Articles

How could the author fail to mention flash memory technology? Are we in the 80s? How can the magazine fail proofing such reports?

...mmm... What about (electro-chemical) memistor, a Bernard Widrow creation? Memistor was the weighting element in early neaural networks implementations. A reference is given in Hecht-Nielsen book "Neurocomputing". Of course nanotech could replace a Captain Nemo-like device with something more practical.

"The reason that the memristor is radically different from the other fundamental circuit elements is that, unlike them, it carries a memory of its past"

This isn't right. Both the inductor and the capacitor have memory. The resistor is the only fundamental circuit element without memory.

THE UNITS OF THE MEMRISTOR IS STILL VOLT/AMPERE, I.E. OHMS, JUST LIKE THE ORDINARY RESISTOR R. THEREFORE, I DO NOT BELIEVE THAT THIS FOURTH ELEMENT DESERVES TO BE CONSIDERED A FUNDAMENTAL ELEMENT. FURTHERMORE, IF THE BEHAVIOUR OF THE MEMRISTOR CAN BE EMULATED BY A CIRCUIT OF PASSIVE R,L,C ELEMENTS AS CLAIMED IN THE PAPER, THEN IT CANNOT BE A FUNDAMENTAL ELEMENT.

In a smaller sense the Chua's postulation story seems to be similar to that of Mendeleev in chemistry. If so, a revolution not only in electronics can be expected. I am happy to be a witness of this historic event. It is really great. My congratulations!

This discovery is being presented (above) in a sensationalistic way that completely ignores (shows ignorance of) a famous paper called
(approximately) THE UNREASONABLE SUCCES OF MATHEMATICS IN THE PHYSICAL SCIENCES, and a sequel to it.

There is / was a standing "37 year mystery"
to be "solved" only if you believe that mathematics has predictive power that can be extended to complete generality, a false method of reasoning by induction that goes back to the 19th century.

Richard Hofstadter's ANTI-INTELLECTUALISM IN AMERICAN LIFE noted in passing that (even our best) universities harbor thinkers insufficiently critical of their own views in precisely the ways needed to avoid such nonsense as that above.

The historical view that the present incarnation of this device was presaged in a "prediction" based on other principles is false too, as explained (more or less) by R. G. Collingwood in his works on historical method.

The actual physical world is much more diverse than the canon we perpetually adopt to explain it, and the habitual process of this adoption dulls many minds to that fact.

All the reports of this research are in fact extremely exaggerated and inaccurate. I liked the work, I like Chua, but the stories are way over the top.

Complementing Mr. Armstrong here, I would like to point out that it is not like Chua introduced a missing term in Maxwell's equations, as Maxwell himself did. The memristor is a class of differential equations that are special cases of the Maxwell's equations, just like the other elements also are (that is one of the brightest points of Chua's 1971 paper). To find "pure" memistor was like a gedankenexperiment waiting to be fabricated.

Many of the properties being attributed to the device are simply characteristics of capacitors, or better, non-linear RLC circuits. The actual phenomenon that could not be modeled passively without a memistor was not explained. So I am still to understand the excitement.

Crazy non-linear passive electronic devices are not news. Take a look at stuff like the Hall effect... The idea that the HP researchers were "struggling" to understand the phenomenon only makes me think they did not study electromagnetism at college. They should have analyzed, inferred the equations (which is a "simple" specific case of Maxwell's equations), and then found out that it was equivalent to Chua's device, and not "finally" applying his idea to model the thing. I hope my suspicions that they are excellent experimental researchers but poor theoretical ones is only a fake suspicion caused by the bad journalism.

I specially hated the concept that it would be possible to complement RLC in LTI electronics, what is assumed by the incautious reader that is the subject of the story, since the non-linearity is not mentioned the way it should. That's absurd. The moment you see R ,L and C simply as special cases of the more general complex impedance, you see that. But many people, such as the reporters and non-expert readers, don't have this notion, and are now thinking that high-school linear electronics must be rewritten. That's outrageous.

Non-volatile memory is also hardly news, as Mr. Teuchert pointed out. Few weeks ago I remember hearing some nice news regarding MRAMs... For this "memristor" thing to be better then a capacitor, it must have a low current leakage, which most certainly exists, and be fast. Where are the figures?...

This all is even starting to look to me as a big HP swindle. I smell shameless capitalism. Where is the scientific scrutiny?...

According to the abstract of the Chua paper, "Many circuit-theoretic properties of memistors are derived. It is shown that this element exhibits some peculiar behavior different from that exhibited by resistors, inductors, or capacitors. These properties lead to a number of unique applications which cannot be realized with RLC networks alone."

James Whong wrote "THE BEHAVIOUR OF THE MEMRISTOR CAN BE EMULATED BY A CIRCUIT OF PASSIVE R,L,C ELEMENTS AS CLAIMED IN THE PAPER, ..."

To which paper is James Whong referring?

What should be kept in mind is that the main subject of the article is the simple and elegant manner in which the concept of a memristor is created. Of course a memristor can be emulated with a cumbersome collection of resistor, capacitors, inductors and amplifiers but the discovery of this device provides a particularly efficient memory device which does not require constant re-freshing nor does it involve current leakage. And as an added bonus, it can easily be explained with an equation which was previously only a mathematical oddity.

The behavior of the TiO2 "memristor" is misrepresented as dØ = Mdq. The operation of the TiO2 device mentioned in the article is not based on changes in stored magnetic flux caused by charge flow, but rather changes in resistance driven by charge flow (which in turn drives mobile dopant flow). Its operation would best be described as
dR = Mdq.

While the device may turn out to be useful as a memory element, its operation is not that of the theoretical memristor.