{"id":411,"date":"2010-06-09T16:49:37","date_gmt":"2010-06-09T20:49:37","guid":{"rendered":"http:\/\/blogs.law.harvard.edu\/niftyc\/?p=411"},"modified":"2011-07-29T22:55:46","modified_gmt":"2011-07-30T02:55:46","slug":"the-slowest-dial-is-mexico-city","status":"publish","type":"post","link":"https:\/\/archive.blogs.harvard.edu\/niftyc\/archives\/411","title":{"rendered":"The Slowest Dial is Mexico City"},"content":{"rendered":"

(or: When Systems Are Engineered, <\/strong>Who Gets the Best Addresses?<\/strong>)<\/p>\n

In a project I’ve been working on about addressing systems in communication infrastructure (excited yet?), I’ve been telling people that early phone numbers<\/strong> were organized in part around the time it took to dial them<\/strong> on rotary telephones.<\/p>\n


\n<\/em><\/p>\n

\"\"<\/a><\/p>\n

[a rotary dial — click to enlarge — photo by <\/em>zen on flickr<\/a>]<\/em><\/p>\n

You see, youngsters, the weighted dial on a rotary telephone requires a fixed amount of time <\/strong>to dial each number. \u00a0It’s about one second per ten values so that the amount of time goes up as the number goes up, with one being the fastest number to dial (about a tenth of a second or 1 click) and zero being the slowest (about one second or 10 clicks). \u00a0These time estimates don’t include moving your fingers around, mind you.<\/p>\n

You could say, who cares? \u00a0But if you add up all of the seconds required to dial and multiply by all of the phone calls, that’s a lot of seconds<\/strong> people spend dialing those nines and zeroes.<\/p>\n

<\/p>\n

The Optimization Persists Long After It is One<\/h2>\n

I told people that when area codes were first invented the lowest numbers went to the largest population centers<\/strong> first to optimize dialing time for the most people. \u00a0E.g., New York City<\/strong> got 212. \u00a0Chicago <\/strong>got 312. \u00a0A neat and sort of interesting historical fact, right? \u00a0It helps to explain how a system optimization at one point can have long term consequences after the reasons for it stop being relevant. \u00a0(No more rotary phones, and surely the largest cities in the US have changed as well in 50 years, but Detroit <\/strong>still has a pretty awesome area code that’s easy to dial on rotary phones [313].)<\/p>\n

I dimly remember seeing this in print somewhere.<\/p>\n

Here’s the problem: \u00a0I’m not sure where I learned about this rotary dial influence and I had no sources<\/strong>.<\/p>\n

So I repeated this story recently in public and I got: “What about Washington, D.C.<\/strong>? \u00a0Surely that was a major population center when area codes were first introduced!” \u00a0Um, I’m sure it was, I said. \u00a0But then why, the questioner asked me, would it receive the area code “202” — one of the slower ones to dial because of the zero? \u00a0(1.4 seconds) \u00a0Um… \u00a0um…<\/p>\n

Okay, so I figured this out — and I have a source. \u00a0The North American Numbering Plan of 1947 says that in the first area codes the middle digit was restricted<\/strong> to either a zero or a one.<\/p>\n

Initially, zero in the middle position meant “whole state area code” while one meant “this state has more than one area code. \u00a0The middle digit had to be restricted in this way (to a 0 or 1) because, somewhat bizarrely, this is how the electromechanical switching equipment would differentiate between an area code and an exchange number. \u00a0In other words, at the time you didn’t need to dial “1” first in order to call long distance<\/strong>, so if you dialed the first three digits of your call:<\/p>\n

NNN<\/p><\/blockquote>\n

…the switch you were using needed a way to know<\/strong> if what you just dialed should be treated as the first three out of seven digits or the first three out of ten…<\/p>\n

NNN-xxxx (it’s the local<\/strong> exchange number)<\/p>\n

NNN-xxx-xxxx (it’s the long distance<\/strong> area code)<\/p><\/blockquote>\n

If the former it would route to a local exchange, if the latter it would route to a long distance trunk. \u00a0It looked for a zero or a one in the middle position to identify a long distance call<\/strong>.<\/p>\n

(Yes that means, for you mad numerologists<\/strong> out there, that exchange numbers never had 0 or 1 in the middle digit.)<\/p>\n

Another fun fact: This whole system of area codes was used internally to the phone system by operators, not externally by users, for about fourteen years<\/strong>. \u00a0Direct dialing by users didn’t become universal until the 1960s.<\/p>\n

Exceptions Break Your Beautiful Optimization From the Start<\/h2>\n

So the solution to the puzzle is that only the first and third digits were optimized for population density. \u00a0Washington, D.C. then gets a very good number because the “2”s in “202” are fast to dial, but it gets a zero in the middle because it is a state — or in this case a “district” — that only has one area code. \u00a0The engineers probably thought that the 0\/1 distinction would be fine because states with population centers in them would tend to have “1”s because they would have more area codes, and 1s are faster to dial. \u00a0But here is yet another example of an exception breaking a beautiful optimization plan<\/strong>. \u00a0DC is a kind of city\/state. \u00a0Not really a state, but more than a city. \u00a0Since it isn’t subdivided further, it gets a slow “0”.<\/p>\n

(An aside: This would never happen today because the telcos now ensure that Washington DC gets the best of everything<\/strong> — they know to keep the legislature and the regulators impressed with their services.)<\/p>\n

Okay all this may seem arcane, but these kind of things matter to people<\/strong>. \u00a0Area codes and phone numbers matter, at least they used to.<\/p>\n

\n

\"\"<\/a><\/p>\n

[This ad explains the shift from numbers to letters in Moscow.
\nClick to enlarge — from MTGS via <\/em>Artemy Lebedev<\/a>]<\/em><\/p>\n

This long path of puzzle solving made me think about the premise beyond the seeming anomaly of Washington DC. \u00a0Is it true that the original codes were optimized by dialing time and population density?<\/strong> Who won and who got screwed in the original numbering scheme? \u00a0That is (besides Washington DC) who got the slowest or disproportionately slow-dialing area codes and who got the fastest? Any anomalies? \u00a0So I made a spreadsheet and compared the dialing times to the <\/strong>1940 census<\/strong><\/a>.<\/p>\n

The Telephone Engineers Played Favorites<\/h2>\n

It looks like the relationship is sort of true<\/strong>, in that the assignments do mostly correlate to population density during the 1940s, but that in addition there are some anomalies. \u00a0In other words,\u00a0the telephone engineers played some favorites<\/strong>.<\/p>\n

The Winners<\/h3>\n

The fastest-dialing<\/em><\/strong> area codes given out in 1947:<\/p>\n

1. New York (212)
\n2. Chicago (312)
\n2. Los Angeles (213)
\n3. Detroit (313)
\n3. Pittsburgh (412)
\n3. Dallas (214) *<\/span>
\n4. Philadelphia (215)
\n4. Austin (512) *
\n<\/span> 4. St. Louis (314)
\n4. Western Massachusetts (413) *<\/span><\/p>\n

[Cities with the same number are equivalently slow\/fast.
\n*= anomalies I’ll comment on]<\/p>\n

The Losers<\/h3>\n

The slowest-dialing<\/em><\/strong> area codes given out in 1947:<\/p>\n

1. North Carolina (704) *
\n<\/span> 1. South Carolina (803)
\n1. South Dakota (605)
\n1. Nova Scotia & Prince Edward Island (902)
\n2. Idaho (208)
\n2. Montana (406)
\n2. New Mexico (505)
\n2. Tennessee (901)
\n2. Vermont (802)
\n2. Virginia (703)
\n2. Wyoming (307)<\/p>\n

[Cities with the same number are equivalently slow\/fast.
\n*= anomalies I’ll comment on]<\/p>\n

Who won:<\/h3>\n

Austin, Western Massachusetts, and Dallas. \u00a0They get super-fast numbers. \u00a0Why did they get these prime numbers? \u00a0Austin, Texas? \u00a0Come ON. \u00a0I smell a payoff. <\/strong> Favoritism. Something.<\/p>\n

Who got screwed:<\/h3>\n

Cleveland, Baltimore, and Boston. \u00a0They were given disproportionately slow numbers for their population. \u00a0Although they don’t make the above list, the many citizens of New Jersey (201) also get screwed because they get a state-level code and the middle zero gives them a very slow number relative to the state’s population density.<\/p>\n

And how about the people of Charlotte, North Carolina? <\/strong> They live in one of the 100 most populous cities in the USA at the time, yet they are lumped in with South Dakota for the slowest area code given out in the USA (2.2 seconds).<\/p>\n

Who got really <\/em>screwed:<\/h3>\n

Let’s not forget our friends in Canada. \u00a0The Canadians didn’t get any good numbers<\/strong> for their cities, even for Toronto and Montreal.<\/p>\n

Who got really, really<\/em> screwed:<\/h3>\n

Mexicans. (Sound familiar?)<\/p>\n

The Poor Have the Worst Address<\/h2>\n

Mexico joined the numbering plan later and for their trouble they were given 905 for Mexico City, the slowest area code then in service (2.4 seconds), slower than everything given in the US and Canada at that time, despite a very large population density. \u00a0This just proves the adage: The poor have the worst addresses. <\/strong>Maybe this rough treatment is what caused them to leave the North American numbering system in disgust in 1991, even though Mexico has not left North America<\/strong> as far as I know.<\/p>\n

I thought that one explanation for the anomalies might be backward compatability. \u00a0Before area codes existed, telephone operators had to look up numbers in a “route book” and cooperate with other operators in real-time to connect calls. \u00a0If you want to hear what this was like<\/strong> it was dramatized in the radio show “Dragnet” in 1949. \u00a0Here is an mp3 of the re-enactment of a long distance call before area codes<\/strong> and direct dialing:<\/p>\n

\"\"<\/a>An operator connecting a long distance telephone call<\/strong>, circa 1949
\n(from the radio drama\u00a0Dragnet<\/em>)<\/p>\n

Listen here:
\n(Windows Media Player format [.wma], 140 seconds, 327k)
\nhttp:\/\/www.porticus.org\/bell\/audio\/1949_LD_telephone%20-call.wma<\/a><\/p>\n

(Audio file from The Bell System Memorial<\/a>.)<\/p>\n

However, from everything I can determine the route numbers had no relation to these area codes. \u00a0Oh well.<\/p>\n

Another possible reason for the anomalies is that it looks like they tried to choose area codes that were very different from each other for places that geographically abut each other. \u00a0I’m not really sure why this would be advantageous. \u00a0(Ideas, commenters?) \u00a0But unlike other numbering schemes like the Interstate Highway System, similar numbers are emphatically NOT spatially adjacent to each other<\/strong> in the area code plan. \u00a0Mostly.<\/p>\n

So California gets 916, 415, and 213. \u00a0Illinois gets 217, 618, and 815. \u00a0In fact, in the entire table no two adjacent numbers refer to adjacent places as far as I can tell. \u00a0So the Dallas number on the chart above is sandwiched between a California and a Pennsylvania number. \u00a0The Austin number is between New Mexico and Ohio. \u00a0Western Mass. is between Wisconsin and Pennsylvania. \u00a0So maybe that’s it?<\/strong><\/p>\n","protected":false},"excerpt":{"rendered":"

(or: When Systems Are Engineered, Who Gets the Best Addresses?) In a project I’ve been working on about addressing systems in communication infrastructure (excited yet?), I’ve been telling people that early phone numbers were organized in part around the time it took to dial them on rotary telephones. [a rotary dial — click to enlarge […]<\/p>\n","protected":false},"author":2132,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_jetpack_newsletter_access":"","_jetpack_dont_email_post_to_subs":false,"_jetpack_newsletter_tier_id":0,"_jetpack_memberships_contains_paywalled_content":false,"_jetpack_feature_clip_id":0,"_jetpack_memberships_contains_paid_content":false,"footnotes":"","jetpack_publicize_message":"","jetpack_publicize_feature_enabled":true,"jetpack_social_post_already_shared":false,"jetpack_social_options":{"image_generator_settings":{"template":"highway","default_image_id":0,"font":"","enabled":false},"version":2},"jetpack_post_was_ever_published":false},"categories":[1321],"tags":[],"class_list":["post-411","post","type-post","status-publish","format-standard","hentry","category-research"],"jetpack_publicize_connections":[],"jetpack_featured_media_url":"","jetpack_sharing_enabled":true,"jetpack_shortlink":"https:\/\/wp.me\/p4M7Bm-6D","_links":{"self":[{"href":"https:\/\/archive.blogs.harvard.edu\/niftyc\/wp-json\/wp\/v2\/posts\/411","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/archive.blogs.harvard.edu\/niftyc\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/archive.blogs.harvard.edu\/niftyc\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/archive.blogs.harvard.edu\/niftyc\/wp-json\/wp\/v2\/users\/2132"}],"replies":[{"embeddable":true,"href":"https:\/\/archive.blogs.harvard.edu\/niftyc\/wp-json\/wp\/v2\/comments?post=411"}],"version-history":[{"count":23,"href":"https:\/\/archive.blogs.harvard.edu\/niftyc\/wp-json\/wp\/v2\/posts\/411\/revisions"}],"predecessor-version":[{"id":428,"href":"https:\/\/archive.blogs.harvard.edu\/niftyc\/wp-json\/wp\/v2\/posts\/411\/revisions\/428"}],"wp:attachment":[{"href":"https:\/\/archive.blogs.harvard.edu\/niftyc\/wp-json\/wp\/v2\/media?parent=411"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/archive.blogs.harvard.edu\/niftyc\/wp-json\/wp\/v2\/categories?post=411"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/archive.blogs.harvard.edu\/niftyc\/wp-json\/wp\/v2\/tags?post=411"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}