CircleID

While doing research for a paper on telegraph codebooks, I was reminded of something I had long known: one could have short addresses for telegrams. A short article in The New Yorker described how it worked in New York City.

Briefly, one could pick more or less any name that wasn't in use, and list it with the Central Bureau for Registered Addresses. Many names were cute, such as Nostrop for the Gillette razor company or Illuminate for the New York Edison Company. Prices seem cheap—$2.50 per year—though this was during the Depression.

Errors were an issue: you couldn't register a name that could easily be confused with another given the error properties of Morse code. For example, A is "._", but if there's too long a space between the two signals it might be received as ET, which is .  _.

The Central Bureau also did error correction on failed address look-ups. To do this, they used things like reversed-name indices, an alphabetical list going right-to-left. The article suggests that most errors were in the first few letters of an address; I suspect it was more that the human doing the look-up at the receiving telegraph company could do any necessary error correction given the first few letters and an alphabetical list.

There turns out to have been a security angle as well, though I confess I don't understand it. This particular service was local to New York City, so perhaps we're really talking about the NYC.NY.US domain, rather than an analog to .COM. Originally, each cable company had its own address list. However, "In 1917, the State Department, fearing spies, abolished all existing lists and set up a uniform one for everybody." I don't know why having a single list would make life harder for spies. After the war, the article says that the cable companies realized that this was more convenient for everyone and set up their own single list. Personally, I suspect it was more a matter of customer demand.

I wonder, though, if some of the same issues we see today arose then. Were there trademark disputes? Telesquatting? It might be fun to research this history.

Written by Steven Bellovin, Professor of Computer Science at Columbia University. Visit the blog maintained by Steven Bellovin here.

Related topics: Cybersquatting, Domain Names, Domain Registries, Internet Governance, Law, Top-Level Domains