The Lie of the Immortal Bit: On the Peril of Perfect Digital Copies
We are told, with the unwavering certainty of a catechism, that the paramount goal of digital preservation is the creation of a perfect, unbroken, and immutable copy. The bit-for-bit duplicate. The checksum-verified, pristine clone of a file, a website, a database. This is the holy grail, the unassailable best practice. But what if this foundational principle, this dogma of digital immortality, is not just incomplete, but actively dangerous to the very thing we seek to preserve: meaning?
The obsession with the perfect copy is a reductionist fantasy. It treats information as a mere object, a static vessel whose only value lies in its flawless transmission through time. It assumes that a string of ones and zeros, perfectly preserved, will always decode into the same human experience. This is the lie of the immortal bit. It ignores the profound truth that context is not data; it is the atmosphere in which data breathes. A perfect copy of a website, stripped of the network latency, the specific browser renderings, the plugins that once made a video play, is a taxidermied bird. It has the exact same feathers, but it will never sing.
The Ghost in the Machine is the Machine Itself
Consider a piece of software art from the 1990s, designed to run on a specific graphics card with known driver bugs that the artist exploited for a visual effect. A perfect bit-for-bit copy of the artwork’s file is made. Decades later, we verify the checksum. It is perfect. We run it on a modern emulator that perfectly mimics the period-correct CPU. But the emulator, in its quest for clean, accurate emulation, “fixes” the graphical glitch the artist relied on. The artwork is rendered incorrectly. The perfect copy has failed. The artifact was not the file alone; it was the file in conversation with its flawed, material environment.
This is the peril. By focusing solely on the bits, we risk creating cemeteries of perfect corpses, meticulously maintained but utterly lifeless. We preserve the ‘what’ with fanatical precision but lose the ‘how’ and the ‘why’. We save the text of a digital novel but lose the proprietary e-reader software that presented it with a specific, intended pagination and typography that the author approved. The experience—the actual thing we read—is gone.
The counterintuitive argument, then, is this: sometimes, the most faithful preservation requires intentional, knowledgeable corruption. It demands we preserve not just the data, but the glitch. The slowdown. The obsolete plugin. The broken link. It requires emulating the entire technological ecosystem, warts and all, because those warts are often features, not bugs. It argues for saving clumsy, inefficient, and ‘impure’ system images over pristine, isolated files.
True preservation is not the science of stopping time, but the art of capturing a moment in it, with all its beautiful, functional messiness intact. It is an argument for saving the noise to hear the signal. Our goal should not be a museum of sterile, identical specimens, but a living record of digital life, in all its contingent, fragile, and gloriously imperfect glory.
Notes & further reading
A few pages I came back to while writing this: