Think about it. The only computers in operation back then were differential analysers, which used wheels and gears to solve equations. These were strictly analog devices and had no real concept of "memory," let alone interactivity. They couldn't store information, let alone allow an operator to enter the text of a news report.
So how was it done with analog technology? The January 12, 1929 issue of the New Yorker explains the process beautifully. The sign was "designed, built, installed, and inaugurated" by Frank C. Reilly, who seems better known for the plays that he wrote than for this incredible invention. The way it works is ingenious...and totally crazy to modern sensibilities:
The frame which encircles the building and on which the moving letters and words appear is controlled by a smaller rectangular affair in a room on the fourth floor. Through this the letters, made of metal, are run after being dropped in proper order down a chute. Each letter, as it moves along, passes over myriad metal brushes which are connected with the bulbs on the outside of the building.Nowadays we'd be encoding -- digitizing -- the letters into a code that could be deciphered by the sign. But back then, the only way anybody could conceive of such a sign was to LITERALLY make it an ANALOG device...one which responded to REAL letters which whizzed around and triggered the lightbulbs. Amazing!
I think this is a situation where the technology looks more surprising to a modern day observer than it would have to somebody in 1929. We've moved so far away from the idea that a REPRESENTATION of something can nearly match its ORIGINAL, that the thought of somebody designing such a sign is totally bizarre. And yet, when you think about it, the design makes perfect sense, and it's probably the only way it could have been done.
The sign is still going, though I'm not sure if it uses the same technology. This New York Times article gives a few more details about how the sign worked (or still works).
PS: After reading the scant details another time I realize that my original conception -- that the letters were full-size models that moved directly behind the lights -- is probably false. It would seem that the letters were made to scale and run through a miniature track, whose sensors extended far out to the actual bulbs on the walls. This explains why the sign wasn't incredibly noisy on the outside, and also how it could gracefully go around corners.