The Babbage machine returns, smaller but more important

Ironic, but just a few weeks before Ada Lovelace Day, celebrating the woman who programmed Charles Babbage's unbuilt Difference Engine, we found that Babbage's design principles may have real 21st century importance.

(This is a close-up of an illustration that appeared in an abstract at Nano Letters, showing a logic gate based on Babbage's design principles.)

The lab of Pritoraj "Raj" Mohanty at Boston University has used the principle structure of the Babbage Difference Engine -- a mechanical logic gate -- and built a nanoscale version of it that uses less energy than traditional chip gates.

This is not just important in a green way. Energy loss is a big problem for chip-makers.

Energy is lost in the form of heat and heat can build up. That's why computers have fans, to dissipate that heat. Otherwise, as in the ill-fated Intel Pentium IV, the chip can over-heat and burn out. These heat problems eventually caused chip makers to move toward low-power chip designs in the last decade.

In the original Difference Engine design, gears set numbers mechanically when a handle was turned. A later version was actually built, and helped solve logarithmic problems. But the revolution in calculation Babbage predicted never happened, due to the costs of manufacturing the machines.

The new nano-Babbage machine uses electricity, of course, but the principle of storing and moving whole numbers remains.

As described in the journal Nano Letters, Mohanty's gate uses electricity to set a silicon strip vibrating, and that vibration can be reliably switched to different amplitudes, settings software can program.

The gate is not as fast as a traditional gate, but since it uses less energy, it creates less electrical interference with other components, meaning it can be made smaller and more reliable.

Mohanty's success also illustrates an important principle in science. Don't confuse it with engineering.

What can be conceived can't always be built right away, but the principle remains important. Discarding science that can't be engineered right away is a mistake, and the past is always teaching us lessons.

This post was originally published on Smartplanet.com