Wednesday, January 26, 2011

(empo-tymshft) Excellent predictions in 1968, but they didn't take Moore's Law into account

[DISCLOSURE: My employer traces its history, in part, to the Autonetics Division of North American Rockwell.]

If you search Google Books for the August 1968 issue of Popular Science, you will find an article by W. Stevenson Bacon that discusses twelve then-new advances in public safety and law enforcement. The items that were discussed in that 1968 article are now commonplace in 2011, in ways that Stevenson couldn't have imagined. For example, item 6 discusses the capability to track crimes by location, but Stevenson didn't anticipate that an entire supersystem would spring up that would allow very precise tracking of any location on earth. And Stevenson didn't anticipate that we would carry telephones that could broadcast their location - even the Motorola rep quoted in item 9 could only envision "tiny personal teletypewriter units that will permit individual police officers to receive messages direct from a computer." (What if the police officer could CARRY a computer in the palm of his or her hand?)

But I paid most attention to item 4, which talked about the beginnings of the industry that has employed me for the last 16-plus years. As you read this, remember that it was written in the summer of 1968.

While it was possible to first identify Senator Kennedy's suspected assassin, Sirhan Bishara Sirhan, through the pistol he used (as a case in point, through a computer record of guns sold in Los Angeles), new fingerprint-matching systems might one day give instant identification of all suspects.

In addition, experts hope to teach computers to recognize partial prints such as were available in the assassination of Dr. Martin Luther King. The few smudged prints left by suspect James Earl Ray made his identification a matter of weeks rather than minutes.

Working under FBI contract, researchers at Cornell Aeronautical Labs and Autonetics Division of North American Rockwell have literally thrown out traditional matching techniques. They're putting into mathematical language such print details as ridge endings, forks in ridges, incipient ridges, islands, and enclosures.

However, Stevenson's analysis was lacking on one respect:

It is hoped computers can be made to chart these details horizontally and vertically, and calculate a set of "descriptors" that would be stored in its memory. The machine would then be able to scan and match fingerprints - including partial prints - at the rate of 500,000 a day.

How would Stevenson had reacted if one of the scientists started talking about matching 500,000 fingerprints in SECONDS, using computers with GIGABYTES of memory?

Stevenson probably would have told the scientist to submit THAT idea to Popular Science Fiction.

In the title of this post, I referred to Moore's Law, which anticipated these rapid increases in computing capabilities. Perhaps Stevenson didn't happen to see the article written by Gordon E. Moore in the April 19, 1965 edition of Electronics Magazine. Wikipedia quotes Moore:

The complexity for minimum component costs has increased at a rate of roughly a factor of two per year... Certainly over the short term this rate can be expected to continue, if not to increase. Over the longer term, the rate of increase is a bit more uncertain, although there is no reason to believe it will not remain nearly constant for at least 10 years. That means by 1975, the number of components per integrated circuit for minimum cost will be 65,000. I believe that such a large circuit can be built on a single wafer.

But neither Stevenson nor Moore could predict the variety of devices that would be created. Not only do we have "smartphones" and large biometric systems today, but these advances in computing power have also provided us with Tamagotchi electronic pets, Sirius XM radio, and devices that can hold thousands upon thousands of long-playing records...wait a minute, what's a "long-playing record"?
blog comments powered by Disqus