In September 1888, an American machine-tool engineer named Oberlin Smith published an article in The Electrical World titled "Some Possible Forms of Phonograph." It was not a patent — and that is precisely what makes the story remarkable. Smith had filed a U.S. caveat (No. 3,683) in October 1878, ten years earlier, describing a method of recording sound magnetically. A caveat was a now-extinct legal instrument: a confidential notice to the Patent Office that an inventor was working on something and wanted a one-year warning if anyone filed a competing claim. Smith let his caveat lapse without ever pursuing a full patent, and then in 1888 simply published the whole idea, gifting it to the public domain.

The idea was this: take a length of cotton or silk thread, embed it with steel dust or tiny iron filings, and pull it past an electromagnet whose coil carried the fluctuating current from a microphone. The varying magnetic field would magnetize the iron particles in a pattern corresponding to the sound waves. Pulling the thread back past a similar coil would induce a matching current — and reproduce the sound. Smith had, in a single eight-page article, described:

• The principle of longitudinal magnetic recording, used in every cassette tape, reel-to-reel, and VHS deck of the 20th century.
• The concept of particulate magnetic media — discrete ferromagnetic particles in a binder — which is exactly how audio tape, floppy disks, and early hard drives stored bits.
• An acknowledgment that a solid wire or ribbon would probably work better than embedded thread — anticipating both Valdemar Poulsen's 1898 wire recorder and Fritz Pfleumer's 1928 magnetic tape patent.

Smith's reasoning was unusually rigorous. He noted that Edison's tin-foil phonograph (patented just months before his caveat) suffered from mechanical wear — the stylus physically deformed the medium each time it played. A magnetic record, Smith pointed out, would be non-contact on playback: the field would induce current without the medium ever touching the pickup. This is the exact argument that, a century later, would justify read heads flying micrometers above a spinning platter in the IBM hard disk.

Why did it take so long? Smith had no amplifier. The signal induced by his magnetized thread was so faint that, without vacuum-tube amplification (Lee de Forest, 1906), playback was barely audible. Poulsen demonstrated a working wire recorder at the 1900 Paris Exposition — Emperor Franz Josef recorded his voice on it — but commercial magnetic recording had to wait for AEG's Magnetophon in 1935, which used Pfleumer's iron-oxide-coated paper tape. That tape, in turn, became the direct ancestor of the iron-oxide coating on the first IBM RAMAC hard drive in 1956, and of every spinning-disk drive shipped since.

Smith's 1888 article is now considered the founding document of magnetic recording. He could have owned the field. Instead, he gave it away — explicitly writing that he hoped "someone with more time" would develop the idea. Someone did. Everyone, eventually, did.