York...
As an aside, in some cases NYC had _8_ digit phone numbers, such as HOllis 5-10254. A manual exchange could have 10,500 numbers. Also, if the phone number had a party line suffix, eg. HOllis 5-9242J, the J was dialed, too. Literature on the panel exchange in NYC confirms the registers could hold and pass 8 digits. I don't know if that applied in other big cities.
When calling from a dial exchange to a manual exchange, the desired number appeared on a display panel in front of the manual operator. She merely plugged into that number.
Manual service continued in NYC until the early 1950s. A late 1940s telephone strike steeply reduced service to manual exchanges--the public was asked to make emergency calls only.
In the early 1950s, NYC subscribers could dial 'toll' calls in Long Island instead of going through the operator. Instead of being billed
10c or 15c for such calls on an itemized basis, the message unit meter was incremented accordingly for distance and time (the meters already existed to do that for local calls). Some subscribers complained about the loss of itemized billing, but it undoubtedly saved the phone company a lot of clerical processing.
What happened to these numbers when long distance direct dialing was implemented? Were they really all renumbered?
"The" panel exchange?
I don't disbelieve this but I would certainly like to see some references.
I remember reading about this many years ago, probably in Bell Labs Record at the time. The devices at the manual end offices were called Panel Call Indicators (PCI). Somewhere I remember one of the illustraions was of the ir use in New Orleans, in addition to NYC, of course.
I have no information on 8-digit numbers or party line suffixes, except to note that party line suffixes were very common in manual offices of all sizes, including magneto exchanges.
Wes Leatherock snipped-for-privacy@yahoo.com snipped-for-privacy@aol.com
DDD required that all subscribers have a uniform ten digit number-- three digit area code and seven digit local number. In 1950 a great many subscribers in smaller towns and cities had phone numers with three, four, five, or six digits, and all had to be converted to seven digits. This was a major undertaking that took years to complete.
Those subscribers with an eight digit number, be it 2L-6N or a party line suffix, also had to get new numbers.
The Bell Labs history volume 1 (op cit) mentions that manual exchanges had the capacity for 10,500 digits. Also, they show the interface panel for the inward B operator having eight digit capacity, either the leading 1 or a party-line suffix.
Around 1930, New York Telephone prepared a booklet from a lecture describing the panel exchange in use in NYC, and this also notes the eight digits. Unfortunately, I don't have a title nor know how it can be found on-line.
I've heard that NYC phone books circa 1950 had listings showing eight digits. Some day I hope to visit a NYC library and look up such a book (they do have them); I'm particularly curious to see the dialing instructions.
There were classified ads in the NYT of that era showed an eight digit number.
Not such a major undertaking. In roughly 1980 I lived in a rural area of New York State that still had 4 digit dialing. Extension to a _longer_ number can be done automatically -- it is fully deterministic, you are just adding predetermined digits.
My number was 687-XXXX for external purposes, but for in-town callers, simply XXXX would do. The "687", obviously, was just added onto the front when numbers went to 7 digits -- and there were towns near us with switches still configured to allow 3-digit dialing, which had had four digits prefixed to the front, too.
However, there is no such mapping from longer numbers (8 digits) onto shorter numbers (7 digits). Since New York City got only one area code in the original direct-dial plan, if there were really 8 digit numbers in circulation, some subscribers' numbers would have had to be completely changed, not just extended. This is what I am a little skeptical about.
P.S. Many times the "conversion" was accomplished at the same time a town was converted from manual to dial, which was a major effort in those years. Subscribers had to get used to a new phone number and billing routing records had to be all changed, but there was no equipment modification.
However, towns that were already dial had to be converted to be 7D.
In step-by-step offices this often meant adding a "digit absorbing" selector. Say a town had five digit numbers, 5-xxxx and 6-xxxx and the town was to become 345-xxxx and 346-xxxx. The switch would simply absorb the initial 3 and 4. This meant that people within the town would continue dialing five digits as they did before, only outsiders needed all seven. In many small towns in isolated areas five digit dialing continued well into the 1970s. It ended when more exchanges were added as population grew.
I don't have any statistics, but I suspect DDD was not offered very widely in the 1950s because the company was too busy expanding toll line capacity and converting local exchanges to dial. I _think_ the Bell System was about 50% dial at the end of WW II, not sure of the percentage in 1960. Anyway, the first step in providing DDD was to providing operator direct dialing, which sped up call handling. That meant adding toll trunk signalling systems for ringing and supervision which was an effort in itself. (That signalling protocol would be compromised later on by the "blue boxes".)
Even in the 1980s with DDD and TSP widespread, from time to time operators still 'built up' toll calls by calling intermediate toll centers and establishing the connection the old fashioned way. Likewise, they also wrote charge tickets instead of AMA.
Other local exchange items that were necessary to provide DDD were increasing the sender length in #1 crossbar and panel exchanges to hold 10 digits, providing an 'exit' link in step by step exchanges, and providing AMA (automatic message accounting) with automated or operator calling number identification (ANI or ONI). AMA was installed either in the local exchange or at a tandem or toll switching office.
Would anyone care to comment on how all this stuff is handled by modern exchanges today? Can a long distance company operator still "build up" a call manually? Are there still long distance company network managers watching their system for overloads and problems? Is there still a big AT&T network control center in Bedminster? Thanks.
[public replies, please]
P.S. Many times the "conversion" was accomplished at the same time a town was converted from manual to dial, which was a major effort in those years. Subscribers had to get used to a new phone number and billing routing records had to be all changed, but there was no equipment modification.
However, towns that were already dial had to be converted to be 7D.
In step-by-step offices this often meant adding a "digit absorbing" selector. Say a town had five digit numbers, 5-xxxx and 6-xxxx and the town was to become 345-xxxx and 346-xxxx. The switch would simply absorb the initial 3 and 4. This meant that people within the town would continue dialing five digits as they did before, only outsiders needed all seven. In many small towns in isolated areas five digit dialing continued well into the 1970s. It ended when more exchanges were added as population grew.
I don't have any statistics, but I suspect DDD was not offered very widely in the 1950s because the company was too busy expanding toll line capacity and converting local exchanges to dial. I _think_ the Bell System was about 50% dial at the end of WW II, not sure of the percentage in 1960. Anyway, the first step in providing DDD was to providing operator direct dialing, which sped up call handling. That meant adding toll trunk signalling systems for ringing and supervision which was an effort in itself. (That signalling protocol would be compromised later on by the "blue boxes".)
Even in the 1980s with DDD and TSP widespread, from time to time operators still 'built up' toll calls by calling intermediate toll centers and establishing the connection the old fashioned way. Likewise, they also wrote charge tickets instead of AMA.
Other local exchange items that were necessary to provide DDD were increasing the sender length in #1 crossbar and panel exchanges to hold 10 digits, providing an 'exit' link in step by step exchanges, and providing AMA (automatic message accounting) with automated or operator calling number identification (ANI or ONI). AMA was installed either in the local exchange or at a tandem or toll switching office.
Would anyone care to comment on how all this stuff is handled by modern exchanges today? Can a long distance company operator still "build up" a call manually? Are there still long distance company network managers watching their system for overloads and problems? Is there still a big AT&T network control center in Bedminster? Thanks.
[public replies, please]
Seven-digit numbers were required for operator toll dialing, too. Operator toll dialing was already widespread before the first DDD was offered to subscribers.
Generally the switch was configuired to accept the full seven-digit number on local calls, too, but was not required on local calls.
Wes Leatherock snipped-for-privacy@yahoo.com snipped-for-privacy@aol.com
I agree with Hancock4 -- it was indeed a major undertaking.
In the example you cite, the situation was straightforward: simply adding 687 worked because there were no local numbers beginning with 6,
7, or 8, and because 687 was available within the area code. The local switch was configured to distinguish the absorbed ("predetermined") digits from actual numbers by segregating them on separate levels. Without knowing the specifics of that particular switch, I can't specify the dialing plan, but it must have looked something like this:1 = access for non-local numbers + vertical service codes
2XXX = local number, blank level, or nearby community 3XXX = local number, blank level, or nearby community 4XXX = local number, blank level, or nearby community 5XXX = local number, blank level, or nearby community 6 = absorbed by digit absorber 7 = absorbed by digit absorber 8 = absorbed by digit absorber 9XXX = local number, blank level, or nearby community 0 = operator (or access code)The digit absorber in this case would have been type "AR" meaning "absorb repeatedly." You could have dialed 6, 7, or 8 repeatedly, in any order, with no affect. You could have dialed 687-0 and reached the operator. Or 8888877776666786886786780. Did you ever try it?
As I've noted in previous posts here in T-D, similar situations existed elsewhere. Example:
CARBONDALE, ILLINOIS, 1971 --
- Levels 3, 7, and 9: local 5-digit numbers
- Levels 4 and 5: repeatedly-absorbed ("AR") digits.
- Levels 6 and 8: NNXs in nearby communities.
- Level 2: unused.
But the situation was more complicated in larger communities where it was not possible to avoid conflicts by segregating functions on separate levels. In such cases, a different type of digit absorber, type "A", was used. The distinction:
A = The selector absorbs the specified digit once only; on the next digit, it "trunks on all levels." This digit must be dialed once (and only once) in order to reach certain specified second digits. However, it is absorbed (ignored) for any other second digit.
AR = The selector absorbs the specified digit repeatedly unless a digit has been absorbed previously on a level designated "A".
Two examples:
ANN ARBOR, MICHIGAN --
- Level 6: repeatedly-absorbed ("AR") digit.
- Level 8: absorbed-once ("A") digit -- see note below.
- Level 4: NNXs to nearby communities.
- Levels 7 and 9: unused.
Note how the "A" digit 8 was used to resolve conflicts:
- 8 followed by 6, 7, 8, or 9 was a local 5-digit number.
- 8 followed by 1, 2, 3, 4, 5, or 0 was absorbed and ignored.
CENTERVILLE, IOWA, 1975 --
- Levels 5 and 8: repeatedly-absorbed ("AR") initial digits.
- Level 4: NNXs to nearby communities (plus one located in Centerville itself).
- Levels 2, 3, 7 and 9: unused.
Note how the "A" digit 6 was used to resolve conflicts:
- 6 followed by 2,3,6,8, or 9 was a local 5-digit number.
- 6 followed by 5 was 658-XXXX in Cincinnati.
- 6 followed by 1,4,7, or 0 was absorbed and ignored.
So, you might ask, why didn't the telcos just segregate all Ann Arbor and Centerville numbers on separate levels, like GTE did in Carbondale?
- Because every dialing plan has to avoid conflicts between local 3-, 4-, or 5-digit numbers and NNX codes in nearby communities reached by 7-digit dialing.
- Because every dialing plan has to consider how the local dialing plans in nearby communities avoid conflicts between *their* local (3-, 4-, or 5-digit) numbers and the NNX codes used by *their* nearby communities.
- Because every NNX in an area code has to be unique. A telco can't pick an NNX just because it's convenient for the local dialing plan if it's already in use somewhere else in the area code.
And ultimately, because all dialing plans within an area code form a continuous web of inter-community 7-digit dialing, each one of which has to avoid local conflicts.
Have you followed all this? Or are your eyes glazed over by now? If you haven't followed it because it's too complicated, that's my point: it is complicated! It's amazing that traffic engineers back in the 50s and 60s were able to figure it all out.
Even more amazing is the fact that they were able to implement it with electromechanical devices: Strowger switches, relays, and time-delay relays.
This technique is discussed in detail in "Notes on Distance Dialing," Section 4, "Typical Trunking Diagrams for Step-by-Step Offices," published by AT&T Engineering and Network Services Department, Systems Planning Section, 1975. A PDF of Appendix A1 (the trunking diagram of a hypothetical SxS switch) is posted at
I had a friend named John who lived in West Quad, which had incoming number 2-4401. West Quad residents were not permitted to make inside calls either.
John figured out that if he asked the operator for 8-2440, the operator would dial it like any other 8-XXXX number. He'd then flash of the switchhook on his room phone, effectively dialing 1. Thus, the central office saw 82-4401, ignored the "A"-digit 8, and sent the call right back to the West Quad PBX. John could make an inside call.
This technique worked well on busy evenings when student operators were on duty. But John screwed up: he tried it on a weekday afternoon when the Chief Operator was on duty. She recognized his voice, and politely informed him the inside calls were not permitted. I've always wondered if she figured out how he did it.
Neal McLain
[ ... ]
[ ... ]
Digit aborbing had many uses in larger step-by-step areas, such as Oklahoma City, Tulsa, Dallas, Fort Worth, Houston, San Antonio.
Certainly it would have been impossible to trunk large multi-exchange cities like these without use of digit-absorbing.
I was on the receiving end of the woes that this could cause when I got the number VIctor 3-6056 when I moved back to Oklahoma City from Dallas many years ago.
The time number in Oklahoma City was REgent 6-0561. We got calls all during the night every night from often-incoherent callers with slurred speech who misidialed the first digit as 8 instead of 7. A call starting with 83, because of the digit absorption from the downtown office, absorped the digit 4.
The phone company soon changed my number when I complained. When I was downtown, where my office was, I tried various combinations when making calls to offices out of downtown and discovered many prefixes where the secon digit could be ignored, since it would be absored, just dialing six digits.
It varied depending on the origination and destination office, since figuring out how to make thr tuning work in step-by-step office varied according to the route. It must have been quite a challenge for the traffic engineers, but they seemed to be quite comfortable with it. Wes Leatherock snipped-for-privacy@yahoo.com snipped-for-privacy@aol.com
Hook-flash dialing was quite a trick to master, if you hadn't quite ever needed to before.
A small performing group I was part of back in the late '60s had been putting on an afternoon show for the inmates of the Yale Psychiatric Institute (YPI), and when the show was over we all retired to the ground floor bathrooms to change out of costume and into civilian clothes. By the time we were in full civvies again, 5 pm had come and gone, as had the receptionist, and the entrance doors were all locked -- as was the dial on the receptionist's pulse-dial (rotary) desk phone.
What to do? Dial 9 for an outside line, then the Yale campus police number -- all by hook-switch flashing (!). Took about half an hour to get all the timing right, but, yes, it *did* ultimately all come together, and the campus cops came to let us all out.
Cheers, -- tlvp
Vol 2 of the Bell Labs history (op cit) describes what had to be done, and it was not trivial. As mentioned, they had to add digit-absorbing selectors in step-by-step exchanges. While that's not replacing the entire exchange, it's not something that simply gets plugged in either.
Further, the new exchange designation has to be wired into other exchanges, intermediate tandem switches, and other equipment so the exchange can be reached.
If someone's phone number was say 6235, how would the switch know if someone was dialing 687-6235 or just 6235?
There were a lot of different dialing patterns and they all had to be uniform so that someone in a town could dial 7 digits and the call would still go through. If fewer digits would work--as it continued to do so in many towns--so much the better, but making that happen wasn't always that simple.
There was also the issue of dialing other towns nearby.
Many libraries offer the NYT on-line as well as microfilm. Check out
12/8/1946, pg 209, classified ad, Houses Queens, and you'll see an ad by Listing Realty of Hillside Ave with HO 5-10412. You'll also see an ad for BE 5-2874J.For such eight digit subscribers, as well as party line subscribers with a letter suffix, they all would need to get a new number when their exchange went dial. In those years many people did get entirely new numbers. There were Bell letters and booklets to customers and special directories issued to reflect those changes.
Lots of people got new exchange designations.
In the mid-1970s I worked for a business served by a small Centrex. All extensions were in the 3xxx series, and we dialed 4 digits internally. I was curious and dialed 417 and reached extension 3417. So, actually only three digits were required in most cases. As an aside, their switchboard was an old style cord board; I had assumed Centrex switchboards would be modern cordless consoles.
Presumably the switch serving us was step-by-step, which ironically was easily adapted to Centrex service. (Panel and No. 1 crossbar could not support Centrex.) The irritating aural signals sounded like step-by-step PBXs; Centrexes served by No. 5 Crossbar seemed to have modern signals like a regular exchange. Also, toll calls were ONI, that is, an operator had to come on to get the calling number, even for suburban message unit calls. On this switch when we transferred a call, we flashed the hookswitch once, and the operator came on to do the transfer. On later Centrexes (presumably served by ESS), a hookswitch flash would generate a stutter dial tone and the person could dial the desired number himself, or even easily set up a three-way call.
Side note: On the older Centrexes, when one dialed nine for an outside line, there would be a slight pause and click before the outside dial tone came on. But on the newer ones the dial tone remained on continuously, no pause, no click.
Side note: On many PBXs and Centrexes there were tie lines to other PBXs of the company. One would dial an access code (often 8 or 8n), then the extension on the remote PBX. While usually one could not get an outside line on the remote PBX, one could get a tie line, and dial back to your own PBX, or another PBX if there were multiple tie lines. Very large organizations had private networks interconnecting numerous locations. In experimenting, I once ended up reaching another company who shared our Centrex prefix--apparently the shared circuitry at the central office did not have the necessary 'block' for tie-line calls. Normally our extensions could not dial between two separate companies. (The business I reached was an entirely separate organization).
Although the Bell System prided itself on standardization, it appears PBX, Centrex, and tie line arrangements of large organizations varied considerably from one instalaltion to another, and probably there was variation in local practices, too. Many dial PBX's started their extension numbering with 2, but others did have a 1 series. Dial codes and procedures for tie line arrangements varied greatly.
In the 1950s and 1960s, the Bell System developed a variety of relatively small modern dial PBXs that had cordless consoles, a compact switch, and more automated features (eg 'camp on'). Vol 2 of the Bell Labs history gives some description of those, though mostly the internals of the switch. Unfortunately, there is no mention of prices. I would love to see a Bell System commercial price list from say 1965 listing the rentals of their various dial and manual PBXs and corded and cordless switchboards. Particularly, for a small _dial_ PBX, I wonder about the relative cost of a cordless vs. corded switchboard. since a cordless board required considerably more circuitry but didn't yield much more productivity to the customer.
Most probable answer: there were no numbers in the form 6XXX, 7XXX, or
8XXX. As I explained in my post of Sep 17, 10:14, the dialing plan segregated local 4-digit numbers from the absorbed digits 6, 7, and 8. Thus:1 = access for non-local numbers + vertical service codes
2XXX = local number, blank level, or nearby community 3XXX = local number, blank level, or nearby community 4XXX = local number, blank level, or nearby community 5XXX = local number, blank level, or nearby community 6 = absorbed at the first selector by "AR" digit absorber 7 = absorbed at the first selector by "AR" digit absorber 8 = absorbed at the first selector by "AR" digit absorber 9XXX = local number, blank level, or nearby community 0 = operator (or access code)But, as I noted previously, the above dialing plan is just a best-guess on my part. I assume Thor will let us know which levels actually were used for local numbers.
As for Thor's original statement, "Not such a major undertaking," it would been a fairly straightforward undertaking at the local exchange level. Each first selector would have to have been equipped for digit-absorbing circuitry, and strapped to absorb the specific digits relevant to the exchange (6, 7, and 8 in this case).
The *really* major undertaking would have been at the area-code level, where the job was to assign each community within the area code one or more NNX exchange codes which:
- Wouldn't conflict with the local dialing plan.
- Wouldn't conflict with the local dialing plan of any neighboring community reached by seven-digit dialing.
- Wouldn't conflict with any existing NNX codes within the area code.
- Were true NNX codes (which excludes N0X and N1X codes, which didn't exist until the '70s, and even then were restricted to larger cities already us > On Sat, 17 Sep 2011 10:14:40 -0500, Neal McLain wrote: > ... [big snip] ... >> John figured out that if he asked the operator for 8-2440, >> the operator would dial it like any other 8-XXXX number. >> He'd then flash of the switchhook on his room phone, >> effectively dialing 1. Thus, the central office saw 82-4401, >> ignored the "A"-digit 8, and sent the call right back to the >> West Quad PBX. John could make an inside call.
Well, yeah, but if you're only dialing a 1, it's pretty easy! The trick is to do it fast enough that the CO won't hang up, and that it won't attract the attention of the PBX operator.
Neal McLain
A list of prefixes was assigned for each area code, and all of those considerations were taken into account when the list was first set up. Prefixes were not just assigned randomly. In locations which already had multiple prefixes, those were incorporated into the first area code-wide assignment.
There were also some locations where a community of interest existed between offices in different area codes, where the prefix had to be "protected" in both area codes.
A prime example was the greater Kansas City Metropolitan exchange, since the metro area extended across two states with seven-digit dialing. The Missouri side had area code 816, the Kansas side 918.
Of course, this consideration ceased to be a factor when mandatory
10-digit (or 11-digit) dialing on local calls came into existence.Wes Leatherock snipped-for-privacy@yahoo.com snipped-for-privacy@aol.com
You mean 913.
Don't those two cities have independent 7-digit dialing plans now? I thought 816 and 913 each had 7-digit dialing internally but 11-digit dialing across the river
Another example: DC/Maryland/Virginia metro area. Before 1953, the entire area had a single 6-digit dialing plan (2L+4D). In '53, 7-digit dialing was introduced (2L+5D). Sometime later, the dialing plan was split by area code, with 10-digit dialing across boundaries. Since then, both Maryland and Virginia have gotten overlays, so the metro area now has five area codes with 10-digit dialing in Maryland (240+301) and Virginia (671+703). DC (202 still has 7-digit dialing.
I suppose someday 202 will need relief, with the obvious choice being an overlay. However, somebody here on T-D once suggested a split of sorts: put the federal government in its own area code -- 666.
Neal McLain
On Sat, 17 Sep 2011 20:28:10 -0400, after tlvp wrote
Moderator added:
1) Hook-flashing a Yale Operator on a YPI phone wouldn't go very well, as the Yale operator would simply respond as if we were YPI inmates having somehow gotten unauthorized after-hours access to the phone. Campus Police perhaps likewise, but they'd come to make sure we were properly confined again, and then learn we weren't inmates at all :-) . 2) A consistent rate is important, of course, but the speed is, too: slower than about 5 pulses per second and your click-stream becomes not a pulse-dial 9, for example, but rather nine pulse-dial 1s.The trick to getting a consistent *and* fast stream of clicks, I found, was (i) to alternate fingers on both hands, (ii) to hook-flash in perceptible, systematic patterns, e.g., three quick triplets for a 9, Click-click-click Click-click-click Click-click-click, with other suitable patterns for other digits, and (iii) to aim for something like 6 flashes per second (360 as a metronome setting) or more. HTH :-) .
3) Q.: Will hook-flash dialing still work on today's DTMF-based PBXes?Cheers, -- tlvp
Oops ... I mean 571, not 671.
Neal
Speaking of Washington . . .
Washington went dial around 1930. Washington grew somewhat during the New Deal as the alphabet soup agencies were formed, but then grew tremendously during WW II. (The Pentagon alone had about a 16,000 station PBX. See David Brinkley's excellent "Washington Goes to War".)
Would anyone know what kind of switch type was used for Washington dial service, especially as the city grew? Also, did Washington get any special long distance switches or trunks to meet wartime demand? Did the military have an early version of Autovon during WW II to interconnect Washington with military bases around the country? I recall reading that they did build a new cross country toll line just before the war, but I can't recall the details. The No. 4 toll crossbar switch first went into service during WW II, but it was based in Philadelphia, not Washington. Also, to meet war time traffic needs, they narrowed the bandwidth on toll lines to squeeze in more capacity.
With the press of wartime traffic, long distance facilities to/from Washington must have been strained. I wonder what percentage of toll calls were completed on a demand basis vs. a delayed basis, and if operators had a priority system for handling calls.
(There's a Bell Labs volume on military service, but it's mostly about fire control, radar, and military radio communications.)
Cabling-Design.com Forums website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.