## Enough to specify you

Go to Google and start typing the name of a website. How far do you have to type before that site’s the number one hit ghosting below, that is, Google’s guess of what you are trying to type?

Here’re a few random samples. (Your results may vary.)

• aardvarch/aeology (also, martin rund/kvist)
• phary/ngula (also, pz m/yers)
• daylight at/heism
• whate/ver (john sca/lzi or scalz/i)
• respectful i/nsolence
• dispatches f/rom the culture wars (ed bray/ton)
• xk/cd (randall m/unroe)
• charles str/oss

Meanwhile, my dear Lemmata had a name too generic to get to the first spot; and Mirrors of Eris got there only with the final “s”. (Dang you, “Places for mirrors of near Millcreek Township Police Airport (ERI), 4411 W 12th St, Erie, PA 16505″!)

I could probably make a point about how just a few words are enough to specify something even from a thing so vast as Google; but the necessary length is a very random thing. Single-word names are hopeless, unless the word is rare or the subject famous; and others appear (or not) depending on the commonness of their word combinations. For example, “deep thoughts” does not distinguish itself because it is a) a common phrase, and b) a Jack Handey humor title; but “deep thoughts and si/lliness” distinguishes.

One would be tempted to make up a word and thus have a unique, easy-to-find identifier; but this can run afoul of Google’s helpfulness. For example, on a whim I suggested on Mirrors of Eris that I could have named the place “snurtlet” instead for this effect — now that I actually checked if it would have worked, Google helpfully told me this:

Showing results for snurtle. Search instead for snurtlet (here).

The first link for snurtle is to the Urban Dictionary; predictably, the fifth definition is about penises. (There is no sixth, but if there was, it would probably be “an unsuccessful attempt to perform an asphyxiatic Boston pancake i.e. a Poughkeepsie Clayface with Lacquer.”)

(Why yes, I am a horrible person; I know.)

Actual calculation shows that with 26 letters and 10 digits, and with $6\times 10^9$ people, seven characters should be enough to specify each living human being thirteen times; each once is around 6.3 characters. (Which sounds crazy wild, until you remember that with numbers that multitude takes 10 characters. Er, yeah.)

Just in case, if humanity ever gets around to assigning a unique identifier to each living person, I’m reserving MOFERIS, so I won’t get stuck with HRERFVV. Or DAMNGIT. (Or possibly I should take FUCKOFF; wouldn’t that be something to be telling telemarketers and social workers and everyone who requested your identity? “FUCKOFF, officer. Is there a problem?”)

An exercise for the reader: Estimate the number of insults and slurs of seven letters or less in all the major languages of the world. Estimate which fraction of the seven-character identifiers would have be disqualified if the inclusion of such a word was grounds for it. Assuming there are 13 times more identifiers than people, does this leave enough “clean” identifiers to not stick anyone with SADFUCK?

### One Response to “Enough to specify you”

1. Bob O'H Says:

Pity the poor sod who ends up with PERKELE. Especially if they’re a fundy Christian.

Actually, I want the UN to set this scheme up, and I want to infiltrate it.