Storage, transmission, whatever
« previous post | next post »
Mike Ives, "Charles Kao, Nobel Laureate Who Revolutionized Fiber Optics, Dies at 84", NYT 9/24/2018 [emphasis added]:
Working in Britain in the late 1960s, Dr. Kao and a colleague played a crucial role in discovering that the fiber optic cables in use at the time were limited by impurities in their glass. They also outlined the cables’ potential capacity for storing information — one that was far superior to that of copper wires or radio waves.
As commenters on the NYT obit point out, optical fiber (like copper wires and radio waves) is used for transmitting information, not storing it.
There's nothing in Kao's Nobel citation or his Wikipedia entry to suggest a source for this error. Nor could I find any similar errors in his other obituaries. So this (potentially serious) substitution therefore remains a problem in attributional abduction.
The author's other pieces seems to be about East Asian politics and culture, rather than science, so it's possible that he's just unclear about the difference between transmission and storage. This strikes me as an unlikely type of confusion, but it's possible that a larger fraction of educated people than I would have thought see data transmission as an aspect of data storage, or perhaps see data storage and data transmission as interchangeable concepts. One little-studied aspect of the "science of science communication" is the set of conceptual spaces in terms of which non-experts interpret key aspects of science and technology.
It's also possible that this error was a slip of the fingers — substitution of a semantically related alternative that the author would recognize as wrong — on the part of the writer or an editor.
In case someone at the Times discovers and corrects the mistake, here's a screenshot:
mikegrubb said,
September 28, 2018 @ 8:35 am
Maybe Mr. Ives is a victim of the "series of tubes" definition of the internet espoused by a certain US Senator and imagines the data being stored by the fiber-optic cable between the moment the impulse is sent and received the way a water pipe could be said to temporarily "store" water until the tap is opened. (I admit I'm reaching here.)
Peter Vanderwaart said,
September 28, 2018 @ 8:38 am
The text suggests to me that transmitting and storing are two separate possible uses of fiber optic cable. For example, some very early computers used sound waves in water as a data storage mechanism taking advantage of the relatively slow speed of sound in water.
[(myl) Early computers used delay-line memory systems, typically based on recirculation of acoustic waves in mercury-filled tubes, and there are recent suggestions about exotic methods for light-wave data storage. But Kao's work on optical fibers has nothing to do with any of that.]
Jerry Friedman said,
September 28, 2018 @ 9:19 am
And along the lines of what mikegrubb said, what it makes me wonder is whether Ives thinks of data being transmitted as being loaded onto trucks or boxcars, which is like being stored, sort of.
By the way, I found "outlined the cables' capacity" jarring. Estimated, calculated? To me you outline something with a lot of parts, such as a plan, not a single capacity or even a capacity as a function of fiber diameter or whatever.
David L said,
September 28, 2018 @ 9:25 am
The final section of the key paper (there's a link in the NYT obit) includes this statement:
Thus, compared with existing coaxial-cable and radio systems, this form of waveguide
has a larger information capacity and possible advantages in basic material cost.
It's clear from the paper itself (section 4.4) that by 'information capacity' they mean transmission capacity but I can see how the sentence above might be misinterpreted.
cM said,
September 28, 2018 @ 10:21 am
Well – while usually not being used as dedicated storage devices, long fiber optic cables have quite a bit of data "in flight", and could of course still be used as delay-line storage today.
A quick search led me to a recently built (hm, are submarine cables "built"? Or rather "deployed"?) Bilbao-to-Virginia cable 4000 miles long and capable of 160TBit/s – meaning it will "store" almost two thirds of a terabyte:
4000 miles * 160 Tbit/s / (2/3 * c) (refractive index ~1.5)
Philip Taylor said,
September 28, 2018 @ 10:37 am
(Submarine cables) — They are certainly "built" (or more commonly "manufactured"), after which they can be "deployed" (or more commonly "laid").
Technical Officer (retired), Post Office Cable & Wireless Services.
Jeremy said,
September 28, 2018 @ 12:23 pm
Clearly not the case here, but there was a time when transmission over a wire was use for storage:
https://en.wikipedia.org/wiki/Delay_line_memory
jpiitula said,
September 28, 2018 @ 1:01 pm
Our computing center used storage medium as transmission medium for a fifty terabyte dataset (OCR data together with the scanned pages in lo and hi res jpeg) only a few years back. They deemed it most practical to move the physical disks to the computers, mount them locally to copy the data to the computing environment, then return the disks.
cM said,
September 28, 2018 @ 1:12 pm
jpiitula: Well, yeah. For really, really big data sets, moving physical media is still the way to go.
Amazon's Snowmobile is a – quite impressive – real-world implementation of the Tanenbaum station wagon.
Jim said,
September 28, 2018 @ 1:47 pm
Could this be a reference to lower signal loss? That's still an awkward way of phrasing it, but of course fiber optic only attenuates over much larger distances compared to other media.
H said,
September 28, 2018 @ 2:44 pm
As an engineer working with transmission and storage of electrical energy, I sometimes say that storage is simply transmission through time instead of space!
As some commenters have pointed out, storage technology might be used for (or mistaken for) transmission purposes, but it's much harder to use (or mistake) transmission technology for storage purposes.
MattF said,
September 28, 2018 @ 4:47 pm
@H
I agree– data transmission is just data storage in spacetime. It's fair to note that spacetime has certain causality constraints… but that's a lecture I haven't written yet.
Lillie Dremeaux said,
September 28, 2018 @ 5:41 pm
Glad readers pointed this out. The obituary was fixed and has a correction on it now. (By the way, the writer's surname is spelled Ives, not Yves.)
~ LL fan/NYT editor