably best told there. Could you finish by talking about your education? How did you become involved in switching?

Joel:

I went to MIT and I worked toward a master's degree. As part of the master's degree program, you had to do two things: You had a master's degree seminar, and you had a master's degree thesis. In both of those I did something in connection with switching. My thesis was a theoretical piece that I did about switching organization and system architecture, but the one I wrote for the history of switching was more or less patent-based because that's what I had been studying. But I'd done a lot of literature searches, too. And I put out a book that was pretty thick, on the history of switching, and which I wrote in 1940.

Nebeker:

Is that right?

Joel:

Yes. At any rate, I produced it as a student. And I have a pretty good bibliography in there of stuff just about in that time frame. One major article on each major system around the world at that time. We're talking 1940 now. That's over fifty years old.