All posts by robertmcgrath

Bees Know How To Make Flowers Bloom Earlier

Bees are neat!  We keep learning all kinds of stuff about these little critters, who seem to be astonishing problem solvers.

This spring researchers from ETH Zurich report a very interesting behavior that Lars Chittka describes as “horticulture” [1].  They observed bees deliberately damage flowers, apparently in order to make them bloom sooner, so the bees can get pollen sooner [3]. <<link>>


This is a very nice study.

They induced bees to do this trick, and also tried to replicate the damage mechanically.  The experiment showed that the bee damage induced flowers to bloom 30 days or more early!  Interestingly, the human efforts to replicate the bees were far less successful, inducing flowers 5 days early.


The researchers demonstrate that this behavior is likely linked to a lack of pollen around.  I.e., the bees are starving, so they prod key sources to speed up the arrival of food sources.

One of the interesting things is how specific, indeed, fiddly the damage is.  Whatever the bees are doing, it really works.  But the researchers couldn’t replicate it very well, so who knows exactly how it works?  There is nothing obvious about the specific patterns that would seem to explain the effects.

Bombus terrestris workers damaging Solanum melongena leaves. (A) Sequential images of a worker penetrating a leaf with its proboscis (taken over ~3 s). (B) A worker cutting into a leaf with its mandibles. (C) Characteristic bee-inflicted damage. (From [3])
This looks more like acupunture than horticulture to me!

Once again, bees seem to know something that we don’t know.

  1. Lars Chittka, The secret lives of bees as horticulturists? Science, 368 (6493):824, 2020.
  2. Matt McGrath, Nature: Bumblebees’ ‘clever trick’ fools plants into flowering, in, May 21, 2020.
  3. Foteini G. Pashalidou, Harriet Lambert, Thomas Peybernes, Mark C. Mescher, and Consuelo M. De Moraes, Bumble bees damage plant leaves and accelerate flower production when pollen is scarce. Science, 368 (6493):881, 2020.

Book Review: “Tacky’s Revolt” by Vincent Brown

Tacky’s Revolt by Vincent Brown

Until this year, I knew very little about the history of Jamaica or the Caribbean in general, so this book was a lot of new material for me.

Tacky’s Revolt was one of a series of slave uprisings in Jamaica (and elsewhere) in the 1700’s.  Jamaica was the most profitable colony in the British sphere, and that profit was produced by slave labor.  Brutal, lethal work, mostly done by captives from Africa.  Surprisingly enough, the enslaved people were not happy with this arrangement, and, from time to time, rose up to take control of their own lives and fortunes.

Brown gives a complicated and nuanced history of these pivotal fights.  He is at pains to portray the multiple simultaneous conflicts in play, imperial transatlantic wars, African imperial wars, and “intimate” master-slave wars.

Within these conflicts, there were many parties on each side with competing interests.  The colonists feared foreign invasion and domestic slave rebellion.  Large planters and absentee owners sought maximum production, smaller businesses and workers needed security, the military forces needed to project power and to see to imperial goals not just commerce.

The slaves were recent captives from different parts of Africa, and different social strata, as well as island born “creoles”.  These people did not necessarily share anything other than the misery of bondage.

There were also “maroons”, descendants of escaped slaves who had achieved a temporary recognition and sovereignty from the colonial government, and worked as war time allies by treaty.

Over this diversity lay the politics and ideology of race, which was conceived to be linked to both ancestry (European versus African) and skin color.  In any case, as always, the concepts were messy, with many people falling into “mixed” categories, and increasing numbers of people born on island no matter what their heritage and skin tone.

The institution of slavery was messy, too.  While whites could employ arbitrary and unlimited violence to keep slaves in line, everyday life was more nuanced.  Slave owners were responsible for keeping their thousands of slaves in order, and this required both force and inducement. Some slaves were relatively privileged, and any who had a semblance of security might resist an uprising that risked what they had.

Relations between creoles and freed men in the cities and the maroons in the hills were complicated regardless of skin colors.

For that matter, the white colonists and government forces were governed by force as well.  The military forces were notoriously brutal, and  in times of uprising. martial law required whites to military service, at the cost of private profit.

If this sounds like hell, Brown will tell us that it was in fact continuous war.  Wars between European powers drove competition for colonies, which needed slave labor to be profitable. Wars in Africa fed the European slave trade, which both needed and sustained the overseas colonies.  Controlling thousands of slaves was essentially a continuous war of the masters against the slaves, one which only intensified over the years.


The actual incidents recounted in the book aren’t actually that spectacular.  The slave rebellions were short lived and unsuccessful. The reactions were predictable and brutal. The stories told are predictably twisted and awful.

However, these unsuccessful uprisings were not without wider consequences.  The Jamaican revolt pushed London to reform the government of colonies, taking more control and demanding more revenue.  Applied to North America, this policy led to extreme consequences a few years later in the 1770s.

The uprisings also fed into the development of the ideology of racism, in part to justify slavery and the suppression of slaves.  Brown points out that the events also fed into the beginnings of the anti-slavery movement, demonstrating the costs and moral peril of the slave trade and slavery in general.

Prof. Brown is an academic, and he is pretty careful with  sources which are imperfect and one sided.  We learn quite a bit about how the Africans in Africa and America might have been thinking, though we have little firm evidence about much of this.  Brown interestingly juxtaposes African perspectives with the thinking of the colonists, who were ignorant, misinformed, and often projected their fears onto the unknown masses of slaves and native Africans.

I must say that much of this book is filled with peculiar fussing about geography and conceptual geography that he calls “spatial history“.   A lot of this seems to be over done, dressing up simple history with academic jargon.

Still, it’s an interesting and gripping story, regardless of these academic flourishes.

  1. Vincent Brown, Tacky’s Revolt: The Story of an Atlantic Slave War, Cambridge, Harvard, 2020.


Sunday Book Reviews

Solar Farms: Location, Location, Location

Back when I was a lad, renewable energy was gee whiz Buck Rogers technology.  These days it is real, cheaper than coal and rapidly displacing fossil fueled power generation.  And, as such, it is now as much economic and political as technical.

One of the growing questions for renewable energy is the impact of large scale wind and solar farms.  Wind and solar are clean sources of energy, but large, sprawling complexes on land and sea do have impacts on the environment and on humans.

So, for instance, a gigantic sprawling solar farm in the Mohave will certainly ve a big source of Carbon free energy [1].  But the area is a fragile and beautiful wilderness, so the project is going to have unfortunately costs.

On the other hand, the good news is that there are plenty of places to put solar farms that are far less problematic, so there is no need to deploy in the most sensitive places.  The trick is to get renewable energy on the best sites. This is a sometimes delicate socio-economic-political dance, not least because there are plenty of interested parties.

The Nature Conservancy and similar groups are stepping up to this task in the Flint Hills and in Appalachia [2] (and I assume, elsewhere).  In each of these cases, the key is solid, very solid, local roots, and a can do attitude.  We want development and we want renewable energy.  And in the case of Appalachia, we have lots of former coal mines that offer opportunities for solar farms as part of a reclamation process.

This is clearly the right approach, and I’m very glad to see that people are working hard in the right directions.

  1. Rachel Koning Beals, Warren Buffett-backed largest U.S. solar project approved as nation’s renewable use on track to pass coal, in MarketWatch, May 14, 2020.
  2. Cynthia Shahan, The Nature Conservancy To Unlock Solar Potential On Retired Appalachian Coal Mine Land, in CleanTechnia, .May 18, 2020.


Hydrogen on Tap

Sun and wind are plentiful and free, but they aren’t really fuel.  So—H is for Hydrogen!  Abundant and clean, for sure, but not so easy to get or use.

Researchers from Purdue have been working on a cool technique that stores Hydrogen in chunks of Aluminum alloy, releasing Hydrogen when drizzled with water [1].

Their company web site for AlGalCo suggests that they are at “Version 5.0” of the concept.  In a sign that the technology is maturing, they have evolved the name from “Hydrogen on Demand” and “Hydrogen on Command” (HOD, HOC) to the much more marketable “Hydrogen on Tap” (HOT!).

This is a really neat idea.

The demonstration is designed for vehicles.  A rack of cylinders contains “buttons” of Aluminum Gallium allow.  When the truck starts, water is drizzled on the buttons, releasing Hydrogen which is fed to the engine.  No petroleum is needed until the hydrogen runs out!

The process uses up the buttons, leaving a chunk of Aluminum oxide that can be recycled to create new buttons (or anything else you want Al for).  The cylinders can be swapped out for recharging.

One thing I like is that the Hydrogen is pretty safe in this form (though you want to keep water away from the buttons).  The inputs are pretty clean, and the outputs are very clean.

I don’t have information about the process of charging or recharging the alloy, which presumably sucks energy.  Obviously, this process should draw from renewable energy, perhaps storing power from peak times.

The current demo only partly supplies the truck, reducing fuel consumption by 15% or so.  And, at 45 kg, it’s pretty chunky.

Kurt Koehler, founder and president of AlGalCo, shows his HOT (Hydrogen on Tap) system on 9 April 2020. It sits in a City of Carmel Street Department truck. (Photo: Kelly Wilkinson/Indy Star/USA TODAY Network/Imagn Content Services) (From [1])
Electric vehicles seem to be far more practical than Hydrogen overall, not least because of a much better support infrastructure.

However, this HOT technology does compare favorably with Hydrogen fuel cells, which are awfully complicated and fiddly.  So HOT may be very useful for commercial trucks or ships, for instance, which run a lot and can deal with the weight.

Neat!  Or, should I say, “Hot!”

  1. Maria Gallucci, ‘Hydrogen-On-Tap’ Device Turns Trucks Into Fuel-Efficient Vehicles, in IEEE Spectrum – Energywise, May 1, 2020.

Abuse of the Lightning Network

One of the more successful efforts to reengineer Bitcoin to make it more scalable has been the Lightning Network.   This approach deals with the bottleneck of the single global ledger by—wait for it—adding another layer of indirection!  In cryptoland, this is called a “sidechain”.

The basic idea is for parties to establish a private channel for, a ledger for two, so to speak, where they can post transactions.  This channel is authenticated by registering (and ultimately deregistering) on the Bitcoin blockchain using multisignature cryptography.  Once set up, the participants can put whatever transactions they want on the side channel without any transactions on the main blockchain.

The point is that these side transactions are much faster, since they are a more direct move between the two parties, and don’t require confirmation (or fees) from the gigantic Bitcoin network.  When business is complete, the sidechain is terminated, and whatever balances are transferred via regular blockchain transactions.

Notably, once closed, there is no record of any of the side transactions, so there is an additional level of concealment/privacy for the parties.

Of course, the downside is that this is a separate network and protocol, rooted in Bitcoin but not really protected or supported by the strength of the Bitcoin network.  There is also a certain amount of complexity in the additional layer of protocol.

But, yeah.  This is certainly one way to get better latency and lower the costs of the main Bitcoin network.  After all, the Nakamotoan notion that every transaction, everywhere, for all time; should be on one ledger is, well, insane.  There obviously should be and will be side channels.

This spring researchers at Florida International University report that these LN side channels are, in fact, side channels that can be used for nefarious purposes [2].

Specifically, they show how to command and control (C&C) a botnet via the LN.  Bitcoin and other cryptocurrencies have been a favorite method for the payment of ransoms to botnets, but public blockchains are not ideal for C&C due to latency and the permanent record left behind. The Lightning Network is designed pretty much to solve these problems for financial transactions, so it isn’t terribly surprising to learn that it is well suited to covert C&C of botnets.

The basic design is to organize the botnet as a collection of small botnets, each receiving commands from a LN wallet.  The commands are sent as part of transactions.  If I understand correctly, the message is encoded in the numerical amount of the cryptocurrency.  I.e., a transfer of ‘7 sataoshi’ is read as ‘7’ in a codebook.

This method bleeds cryptocurrency, of course, but the transactions can be looped back to recycle most of the funds.  So, contrary to some naive arguments, the fee structure is no deterrent to misuse of the network.


This messaging is fast (compared to Bitcoin), scales to very large numbers of bots, and with recycling, it is really cheap to set up and operate.

Finally, when the botnet is torn down, the LN channels are closed the funds are recovered and all record of the transactions is gone.   So the history of the activity is concealed.

This kind of botnet is difficult to disable, because the Bitcoin and LN are very robust and “censorship proof”.  The researchers discuss the difficulty of identifying the location of the botmaster, due to the strong privacy of the LN.

One interesting point is that the C&C network is a completely normal and legitimate use of LN.  The botnet pays its dues, and follows the rules.

To me, this means that any users of the LN and Bitcoin are indirectly supporting the criminal activities, probably with little way for other users to know what is happening.  That’s kind of annoying, and also kinda makes using the LN a criminal activity.

  1. Diana Hernandez-Alende, Researchers find Bitcoin’s Lightning network susceptible to cyberattacks in FIU News, May 11, 2020.
  2. Ahmet Kurt, Enes Erdin, Mumin Cebe, Kemal Akkaya, and A. Selcuk Uluagac, LNBot: A Covert Hybrid Botnet on Bitcoin Lightning Network for Fun and Profit. arXiv arXiv:1912.10617, 2020.


Cryptocurrency Thursday

Dogs and Robots

My cat will tell you that dogs are stupid.  They are certainly a lot more likely to pay attention to what humans say to them than cats are.

But both cats and dogs are quite able to ignore voices from a speaker, e.g., a TV or computer.  They do pay attention to cat, dog, and other animal sounds, but generally don’t worry about the people talking.

So we know that dogs have some concept of a difference between speech from a person who is present, and the sound of a person who is not present.

But what about “human like” robots?

A “social” robot is designed to present various human attributes and behaviors, so as to appear sort of human—to a person.  What these attributes and behaviors should be is an open question, as is the question of how different people perceive them.

This spring researchers at Yale explored how dogs think about human-like robots [2].

The new study examines whether these human-oriented attributes are perceived and acted on by dogs.  Clearly dogs can tell the difference between a human and a machine.  They can also perceive the simulation of a simulated human, though they may or may not perceive the “humanness” of the simulation.  For one thing, a robot does not smell like a person, does it?

The study compared a loudspeaker with a (very toylike) robot.  The device called the name of the dog, and gave a “sit” command.  In the first condition, the dog was observed to see if he or she gazed at the source of his name.  The second condition observed if the dog sat in response to the command.

The results showed that the dog seemed to attend to the robot more than the disembodied loudspeaker.  I.e., the dog was more likely to gaze at the robot, and to sit.  The researchers conclude that “The dogs in our study reacted socially to a social robot, and the robot seemed to affect the dogs’ behaviors.” ([2], p. 22)

The researchers speculate that the dogs perceive the social robot as an “agent”, akin to a human.

Well, maybe.

It certainly seems as if the dogs could perceive the difference between the social robot and the bare speaker.  However, there were two people present with the dog in each trial, and the people surely knew the difference—not to mention the aim of the study.  The video does not give us a good view of the humans, so it’s hard to tell what they were doing.

Can you say, “Clever Hans”?

We also don’t really know how much prior experience the dogs may have with robots (probably not much) and speakers (possibly a lot).  So there may be issues of familiarity and novelty.

And above all, the robot itself isn’t particularly human-like, is it?  I would never mistake it for a human, so why would a dog?  So, whatever is going on, I have to really wonder if the imagined “social” cues even exist.

This is an interesting study, and it is certainly nice to consider a broader notion of the perception of “social” robots, as well as a proper curiosity about what non-humans might think of our silly toy robots.

To me, it is an open question whether robots designed to be “social” for humans should or should not be perceived as human-like by dogs.  I tend to think not.  However, I strongly suspect that dogs can pick up what the people around them think about the robot, and might well play along with the game.

In short, this may be a complicated and indirect case of Clever Hans in the twenty first century.

  1. Evan Ackerman, Dogs Obey Commands Given by Social Robots, in IEEE Spectrum – Robotics, May 13, 2020.
  2. Meiying Qin, Yiyun Huang, Ellen Stumph, Laurie Santos, and Brian Scassellati, Dog Sit! Domestic Dogs (Canis familiaris) Follow a Robot’s Sit Commands, in Companion of the 2020 ACM/IEEE International Conference on Human-Robot Interaction. 2020, Association for Computing Machinery: Cambridge, United Kingdom. p. 16–24.


Robot Wednesday

Riehle on ‘The innovations of “open source”’

Over the course of my career, the concept of “open source software” emerged from the cultures of academic research and hobbyists (which heavily overlapped in early years).  It has now grown and been embraced by governments and corporations, and the concept has spread beyond software, to hardware, social processes, and politics.

This idea of open source, as well as a handful of super successful cases (Linux! Gnu C++!), have been heralded as major technical and social innovations. This rhetoric is often muddy, and incoherent.

Dirk Riehle writes in IEEE Computer summarizing the main categories of innovation as “legal, process, tool, and business models” ([1], p. 59)

The legal innovations are the well-know, if massively confusing, “open source licensing” for software.  These licenses have been brilliantly organized by Creative Commons, and now can be applied with precision and finesse to lots of things besides code.

The process innovations pioneered decentralized collaboration and peer review. These methods employ digital network technology to implement some form of a self-organizing meritocracy to develop and maintain code that is “owned” by everybody.  This approach is certainly a radical departure from centralized requirements driven software development.  If nothing else, it is impossible to know what the actual product will actually become, because that depends on the emerging consensus of the developers. : – )

One of the huge successes of open source software has been in tool development, in two senses.  Software tools are notoriously expensive to create and sustain, but access to good tools is essential. Many open source projects are tools that everybody needs.

Open source projects have also pioneered tools for collaborative development.  These tools make it possible, indeed, easy, to create and run a distributed work group.  Originally designed for code development, these tools have become the basis for a vast array of collaborations (for better or worse).

The fourth area of innovation emerges from the combination of the other thrusts into business models.  There are plenty of controversies here, not least because some companies have chosen to torture the terminology, declaring their proprietary systems to be “open”, or even claiming private ownership of open source software.

At the base of these business controversies is money; how to economically sustain software and other products.  “Free” is nice, but it doesn’t pay the bills.  So how can the bills get paid?  There are different approaches.  Again, Creative Commons has some of the best analysis of these issues [2].

This article is a pretty reasonable description of the main contributions of open source software.  These innovations are interconnected, but logically separate. You can release anything under an open source license, and you can use open source tools or processes for proprietary work.  And that’s fine:  you should fit the tools and practices to the project.  (See Creative Commons [2.)

Most open source software (as is the case for software in general) is never finished let alone used.  But “open source” has been a very significant innovation indeed: the concept of open source software has become a major cultural meme, called upon in many contexts.

For example, in early days, coworking was conceived as an “open source” work environment [3].  Bitcoin and other cryptocurrencies not only use open source software and methods, they enshrine decentralized “consensus” in the very core of the protocols (for better or worse).  And, as noted, Creative Commons has extended the concepts to digital objects of all kinds.

  1. Dirk Riehle, The Innovations of Open Source. Computer, 52 (4):59-63, 2019.
  2. Paul Stacey and Sarah Hinchli Pearson, Made With Creative Commons. 2017, Ctrl+Alt+Delete Books: Copenhagen.
  3. The Coworking Wiki, Coworking Manifesto (global – for the world) in The Coworking Wiki, 2015.