Best terminal branches I found are:
* HP created something they claimed was a memristor [1]
* there were thermodynamic arguments about why it wouldn't work [2]
* HP discontinued the architecture based on the component [3]
1. https://arstechnica.com/information-technology/2014/06/hp-la...
2. https://arxiv.org/ftp/arxiv/papers/1207/1207.7319.pdf
3. https://www.theregister.co.uk/2016/11/29/hp_labs_delivered_m...
One thing about the OP paper smells very fishy to me:
About two years ago, an extremely important result about Memristors and the notion of circuit elements in general came out. Unfortunately, it didn't get a lot of attention, and remains not-that-well-known even among experts. But: Anyone making claims like the OP paper would HAVE TO consider it in their model if they want to get taken serious. Here's the paper:
https://hal.archives-ouvertes.fr/hal-01322396/document Abdelouahab, Mohammed-Salah; Lozi, René; Chua, Leon O. - Memfractance: A Mathematical Paradigm for Circuit Elements with Memory [2016]
The fact that the OP paper doesn't even mention Memfractance makes me highly suspicious, and doubly so as the OP paper also brings up a "periodic table" of circuit elements - something the Memfractance paper also does (embedded in ℝ²), but uh... let's say with a lot more behind it. In fact, I'll quote the entire "Conclusions" section of the Memfractance paper here:
>In this paper, we have used fractional calculus in order to generalize and provide a mathematical frame for circuit elements with memory: memfractance. We have emphasized that the memfractance is a general paradigm for unifying and enlarging the family of memristive, memcapacitive and memin ductive elements. The motivation and significance of this paper is that there may exist future nano-electronics devices that are more realistically modeled with memfractance elements
>We have generalized the definition of fractance which was first introduced in 1983, and after that, introduced the paradigm of memfractance which is fitted for circuit elements with memory such as memristor, meminductor, memcapacitor and second-order memristor first introduced here. We have defined a new element called memfractor which possesses interpolated characteristics between those four circuit elements.
>We have then generalized Ohm’s law to memfractor and proved it. A particular, albeit wide-ranging, case of memfractance: the interpolated memfractance has been carefully studied through several numerical illustrative examples. Special attention has been devoted to the interpolated char- acteristic of a memfractor lying between memristor and memcapacitor which exhibits an unexpected new behavior of time variation of flux (φ–t curve). This phenomenon has been studied very carefully by the means of rigorous proofs.
>Finally, following Chua’s recent work in which an infinite discrete family of circuit elements: the (α,β) element is introduced in the scope of developing a rigorous mathematical theory of nonlinear circuits, we extend the previous generalized Ohm’s law in order to embed memfractors elements into this periodic table. For this aim, we define an infinite continued family of circuit elements including circuit elements with memory (such as second-order memcapacitor and meminductor and third-order memristor), with a special metric. We call this family: fractional circuit element family.
Note: That last passage is a reference to this specific paper:
https://www.semanticscholar.org/paper/The-Fourth-Element-Chu... - Chua, Leon O. - The Fourth Element [2012]
Which the OP paper also doesn't cite.
To me, this looks like a FUD campaign by Intel.
I also briefly glanced over the OP paper besides checking for citations in it, and found this passage:
"We have borrowed the expression “periodic table” as attributed to Chua" - which cites a secondary resource from 2008, before Chua's two above papers. Looking for any words starting with, or containing a fragment of, "fract" in the paper also yields no results.
Looks like the typical outcome of someone who thinks they don't suck at literature research, but does suck at literature research, which, quite honestly, the vast majority of scientists do. Ask a randomly sampled researcher if they ever heard the term "Main path analysis" and they'll stare at you like a deer staring into headlights.
Having said that:
The paper does seem to bring up a few valid points, e.g. people making errors & drawing bad conclusions from mislabeled diagrams. This isn't an issue in Chua's work, however.
Oh, and: Here, some slides on memfractance:
http://slideplayer.com/slide/3869533/
Edit:
Actually, this got generalized even further in early 2017:
https://www.sciencedirect.com/science/article/pii/S100757041... Machado, J.Tenreiro; Galhano, Alexandra M. - Generalized two-port elements
That sounds like a hoax to me. I had to look up fractional calculus, but it appears to be a real thing. It still sounds ridiculous and appears to bring in memcapitive and meminductive elements to their odd new world. At the end of the day, writing papers with strange terms isn't nearly as useful (to the rest of us) as actually producing a new device.
(1) https://www.mathpages.com/home/kmath616/kmath616.htm (2) https://arxiv.org/abs/1009.5533
Fractional calculus is a well established tool to develop physical models, particularly rheological and damping models. It makes sense to apply the same technique to model circuit components as rheological and electronic models do share quite a bit of common ground, and in some cases even symbolic representation.
I've heard people suggest Intel might be acting like this due to (HP?) patent issues? (Edit: to be clear this was just interesting musing by an outside observer like the rest of us, not some source with any authority.)
https://news.ycombinator.com/item?id=9962755
Seems possible that this might indeed constitute part of some ply by Intel, trying to bypass having to pay patent/licensing when actually using their 3D Xpoint tech.