Brooklyn artist Daniel Bejar made a copy of the key to his apartment (hopefully not his current one, because of stuff like this). Then he made a copy of the copy. And then he repeated the process, by my count, 67 times. At which point the copy was an uncut blank. Kind of like I Am Sitting In a Room, but with keys. He calls it Visual Topography of a Generation Gap. [via Boing Boing]
More:
24 thoughts on “Generational Loss In Housekey Copies”
Comments are closed.
That’s some pretty atrocious loss going on. His key-copying would appear to have a problem.
I have no doubt that generational loss is real, given that my own housekeys are copies-of-copies … and three separate sets of ‘backup keys’ can’t open my doors. But the ‘failure mode’ isn’t “hey, everything rightshifts one mm!”, it’s a more irritating ‘grooves widen, peaks flatten’ sort of mushiness.
I agree it seems far too regular an error to be truly random.
I think a more interesting experiment would be to make a family-tree style graph of keys where you make two copies of each key, then once a key stops working it can no longer be duplicated and that family line ends.
Perhaps it is a different setup? Someone should try to reproduce.
So, more of a lowpass filter keycutter, then?
I’m guessing at some point it didn’t open the door anymore.
indeed seems more like a systematic error than an accumulation of random errors.
I don’t know that we should trust this man. He is a known pornographer.
Just kidding, this is great. Interesting that the loss seems to be solely dependent on the position of the notches, and not their size.
That is due to a gross misalignment in the key cutter. a well adjusted key cutter can go a lot farther in generations.
A key cutter that badly aligned will have problems making even a single key that will work properly.
Looks like the lateral offset of the copier is off slightly, causing a forward marching of the tooth pattern.
Wouldn’t be hard to do, and probably faster to get an effect than an actual generational loss, i.e. random edge errors in the cuts.
Key copiers are pretty accurate, if they’re setup and run correctly. So you have to intentionally botch it somehow to get something visible without doing thousands of keys (which gets a bit expensive, I’d wager…)
Yes it is offset. The “old” key is farther in the machine than the “new” key causing this pattern in the cut to be transfered to each “new” key when it is cut.
Phase error .NE. 0
I don’t know that we should trust this man.