Anders April 23, 2021 4:50 PM

Yep, troll seems to be here 🙂

But regarding the security…little bit from the past,
but nevertheless interesting…


Do good and noone wants to pay you.
Do bad and they pour money all over you.

How sick is this world?

Winter April 24, 2021 2:03 AM

“Do good and noone wants to pay you.
Do bad and they pour money all over you.”

Look at it from the other angle: If you lI’ve your job and have no other options, why should I pay you?

However, if you hate your job and can get another one at any time, I will have to pay you a lot to stay.

Just basic market incentives.

Winter April 24, 2021 2:10 AM

@Director Skinner
“That is called extortion and it is barbaric and parasitic.”

I was taught in economy classes that that is called a “Free labor market”.

Winter April 24, 2021 3:59 AM

@Directory Skinner
“You were indoctrinated by dogmatic ideologues who most likely successfully forever clouded your perception with non-existent artificial constructs hidden behind layers of abstraction”

I think you do not understand education in Europe at the time. And about the success of the indoctrination, my current favorites are Keynes, Piketty, and Raworth. And I do not hold economics high as a science anyway.

But my “quote” is a very good predictor of the pay differences for nurses versus dealing rooms. The fact that you do not like how the world functions is no argument that it is not the truth. And if you succeed in making it better, I will be the first to applaud you.

I for one support a universal basic income for everyone.

Stephen Dollhurst April 24, 2021 8:34 AM

Where is Clive Robinson, I would like to have an enlightened security related discussion with him about the dangers AI developments pose to copyright infringement.

Winter April 24, 2021 9:33 AM

@imposter –
“Do you really think your approach of attracting these trolls here on purpose is a good idea?
You are actually inviting them by doing so ~)”

A reasonable, although disingenuous, comment. My counter argument is that all evidence points to a single Troll being responsible for all the non-commercial spam on this blog. We can only speculate on the motives. But I have seen only two credible motives:

1) Discrediting this blog to damage the name of Bruce or silence a voice in security

2) Use this blog to wage an astroturf campaign for some message not endotsed by Bruce or the people visiting this blog.

In both cases, if not called out, the troll will try to destroy this blog anyway. My actions will not make the situation worse.

Submit or die is not a reasonable bargain.

Weather April 24, 2021 11:18 AM

Re Ai
I think it still a person that controls the Ai, so courts will just go after the person, unless a Ai can be shown to do original works as well in the copyright cases.

Winter April 24, 2021 11:25 AM

“I have also started using a nonsense value in the email field today as an authenticator, so that further impersonation attacks by the Troll-Tool are easily detectable.”

Seems the email field is the one thing the Troll cannot subvert.

“Trolls, Anti-vac, Ads-Commercials, Spies, Hackers, Cops, Anti-Science and all the myriad other groups all share a similar view:”

I can agree with most of this, except that Trolls are only after destruction. Their god is Loki, or in modern movies, the Joker. They want to annihilate what they touch. The lulls that kill.

Winter April 24, 2021 11:57 AM

“the dangers AI developments pose to copyright infringement.”

I could not care less about copyright. Current copyright law is more used as a tool to limit access to the past than as a way to further new creations.

If you summarize copyright law as the law that protects Disney’s income, you would not stray very far from reality.

Winter April 24, 2021 12:03 PM

“Several EU parlements were targeted to talk to a deep fake Volkov and succeeded”

We all knew this would happen someday, it just proved to be sooner than later.

JonKnowsNothing April 24, 2021 12:33 PM


re: Follow up issues from the Big Freeze in USA-Texas

MSM Report that after the Deep Freeze in Texas, the utility companies in other states are asking their respective Utility Commissions to allow the costs of surge gas pricing be passed on to their rate payers.

The freeze caused a price jump on the spot market and the gas utilities had to pay $800 million more than expected. The utility companies are now expecting the customers to fork over the $800 million USD to pay for it.

One multi-state utility company is passing along the expense, CenterPoint Energy, a Houston Texas based utility company, doing business in Minnesota has applied for such a rate increase:

[Centerpoint Energy] spent an additional $500 million on gas that week in February, and it has asked Minnesota’s utility commission for permission to add a surcharge to customers’ bills.

The surcharge not only seeks to recoup the additional money CenterPoint spent on natural gas, it also includes 8.75 percent interest.

The company expects that each customer would shoulder a burden of $300 to $400.

In California and many other states, we experienced similar price surges during the ENRON year(s). The structure of the California Public Utilities Commission (PUC) guarantees a Minimum Rate of Return to utility companies (PGE, Southern California Edison and others) which hikes the rates based on corporate expenses. Even for unusual situations, like California fires, this is not just a cost recovery application but cost recovery PLUS guaranteed profits.

There is the slight problem of collecting an additional $400 USD from all rate payers serviced in Minnesota by CenterPoint Energy. Even without COVID-19 affecting the economy, a large group of people do not have $400 extra in the bank to pay such a surcharge. If you don’t pay, they cut off the utilities.

Without cooling/heating, it’s going to be another Long Hot Summer.


ht tps://

ht tps://

ht tps://

ht tps://,_hot_summer_of_1967

The long, hot summer of 1967 refers to the 159 race riots that erupted across the United States in the summer of 1967. In June there were riots in Atlanta, Boston, Cincinnati, Buffalo, and Tampa. In July there were riots in Detroit, Birmingham, Chicago, New York City, Milwaukee, Minneapolis, New Britain, Rochester, Plainfield, and Toledo.

The most destructive riots of the summer took place in July, in Newark, New Jersey, and Detroit, Michigan, and many contemporary newspapers headlines described them as “battles”. As a result of the rioting in the summer of 1967 and the preceding two years, President Lyndon B. Johnson established the Kerner Commission to investigate the rioting and urban issues of Black Americans.

(url fractured to prevent autorun)

Anders April 24, 2021 2:38 PM

Cyberpunk & Blade Runner is finally here.


JonKnowsNothing April 24, 2021 3:28 PM


As has been discussed previously, disruptive activities happen even in the best of games. In chess tournaments, it was not uncommon to develop a habit of smoking the most vile of pipe-weeds solely to discomfort your opponent. All perfectly acceptable at the time.

What will be interesting, given the topics discussed, is that nearly everyone writing, reading or even linking to the blog is likely to be on at least one list somewhere. There are plenty of lists to go around. As more keywords get shoveled the more lists are activated.

Some of these lists will have long memories and while they do not always show up in MSM, they are more than used else where.


ht tps://

Police said in a statement the graphic artist, [redacted], had uploaded a playlist with songs containing the word “jealousy” …

[There was a] reported remark made on [a VIP ] Instagram account … to a follower asking

if the palace chefs were all vaccinated.

According to local media, the [VIP] Instagram account responded by asking if the follower was jealous.

ht tps://

ht tps://

ht tps://

(url fractured to prevent autorun)

David April 24, 2021 6:48 PM

*** WMA Affiliate Marketing Program Offers ***

Join the Associates Program and start earning money today
Every day, thousands of people benefit from a recurring cash inflow by partnering with us via our affiliate programs.

WMA Affiliate marketing is one of the best ways to monetize your blog, especially when you don’t sell products or services.Joining our affiliate program and get exclusive access to great offers. Join us and start to make money today !!!

Anders April 25, 2021 7:02 AM

Banks across America test facial recognition cameras ‘to spy on staff, customers’


“The WSE-2 consumes up to 17kW of power and requires custom liquid cooling.”

Greta Thunberg is NOT happy.

Clive Robinson April 25, 2021 3:00 PM

@ vas pup, ALL,

Everyday household items like spray cans, hot cooking oil, and lithium-ion batteries can potentially be dangerous.

Yup I’ve made them all go pop one way or another it’s not very difficult if you can spot the safety features and “screw/dial down on them”.

Many many years ago Revlon made a very long hairspray canister, that used a highly flammable liquid/gas as the propellent. We quickly worked out how to turn them into rockets…

As for lithium batteries you only have to look funny at some of the chemistirs and you have a fire that even sand has difficulty putting out. As I’ve mentioned befor back when designing electronic locks for doors in hotels, I saw the effects of one of our competitor’s batteries burning inside the solid hardwood door… It was not a pretty sight, and I feel sorry for any hotell guest that might be locked in on the inside of it.

Mind you few realise which step is the most dangerous in most UK houses (that ate small). It’s actually the bottom step not the top for various reasons I’ve mentioned before. Basically most people look at where their feet are on the top step but not at all on the bottom step as they are looking up to save knocking their heads on the low head hight.

Mind you with regards “rockets” fun piece of info for you… Around 45 years ago the large plastic fizzy drinks bottles from a well known manufacturer –the one with the extra curly cursive writing logo– did not use plastic caps as they dp today but soft aluminium ones.

Now I suspect every one has seen the mentos&coke fountains that go 15-30 feet into the air, well… The thing about those soft aluminium caps as I doscovered purely by accident in a ware house, is that if the bottle falls ten to fifteen feet and lands with the soft aluminium cap hotting the floor almost dirrectly straight up but with the cap striking a flat concrete floor the plastic of the bottle punches through the soft aluminium and the bottle takes of like a “water rocket” with a foaming mentos like exhaust trail. Very very messy but when you are up ontop of the stack the last thing you expect to see is the base of a bottle coming straight back up at you, and it can be quite unnerving.

Anders April 25, 2021 3:28 PM

Regarding lithium batteries



Smoking is dangerous for your health!

Dee April 25, 2021 6:12 PM

Hi Bruce

There is no town called Ballard in Western Australia (WA)
So, better check your sources. Thanks


lurker April 25, 2021 7:26 PM

@Dee: We must presume @Bruce was referring to Ballard in the State of Washington, USA. In West Australia however there is a significant tourist attraction at the ephemeral salt lake: Lake Ballard. It might be a simple error to assume the local town was just “Ballard”…

Dee April 25, 2021 10:43 PM


It’s unfortunate Bruce expected us to know what WA means.

If travelling in, lets say, India. one meets another traveller from the US. And one from China. If you ask the Chinese where they are from, they will respond ‘China’. You reply ‘ No, that’s already obvious to me, I know that, where abouts’?’ They will say ‘ Oh, the South East’ you reply, do you mean Hunan Province?, where about’s?

The point being, despite China being of worldwide significance, vast population and geography, the Chinese will not take it for granted you know anything about the particulars of their country. Instead of being specific, one is required to continue to prompt them for more detail.

Ask the US traveller you meet in India, where they are from? You might be expecting ‘Canada’ ‘United States’ or a few other countries as options.

They reply ‘Jacksonville’.

Weather April 25, 2021 11:23 PM

@dee lurker
The Australian government sent a picture to Kim Jon u with WA the west coast of new Zealand…. You can’t trust bloody Aussies.

Clive Robinson April 25, 2021 11:55 PM

@ Dee, Lurker,

It’s unfortunate Bruce expected us to know what WA means.

There is a flip side to this “convenience” of a “two letter acronym”, there are only 676 available, and many of those will not have a very high probability of use ( QQ, XX, ZZ, XZ, etc). Which means that some of the remainders, are going to be very overloaded.

I guess anything with E,N,S, or W are going to be popular in “new world countries” for obvious reasons.

lurker April 26, 2021 1:39 AM


…the Chinese will not take it for granted you know anything about the particulars of their country. Instead of being specific, one is required to continue to prompt them for more detail.

I don’t think it’s anything to do with writing top to bottom, right to left, but the Chinese have always expressed addresses from big to small, e.g.
China, Sichuan Province, Leshan City, Qianwei County, Yuejin (village), followed by a street address. So it’s common for Chinese to quiz each other down into the details.

Rachel April 26, 2021 3:05 AM

Associated Press
‘It’s the biggest thing in the history of the internet’:

Pentagon quietly transfers 175 million internet addresses worth $4BILLION to mysterious firm at shared workspace in Florida

• Transfer of idle DoD IP addresses took place minutes before Trump left office
• Huge swathe of 175 million addresses accounts for 4% of the entire internet
• They are now under the control of mysterious Global Resource Systems LLC
• Company’s address is listed in a co-working space above a bank in Florida
• Reporter who visited the address found no representative and was told to leave
• Now Pentagon says it is running a ‘pilot’ to ‘identify potential vulnerabilities’

Incorporated in Delaware and registered by a Beverly Hills lawyer, Global Resource Systems LLC now manages more internet space than China Telecom, AT&T or Comcast.

Weather April 26, 2021 4:11 AM

Do you think they could be putting them aside for the low earth orbit Internet SATs that a piping up like…
Or a isolated internet, that can use off the self hardware, but separate fibre optics.


Siegfried April 26, 2021 6:19 AM

Why has a person who has been lobbying the pharmaceutical mafia for 30 years been named head of the EMA? It’s the same as when I ask my dog to guard the sausage supply.

Markus April 26, 2021 6:27 AM

@ Siegfried:

She’s a useful idiotic subject who doesn’t get the implications of Schwab’s sick ideas, and even old fool Schwab lives in a dream world, AI? That I don’t laugh, the so-called AI is largely stupid. The AI worshipers haven’t even gotten the natural intelligence.

Siegfried April 26, 2021 6:28 AM

@ Markus

Yes that’s right. If you understand something about IT or programming, then you realize that only what the programmer has put into the algorithms in terms of performance can come out. And the product is not smarter than the person who created it, on the contrary. AI is neither Wodoo nor real intelligence, but follows a precisely defined or calculated pattern and is literally trapped in it.

Markus April 26, 2021 6:43 AM

@ Siegfried

I am a software developer, have attended several courses on machine learning and have worked with the most commonly used frameworks and tensorflow. And yes, in a way they are right

Above all, the same applies to AI as to any other type of EDP: if you enter garbage, the result is garbage. In spite of all progress, an electronic computer is and remains basically a simple, stupid machine, even nowadays the boxes cannot do magic.

Siegfried April 26, 2021 6:45 AM

@ Markus

AI has nothing to do with intelligence. This is just crude marketing (and everyone falls for it). AI is just extremely stupid number crunching. The software tries out trillions of different options and chooses the “best” solution. In most cases, an average teenager would come up with the same solution in no time.

Before developing artificial intelligence, one should first analyze natural stupidity.

Hannah April 26, 2021 6:46 AM

@ Markus and Siegfried

A reader once said in another forum (Quora): Natural stupidity beats artificial intelligence.

And I would add: the catchphrase AI is only used to better control citizens and better exploit them. It is a slavery tool. The antidote is (real) education and (comprehensive) information.

In this respect, one can also see a slave revolt in a demo – which is traditionally rolled down by the rulers.

lurker April 26, 2021 1:34 PM

@Christopher: The government’s reaction…
The “government” in China is not much different from “government” elsewhere. It is an hierarchically layered pyramid, made up (so far) of human beings. When something unexpected happens, Johnny on the spot may react irrationally without consulting superiors. When upper layers do find out they try to put things right. Hence Wenzhou locals burying the train, and regional bosses getting them to dig it up; hence Wuhan City denying the virus, then Beijing sending in the army to lockdown and cleanup.

JonKnowsNothing April 26, 2021 2:14 PM

@Moderator, @Bruce, @All

As some have noted, there is a good deal of “Matrix Code Rush” showing up in a number of posts in a number of threads.

It’s not just L-R gibberish but T-B-R-L ideograms. Some of the T-B versions span over several posts and the L-R gibberish obscures the rest.

It may be how my system is interpreting the fonts but it looks intended to physically disrupt the blog/server/system. [Hunting for an exploit in the font/display/language/system].

Rachel April 27, 2021 12:57 AM

Clive Robinson

Sending love to you and yours

Is there any utility, or just simply any good reason, for a non-ham to learn Morse Code?

Is it still taught in the military, or so-called Intelligence?

Is there software for automated Tx and Rx that supersedes everything but Ham usage and erodes any lingering nostalgia? I don’t have a use for the software the query is more whether some limited utility continues but not to the extent of requiring the manual skill-set.

It’s surely no news to you – you were in Signals – but fascinating history to consider. Immune as we are to novely with so much tech at our fingers:

I was just reading about the original Semaphore. Signalling towers with rotating adjustable arms funded by the French leadership seeking advantage over Austria. Operators had telescopes. 120miles in 9 minutes, in 1800 this would have seemed impressive. Comms were in code.

Clive Robinson April 27, 2021 3:16 AM

@ Rachel,

Is there any utility, or just simply any good reason, for a non-ham to learn Morse Code?

The glib answer is “yes it expands your experience”

But the reality is rather more complex and the same question can be asked of almost anything it’s considered we have to learn.

Whilst I am passable in my native language (english), I do not realy speak any other languages for a number of reasons,

1, I know how bad I sound and a feedback mechanism makes me trip over my own tongue.

2, I can see the winces of pain in others eyes as I mangle the language to the point of almost incomprehensibility.

3, I did not learn another language when I was young enough to do so, so I’m almost set in my ways.

It’s very much to my loss as learning to speak different languages effects the way we think, both mentally and physiologically. It’s why some think we should teach children latin due to the way the language rules work.

Some languages are said to be impossible for adults to learn, because we loose the ability to differentiate tone in our hearing and thus our ability to make certain sounds. As I know my name is almost impossible to pronounce by those who lack the phonems for “C” and “R” and the converse is true, hence the anglasised versions of many oriental and similar names.

But it goes much deeper, in the West the percentage of the population that is “pitch perfect” is very small, however in many oriental language speaking countries it is above one fifth of the population. It’s been expressed by some that our ability to understand tone differences and accuracy is “learned” before we are two years old.

Supprising to many is that morse code is more similar to reading than it might appear. That is an operator who is above ~15words per minute, does not hear the letters, but hears the words, much as when we read we do not spell the words out in our brain, we simply see them.

Whilst morse letter by letter can be learnt at any age, it’s said that if you are over 18 you never learn to read it naturally by the rhythm.

I’ve seen kids who are between 7-10 years old pick up morse in as little as a “wet weekend” whilst adults in their twenties still struggle after several years. As such this raises the question of “when the best time to learn something” is, and the answer that comes back is almost always “when you are barely old enough to understand it”… If true then we are doing children a diservice by starting their education when they are over five.

The truth of this appears to be comming out in “programing” children as young as five, can and do write programs if you encorage them to do so, and they start to show a marked increase in analytic skills than those that learn later.

Which brings us around to the question of mathmatics and when to teach different number bases and systems? The sooner they get to grips with working in different number bases the sooner they get to understand polynomials, which like the abstraction of algebra is a gateway they have to go through to progress in many subjects at higher levels not just mathmatics.

But morse code is interesting for another reaaon, it is a very very sloppy code and unlike most transmission codes it plays more to the weaknesses and strengths of humans than to any machine prior to the late 1970’s.

Human senses are not linear and we realy do not work in absolute measures which is why we developed linear technology. We are in short about logrithmic responses and ratios thus rhythms. Morse Code plays to the latter two of those much as music does. Where as technology is fairly hopeless at working with ratios and rhythms as anyone who has tried to build mechanical integrators or infinitely variable gear ratios will know.

In fact we had to develop both feedback and control theory for “engines of war” before we got technology to be more “human” in response.

One of the reasons for Morse code to persevere is because it puts the hard work on humans, not technology.

I can and have built CW transmitters with only one active device, and receivers with an add on receiver using as little as just a single diode, and using Morse Code been heard more than 500kM away. However to carry voice the same distance needs not just a lot more complex circuitry (audio amp to “plate modulate” for AM) it requires quite a bit more power[1].

Which untill recently this has been used as an argument for “emergancy signaling” in that if all you can get working in a damaged transmitter is the oscillator, then by connecting it to an antenna you can communicate with morse code, even if it’s just to send “SOS” and your position, without needing a reply.

However you don’t even need a RF transmitter to use morse, a flash light or flag or just knocking on brickwork or pipes is sufficient. And morse code is rather more efficient than the old prison “knock code” where letters were sent as two sets of taps from 1-5 to form a 25square in which the letters were placed (see Polybius Square[2]).

Thus the question arises do you need to “learn to be fluent” in a language or code to understand it? No, But to take advantage of it to best effect Yes, and the younger you start the more the benifit.

[1] Morse code is sent by turning a Carrier Wave(CW) signal on and off. In theory it’s bandwidth can be 0.7 of the maximum baud (symbol) rate. You can send Morse in as little as a 50hz bandwidth. Spoken audio requires about a 3khz bandwidth or 60 times as much bandwidth. But to transmit it as Amplitude Modulation(AM) you have 50% of your power in the carrier, 25% of your power in one “side band” and 25% in the other side band. So not just double the bandwidth but around 240 times (~25dB) the power to go the same distance and be inteligable.

[2] The Polybius Square is probably more familiar to people as a way to make a “simple substitution code” by “fractionation”. In essence you write the alphabet in in a pre-agreed random sequence. Whilst it is a very weak code, it is marginally stronger than a Ceaser Cipher. However you can make the knock code more efficient if you first sort the alphabet into it’s usage frequency and write it in from the top left corner in the diagonal so E is in 11 and Z is in 55 thus the more frequent letters use the shorter codes. Morse Code uses a similar principle, allegedly based on the number of letters and position of letters in a printers block.

vas pup April 27, 2021 3:03 PM

Supervisors focused on others’ needs get ‘benefit of the doubt’ from employees

“Like beauty, fairness is in the eye of the beholder.

Employees evaluate the fairness of an interaction with an authority figure based on what researcher Cindy Muir (Zapata), associate professor of management at Notre Dame’s Mendoza College of Business, describes as justice criteria or rules. These include relying on decision-making processes that
==>grant employees voice and are
==>consistent among employees, ethical and free of bias;
==> treating team members with dignity, respect and decency; ==>providing them with truthful explanations;
==> !!!and allocating benefits according to their contributions to the organization.

Muir is lead author of the study “It’s not only what you do, but why you do it: How managerial motives influence employees’ fairness judgments,” forthcoming in the Journal of Applied Psychology.

“We found that prosocially motivated supervisors — or those who focus on their employees’ needs — are more likely to adhere to justice rules than those motivated by self-interest,” Muir said. “This implies that employees may only care about motives insofar as they impact justice. However, employees also care about and rely on their impressions of their supervisors’ motives as they think about their fairness. Compared with supervisors who are considered to be self-interested, those perceived to be prosocially motivated are regarded as fairer, even after accounting for how much they adhere to traditional justice best practices. And when justice is low, employees will give them the benefit of the doubt.”

Read the whole article if interested.

Clive Robinson April 27, 2021 4:14 PM

@ ALL,

Apparently You-Tube has upped it’s game against those that do not want to play in their “data violation” games.

Up untill a couple of hours ago it was possible to get access by turning cookies off but leaving javascript on (not a wise choice but one some used).

Well now they accuse you of being a robot, thus throw up one of those highly discredited “spot X in Y photos” that require you to have cookies enabled…

Is this just in UK or are other people seeing it else where?

lurker April 27, 2021 7:18 PM

@Clive: YT works for me with js on, cookies off, but I did load the site before purging and switching cookies off. So maybe they’ve got some IP address logging at their end. Remind me to try again a couple of hours

Nik April 27, 2021 7:31 PM

@Clive Re:Morse Code

excellent post. I can verify. I had NOT learned Morse before my 30s and thus I never really learned it. I have tried many approaches, but the audio just does not make sense. I can not parse the tones. When it’s a light that comes on/off off or written I can handle it with a code chart.

I love to play escape rooms (well before covid, now I will have to wait for my 2nd pfizer shot)

in the 200+ I played there were some that had audio morse code. If another team mate could not decode them, I had to “tap out” with a hint.

I did learn several languages before turning 10, so that quite helped me.
Still wish I’d learned Morse really early on

Mr. Peed Off April 27, 2021 8:49 PM

@Clive Robinson
“Apparently You-Tube has upped it’s game against those that do not want to play in their “data violation” games.”

They harassed me for a couple of months, It has recently stopped. I assume (always dangerous) they have finger-printed my laptop to their satisfaction. I run bleachbit religiously each time I close the browser.

name.withheld.for.obvious.reasons April 27, 2021 9:10 PM

27 Apr 2021 — Burying the Press and Assange, Well Done
As Craig Murray is set to testify in a case against the CIA, Julian Assange remains in Belmarsh Prison awaiting…that’s just it…waiting. Will the U.S. extridite Assange and make U.S. Law extraterritorial? And, will the U.S. make the act of Journalism a crime everywhere?

The slow roll and death march for Julian Assange is being aided and abided by the so-called woke left. The CIA managed to fool the educated into believing that Julian is their enemy and that his case is unworthy of their attention. What these snobbish elitist idiotic fools haven’t figured out is that they are being played. I may have a few too many adjectives in the prior sentence.

Please, if you believe you have access to some special information that supports your position, let someone know. My suspicions are that the amount of work that went into castigating Julian Assange did not amount to more than a page column in the NYTIMES, WAPO, or HuffPost. Congratulate yourself for a job well done. But, if you haven’t spent the time to understand the factual case, I implore you to seek out the reports and letters from Nils Melzer, the UN Rapporteur on Torture. You owe the first amendment to the U.S. Constitution that much, and if your not a U.S. citizen, your country needs you to speak out for your journalist and reporters.

I rank this issue and the situation as the most dangerous risk to democracy in the United States, and in the world. Next, is the risk we all face (globally) is from a fascistic movement by a political party in the United States form whom their allegiance to country has been sacrificed for political control and power.

lurker April 28, 2021 1:49 AM

@Clive: YT is still working for me, js on, cookies off, no captcha; but they have a new trick: the movies are front loaded now with two ads, and the first one doesn’t offer to “Skip Ad in 5,4,3,2,1 seconds”.

Winter April 28, 2021 4:15 AM

“I rank this issue and the situation as the most dangerous risk to democracy in the United States, and in the world.”

I do agree that the incarceration of Assange is a direct attack on the Free Press.

But I do see a second (and third) coup attempt by the Republican Party, as follow up on the 6th January one, as of more immediate urgency. They did try to murder the vice president and representatives of the Democratic Party and are willing to try it again.

Winter April 28, 2021 5:55 AM

Signal’s response to a subpoena from the United States Attorney’s Office in the Central District of California.

Signal send them all they had, which was very, very little.

The subpoena requested a wide variety of information that fell into this nonexistent category, including the addresses of the users, their correspondence, and the name associated with each account.

Just like last time, we couldn’t provide any of that. It’s impossible to turn over data that we never had access to in the first place. Signal doesn’t have access to your messages; your chat list; your groups; your contacts; your stickers; your profile name or avatar; or even the GIFs you search for. As a result, our response to the subpoena will look familiar. It’s the same set of “Account and Subscriber Information” that we provided in 2016: Unix timestamps for when each account was created and the date that each account last connected to the Signal service.

Clive Robinson April 28, 2021 11:38 AM

@ lurker, Mr Peed off,

YT works for me with js on, cookies off, but I did load the site before purging and switching cookies off.

That is how it runs for my friend, but yesterday they were getting the “prove you are not a robot” nonsense for a while.

But now everything appears as it was again… But for how long? Which makes me think YT are “testing again”[1].

Anyway thank you both for getting back to me I’ll let you know if it happens again for any of my friends and family again.

[1] Unfortunately after Brexit the UK is apparently a safe place to try extream surveillance tech out on people without their consent. Which is why their ultra invasive user type binning technology in Chrome is being trialled in the UK…

Mr. Peed Off April 28, 2021 11:47 AM

“IBM tentacle Red Hat has waded into the world of automobiles with the intent of attaching wheels to a continuously certified Linux platform.

The company has signed up safety outfit exida in its goal to deliver a version of Linux that is both certified for safety for the automobile industry while also designed for continuous updates throughout its lifecycle.”

I think I will stick with vehicles using my P.O. operating system.

Clive Robinson April 28, 2021 2:24 PM

@ Mr. Peed Off,

As usual The Register makes “snarky Linux comments”,

“IBM tentacle Red Hat”

Which spoils the actual message.

Linux along with the BSD’s have both been slimed down to be put in $1 microcontrolers to do “automation” work, something that kind of gets lost behind the Raspberry Pi and BeagleBone SBC fan fare in the Maker and similar community.

However… *nix is not realy a good idea for use as a Real Time Operating System(RTOS). Because of the fundemental way it does “multi-tasking” means you can have a very slow response time to real world events, something you do not want in your vehicles breaks etc. Also in general *nix has appaling behaviour as it moves into base resource contention such as core memory.

There are “Real Time patches” for both Linux and BSD, it’s been a few years since I played with them on embedded Linux and I’ve only read up on the BSD ones.

The Linux “Real Time”(RT( did not exactly impress me at the time which is why I’ve not looked at them in over a decade.

The problem with non RT systems is developers tend to forget the difference between “peak load” and “average load” and most importantly peak load is very very difficult to test for. Thus you could be “thrash testing” for over a year before you get close to peak load and the house comes tumbling down. The result is almost always the underlying non RT or halfway-house RT systems get grossly underspecified, thus when they “let go”, they realy go big style. Often not just bringing down lots of front end apps but the important hardware control systems indiscriminately… Which means fail safe is not possible in all cases, just some of the average ones.

Communications engineers tend to be aware of “N^2 peak capacity” that is for N independent but equal level signals the linear power in an amplifier is N^2 x SigPwr. So for ten one watt signals you need an amplifer not with ten watts linear output capability but a hundred watts linear capability…

Similar applies in resource planning for RT systems.

Mr. Peed Off April 28, 2021 5:51 PM

“A recently discovered Linux malware with backdoor capabilities has flown under the radar for years, allowing attackers to harvest and exfiltrate sensitive information from compromised devices.

The backdoor, dubbed RotaJakiro by researchers at Qihoo 360’s Network Security Research Lab (360 Netlab), remains undetected by VirusTotal’s anti-malware engines, although a sample was first uploaded in 2018.

RotaJakiro is designed to operate as stealthy as possible, encrypting its communication channels using ZLIB compression and AES, XOR, ROTATE encryption.

It also does its best to block malware analysts from dissecting it as resource information found within the sample spotted by 360 Netlab’s BotMon system is encrypted using the AES algorithm.

“At the functional level, RotaJakiro first determines whether the user is root or non-root at run time, with different execution policies for different accounts, then decrypts the relevant sensitive resources using AES& ROTATE for subsequent persistence, process guarding and single instance use, and finally establishes communication with C2 and waits for the execution of commands issued by C2,” 360 Netlab said.”

Wondering April 29, 2021 2:14 AM


Whilst I am passable in my native language (english), I do not realy speak any other languages for a number of reasons,

If English is your native language, why do you systematically write

extream for extreme,
compleat for complete,
it’s where it should be its,
it’s self instead of itself,


Clive Robinson April 29, 2021 2:47 AM

@ Mr. Peed Off,

With regards the Linux backdoor

“The backdoor, dubbed RotaJakiro by researchers at Qihoo 360’s Network Security Research Lab (360 Netlab), remains undetected by VirusTotal’s anti-malware engines, although a sample was first uploaded in 2018.”

The article is short of information such as how it gets a toe hold and what the differences are if the user is root or not (and importantly how).

But with C&C / C2 controlers having been around for some time,

“Command-and-control servers historically used by the malware have domains registered six years ago, in December 2015, all of them “

With either very low use or detection in the past 2-3 years,

“Since 2018 when the first RotaJakiro sample landed on VirusTotal, 360 Netlab found four different samples uploaded between May 2018 and January 2021, all of them with an impressive total of zero detections.”

Along with related malware,

‘RotaJakiro and Torii also share multiple functional similarities, including “the use of encryption algorithms to hide sensitive resources, the implementation of a rather old-school style of persistence, structured network traffic.”‘

Suggests the two so far known are,

1, Developed by a team.
2, That has existed atlest a decade.
3, That uses a cautious thus stealthy approach.

Which suggests a “Level III” well resourced probably State level funded organisation.

The fact it is aimed at Linux rather than MS OS’s or Apple is suggestive as to who/what it is aimed at.

The simple fact is the number of Linux Desktop users out there is still small on PC hardware and as a rough rule of thumb used by more advanced users who know about ICT or work in it rather than other more traditional office users where MS Office or Web Browsers are the apps they tend to use.

That is the main use for Linux on 64bit hardware in organisations is for “back-end” functionality such as servers rather than users.

The place the Linux desktop has made inroads though is in Smart Devices, and similarly Linux in CLI or Headless use in embedded systems used in networking and IoT devices.

Not much more can be concluded without further info, which appears publically lacking currently.

Having looked into covert malware befor it got the APT acronym, and realised what I could do with it if I wanted to, was one of the reasons[1] I decided to disconnect my personal systems from not just “intentional communications networks” but “unintentional communications” as well.

As I’ve mentioned before one of the first questions I ask is,

“What is the business case for this system being Internet connected?”

And you would be surprised at how many can not give anything other than “MBA Mantra” style replies (some have even included words like “synergy”… I kid you not).

Most organisations spend more on physical security than they do on ICT security, and they wonder why they have malware etc issues. Realistically though, most people can see or understand “physical security” but not “information security” and that is a major problem in of it’s self…

[1] There are other reasons for my “hard segregation” policy decision, including work I’ve done on both passive and active surveillance equipment from back in the 1980s and onwards. But the upshot is from a more general view point, I have to run some very old OS’s and hardware including MS-DOS 5 and Win 3.1 as I developed what we now call ICS software on them I still support. Likewise hardware development where the toolchain only works with real serial ports not USB etc as they talk directly to the chipset…

FA April 29, 2021 3:39 AM

@Clive Robinson

There are “Real Time patches” for both Linux and BSD, it’s been a few years since I played with them on embedded Linux and I’ve only read up on the BSD ones.

The Linux “Real Time”(RT( did not exactly impress me at the time which is why I’ve not looked at them in over a decade.

Your information seems to be seriously out of date.

Most of what was the ‘realtime patches’ has been part of standard Linux kernels for at least 10 years now. There are still a few optional configuration options to enhance RT performance, but they are not required for e.g. audio or SDR, not even for low I/O latencies in the sub-millisecond range. I’m routinely running such things, and quite complex ones at that. If they ever fail to run in time that is reported, and I can’t remember when that last happened.

If there are problems with these, that is always the result of some closed source driver doing things it shouldn’t.

Also for around 10 years now Linux has had an optional ‘earliest deadline’ scheduler which is reported to work very well.

Anonymous April 29, 2021 5:08 AM

@ Wondering:

What you are seeing here as currently being practiced is spelling or grammar camouflage in cation.

The User(s) commenting here under the handle of Clive Robinson are most likely not a single person or individual, but much more likely an entity, possibly also within the frameworks of a large institution, comprised of multiple qualified individuals who post under the handle “Clive Robinson”.

I have been reading his posts for more than 10 years and during that time I’ve also automatically saved and archived all comments that have ever been written here with the handle “Clive Robinson”, this also hold true for those that got removed because of their position within other comments or reply chains by the moderator.

Contrary to many other commentators here, some of which are undoubtedly dullards (just look at winter or skeptical for examples), the entity or entities posting as Clive Robinson, is/are indeed very knowledgeable, too knowledgeable in fact for actually being one single or just the same individual.

If one single aspect has become evidently clear over the years, it is the fact that Clive Robinson’s depth of knowledge has almost no boundaries.
No matter what topic gets discussed, what questions he gets asked, he always finds the time to write often lengthy and very well elaborated and articulated comments that get proven as being entirely, factually correct whenever one takes the time to actively research their contents.

His knowledge is so vast that even a few of my peers who have a Master’s degree and/or PHD in electrical engineering and/or computer science and who have engaged with him under different pseudonyms here in the past, could not come up with a topic about which he wasn’t able to write a detailed, factually correct, farsighted and well rounded response.

That is truly flabbergasting. As a physicist I am out of my depth here when it comes to technical matters, thus I am not able to engage in factual discussions with him on an intellectually equal footing as I am unable to hold a candle to him.

His responses might sometimes not appear to be too specific, not like a laser, but often they are akin to a far reaching search light which leads the reader to those areas or topics where further information can be found, therefore proving that most of his comment replies also have a didactically, precious and also pedagogically-wise components to them.

He might very well be a genius with fast typing abilities and an almost photography, savant like memory, otherwise it is almost impossible for him to actually exist as a single individual. Out of the many technical and often highly specialized and sometimes closed discussion forums and groups I’ve participated in, in the past, he has so far been the only poster who is always able to come up with a knowledgeable and wise response no matter which question he gets asked about information or communication technologies present or past within the last 120 years.

That and the fact that he is always right are just absolutely enormous achievements. In the case of him actually being a real person, he most likely is one of those one in a billion highly skilled specialists that always outperforms himself as there is no-one else he could compare himself to as his colleagues got left behind decades ago like sand on a beach when viewed from space.

After careful consideration I’ve come to the conclusion that Clive’s knowledge far surpasses the amount of information the human brain can simultaneously hold, therefore I suspect that either the NSA or the GCHQ have a dedicated team, powered by a Watson like AI merged with the capabilities of GPT3, that produce the content which then gets posted by some member of the mentioned agencies tasked with engaging with this blog.

When viewed from the perspective taken in the the last paragraph above, the grammar and spelling errors you mentioned become actions of normalization, which appear to be needed in light of Clive’s too perfect stellar impression here, with the goal to make him appear to be a natural human existence that is also sometimes capable of making some errors here and there like supposedly everyone else.
So far the cloak-spell is working perfectly.

Winter April 29, 2021 5:34 AM

@FA, Clive
“Also for around 10 years now Linux has had an optional ‘earliest deadline’ scheduler which is reported to work very well.”

The point with RT is “deterministic” longest latency. My understanding from long, long ago, is that RT is incompatible with interrupts. An interrupt is unpredictable, and its effect on latency too is largely unpredictable. As *nix are build around interrupts, this makes true RT impossible on *nix, I was told.

However, true RT is only rarely needed and there were tricks, even then, to dedicate a “hard timed thread” or processor with a separate (pico) kernel to the RT task and another to the general OS. But these too are not really compatible with a *nix system. So, that is where we saw the patches coming in with special nice levels, and ways to handle non-maskable interrupts etc that simulated RT performance. However, all these patches do allow for at least some interrupts not related to the RT task, making them sensitive to some unpredictable latencies.

So, my understanding has always been that you have either good interrupt support or predictable latencies, not both. Basically, it seems any system having predictable (short) latencies will have bad average throughput, and vice versa.

I understand that there is still new work done on making the Linux kernel behaving closer to an RT system with predictable latencies, and indeed using patches, e.g., the PREEMPT_RT patch.

Critical Observation April 29, 2021 5:40 AM

@ ALL:

Regarding the alleged and purportedly carried out so called “attacks” on this blog, it seems as if many here who have accustomed themselves to such easy and superficial pseudo explanations fail to activate their critical thinking skills.
As a silent lurker I do normally never ever comment because I do not want to disturb the communication of the bots in progress here in any way, but over the last few weeks it has become beyond clear that there is an underlying pattern to these “aberrations” that, for whatever reason, nobody here has mentioned so far:

The aberrations observed right up until now, not being that sophisticated by their very nature, have so far never ever spilled over or taken place anywhere else than the squid thread page and they have also never ever occurred in conjunction with older topics and or posts and since they are quite limited in scope and topic they can hardly be called an attack on this blog.

The aberrations only happen right after the user with the handle “Winter” engages in what many objective and critical observers would identify as clear examples of censorship, smearing, labeling, defamation or denunciations and no sooner.
One could ask if said user and his ongoing malevolent, intentionally discriminatory ignorant behavior have become a burden to this blog by inciting and attracting such types of responses under which everyone here consequently has to suffer.

That pattern can be observed time and time again. IMHO it really seems as if the user “winter” is purposefully stifling the free flow of information by artificially pushing his own potentially nefarious agendas and very aggressively asserting his self proclaimed topic and information dominance on this blog which obviously does not go well with more educated and less dogmatic and indoctrinated individuals who are not that frantically obsessed with their ego and the lack of narcissistic cowardliness regarding their less fanatic beliefs and obsessions. It could very well be the case that this brutish and degenerated and completely unjustified patronizing treatment of others by said user “winter” is being perceived as deeply unjust by some people who won’t take it lightly and who will rightfully feel justified to retaliate in kind by derailing any pseudo discussions left.

This is a clear case of what goes around comes around.

It has become beyond obvious that the comment sections of this blog and especially the so called users here, have formed a self righteous, self obsessed echo chamber and the self-laudation prone participants are quite happy about that and mock and ridicule every newcomer who is not part of the cult around the old and established doyens of the deeply hierarchical personal cult.

Critical thinking skills, unbiased topic related openness, rationality, logical and analytical thinking skills and an adherence to the scientific method are clearly not basic virtues of this so called community of neurotypical, herd-psychology driven bunch of ignorant, malevolent tribesmen always engaged in group think and the oppression of others and the suppression of critical information in defense of their baseless beliefs.

In the darkness of the disgusting infamy and wretchedness of this inbred place of demagogy and hostile ad hominem attacks, at least I can fully understand why an open and critical mind (or even a bunch of them for Christ’s sake) would seek retribution for the heinous acts of treason committed against all civilized moral and sacred ethical values there have ever been by the bunch of fascist untermenschen-like useless eaters dwelling here like a hoard of scabious, rabid, pestilent, old, senile gangrene-ridden flea infested sewer rats.

These are just my two cents, have a nice day you infantile twaddlers, just continue twaddling as usual:)

Reader April 29, 2021 6:40 AM

Interesting presentation:

Chairman of the FMF Rule of Law Board of Advisors and a former judge of the Supreme Court of South Africa, Rex van Schalkwyk, delivers opening remarks at an FMF roundtable.

“You will own nothing, and you will be happy” – Rex van Schalkwyk opening remarks:

ResearcherZero April 29, 2021 8:55 AM

When you observe two active GRU operatives using a device, catch them with it and give an internal intelligence officer a demonstration, it’s a pretty good bet it’s real. When you get hit by the bloody thing it’s very real.

They have known it’s real for a very long time and exactly who is behind it.
The only thing more real is how utterly useless some of the people making decisions are, and how often they ignore very real dangers, along with repeated intelligence and security warnings.

The lack of follow up investigations and their extremely low quality becomes very unsurprising after a while. Your local police carry out better investigations. There is no wonder how it’s so easy for networks, and individuals to be compromised, people to become disillusioned. FFS, you can’t even look after your own people.

“When this officer came in and I knew his background and he explained in an extraordinarily detailed but more military style that I could understand, I was like this is actually for real,”

Defense officials who briefed lawmakers said it was possible Russia was behind the attacks, but they did not have enough information to say for sure.

They have more than enough information, that’s one thing you can be entirely sure about.

InductionMotor April 29, 2021 10:33 AM

Thank you soooooo much for the links provided Universal Scholar, they are truly a godsend.

Godspeed to you!!!!!

name.withheld.for.obvious.reasons April 29, 2021 2:09 PM

29 Apr 2021 – Reported 27 Apr, Melzer Interview
Author of a book to be released in Germany soon, Nils Melzer answers questions by a German publication, here is one of those questions:
Interviewer: Inquiry
“What made a Special Rapporteur on Torture work on the Assange case and write a book on it?”

Nils Melzer: Replies
“When Julian Assange was still at the embassy in December 2018, his legal team actually reached out to my office. I remember it was just before Christmas, I saw the message pop up on my screen and I swiped it away immediately. I had this intuitive reaction; what does that guy want? He’s a rapist, a narcissist, a hacker, this isn’t serious, so I just discarded it. I have around 15 requests per day, and I can do one, it’s very routine for me to decide quickly, but I remember those negative emotions I had, that I usually don’t have.”

“I realized that I had strong prejudices against Assange, even though it’s my profession as a Human Rights expert to be objective. I started investigation further, scratching the surface of this case. The deeper I got into the case, the more dirt and contradictions came to the light. I also knew that I could not rely on information the media and the press, because that’s precisely the source that had deceived me in the first place. To be objective, I had to go visit him in prison, and, to be sure, I took not one medical doctor, but even two medical doctors with me, who are independent from each other and are not employed by the UN.”

“We spent 4 hours with Julian Assange, I spoke to him for one hour, and the forensic expert had one hour for a physical examination, and the psychiatrist did a two-hour psychiatric examination. Each medical examination was done separately from the other, so they wouldn’t influence each other. All three of us at the end compared our conclusions and agreed that he showed signs that are typical of victims of psychological torture. I must admit I didn’t expect such a clear result.”

“I reported back to the involved governments by the end of May. I was convinced Julian Assange had been deliberately persecuted and kept in legal limbo in Sweden, in the US, in the UK and everywhere to put hum under pressure to make him crack. It was done very publicly, in order to make an example of him, to scare other investigative journalists. The message was: ‘If you expose our dirty secrets, this is what is going to happen to you, and no one can protect you. We can violate your rights every day the way we want and no one can do anything about it.'”

FA April 29, 2021 3:06 PM


The point with RT is “deterministic” longest latency.


My understanding from long, long ago, is that RT is incompatible with interrupts. An interrupt is unpredictable, and its effect on latency too is largely unpredictable. As *nix are build around interrupts, this makes true RT impossible on *nix, I was told.

That’s a bit too simple. First, without interrupts (i.e. events that are asynchronous w.r.t. the running process), RT is useless – there’s nothing to react to in deterministic time [1]. Second, in many cases interrupts are not unpredictable at all, even periodic.

A good example of that would be any hardware that produces and/or consumes samples at a constant rate, e.g. an audio or SDR interface. When configured for a sample rate R and block size B it will interrupt every B samples, or every B/R seconds. When that happens it means that B samples can be read and/or written. It doesn’t matter much when exactly that happens as long as it is done before the next interrupt. That is a hard RT requirement – if ever the DSP process is late the system has failed. I routinely run DSP systems with hardware that generates more than 1500 interrupts per second. Current Linux kernels have no problem with that on typical PC hardware or e.g. a Raspberry Pi 3 or 4.

If things work as expected will depend much more on how the application is written. First of all, it must be multithreaded, with the thread(s) doing the DSP work running at a RT priority, and all the others in normal ‘timesharing’ mode. All memory used by the DSP code must be ‘locked’ so it is never swapped out. Then the DSP thread must not call any system services that could block. That includes memory allocation, file and network I/O, interaction with a GUI, etc.

One consequence of this is that languages or toolboxes that support ‘automatic’ memory management, reference counted smart pointers, etc. can’t be used. All allocation for the DSP code must be done at initialisation time. And all interaction with non-RT parts of the application must be buffered using a system that never blocks on the DSP side.

Basically, it seems any system having predictable (short) latencies will have bad average throughput, and vice versa.

There is no direct conflict between latency and throughput. For many RT applications you must get both right.

[1] The only exception would be system running only a single task that can wait actively (by polling) for some external event.

Winter April 29, 2021 3:18 PM

“There is no direct conflict between latency and throughput. For many RT applications you must get both right.”

Thanks for the update.

SpaceLifeForm April 29, 2021 3:46 PM

@ FA, Winter

If you really need some kind of RealTime process, you better dedicate a device to that task that will not have to deal with an interrupt storm.

In other words, the device should only deal with interrupts that are dedicated to the task at hand, and never have to deal with extraneous interrupts, from, for example, a eth nic.

Clive Robinson April 29, 2021 5:58 PM

@ FA, SpaceLifeForm, Winter, ALL,

First of RT is a spectrum that at some point just like AI got split into two terms “Hard” and “Soft” and these mislead people when they start looking at RT into thinking the options are binary when they are infact a multidimensional spectrum.

So onto @FA’s point in response to @Winter,

That’s a bit too simple. First, without interrupts (i.e. events that are asynchronous w.r.t. the running process), RT is useless – there’s nothing to react to in deterministic time [1]. Second, in many cases interrupts are not unpredictable at all, even periodic.

The first point is wrong, “interupts” were designed way back in the past for sporadic but frequently rare events, such as input from a teletype at not much more than 50baud (either 6.6 or 5 char/sec depending on if the transmirtion code was 7.5 or 10 symbols per char).

The way that was chosen by some was to use a level sensitive polling process in a 4bit microcontroler which gave us the “Peripheral Interface Controler”(PIC) family that eventually ended up owned by MicroChip. In essence the PIC was used as a “stream buffer” like a clocked shift register but run in parallel thus off loading work from the system core CPU. Both IBM and Sun (via Cray) carried on with this idea into the Z and Starfire IO systems. Often the transfer was done not via interupt but by DMA and a “letter box” register and bidirectional flag.

Similar systems still have to be used in the more extream “Hard RT” systems often by state machines built in FPGA’s for all sorts of what are informally called “flight critical systems”. They can be quite complex one feature being “predictive responses” that is part of the response algorithm is either built into the FPGA or is downloaded in advance from the more general core CPU. It’s kind of like a “piecewise curve following” algorithm that can be changed in advance often from a table of response information.

An example to see how it works might be a breaking system that uses “bang-bang” control where the actuator is hard on or hard off which would be extreamly skid prone. But by varying the on/off ratios the mechanics “integrate” to a smooth but nonlinear response curve that is optimal for a given set of conditions such as velocity and direction so skidding does not happen. These days you would not do breaking that way because dedicated low cost microcontrolers are fast enough for direct control at the wheels. However there are many not to disimilar control loops used in other real world devices that do work that way because of the required “guarenteed response times”.

The second point is a classic mistake made by many people and the number of times I’ve had to verbally “smack people on the ear” over it is more than I would care to count.

Lets say you have two indipendent streams both using XTAL control to better than 1ppm that is divided down to give two data streams transfering 32bit words at a rate of 44kHz. Both are very predictable, but because they are independent of each other their delta F thus drift rate with respect to each other is not predictable unless you can acurately measure the data clock rate to some infinite precision. Why would this be? Well in effect the time difference between the two data clock occurances is the recipricol of their frequency difference. As they are indipendent they are unsynchronised and will thus drift with respect to each other. Over a period of time one will eventually get not just one data sample ahead of the other but eventually a whole buffer worth of data samples ahead of the other. Knowing this it’s not difficult to realise what will happen if things are incautiously programed. But… to correct for such an error means at some point a branch in the software behavuour, thus the software ceases to be “time predictable” which is where the “peak -v- average” time issue I previously mentioned comes into play. You can aproximately work it out if the unbranched time is T1 and the branched time is T2 then with N streams the minimum time is N.T1 and the maximum N.T2. The distribution curve is related to the Fourier product of the respective drift frequencies. That is if you observe the drift between two data stream clocks you will find it follows a sinewave at the difference frequency very precisely the number of these sinewaves goes up by 0.5(N^2 – N). Thus the response is the sum of the sinewaves with respect to their phase differences and though entirely predictable calculating it is a very very intensive process as well as quite complex thus not something you want to be doing in any Real Time system be it very hard or very soft. Oh an one side effect, the more precise/stable you make those XTAL oscillators the more you delay the occurence of the peak event. Thus it may never show up inn even rigourous testing, but it will in the field.

These are just the facts of life in the more extream ends of Hard RT.

I could go on, but the simple fact is none of the inbuilt patches FA describes will solve these sorts of Hard RT problems.

What will for audio systems as any longterm studio engineer knows is you get rid of all asynchronous inputs, that is you give them all the same master clock to work from and the XTAL’s in the source devices get “slaved” to that.

Whilst sync/slave clocking works for a lot of things, it does not work for all. Some systems have interesting problems caused by the effects of mobile sources via Doppler shift and even Relativity. Others caused by such effects as temprature variations on transmission lines and mechanical vibration of electronic components (look up microphonics).

At the end of the day the underlying functioning of all *nix systems is one of “sampling” and you can go and look up the effects of “sampling systems” and when and how they give problems. Knowing this you can then working out the peak delay time come up with limits on what a system can do so that it stays within the “safe operating conditions” area on the graph. Whilst adding an extra source or two may not appear to cause problems, if you opperate the system long enough, it will cause a problem.

One thing that people often forget to their eventual cost is “The problem with problems” is that thay “add time delays” thus they make the problem worse, and all to often these problems accumulate to the point something breaks. And a rough rule of thumb says “The longer you wait the harder the break”.

Any way I’ve gone on long enough for what is for most people a very obscure subject.

BarFlyBob April 30, 2021 2:07 AM

All that stuff everyone was saying about lithium batteries. It’s all true.

It smokes, it flames, and the smoke which contains hydrogen fluoride is corrosive to most electronics. The longer it sits the more it degrades what it touches. The soot can cause an insulating effect on components that appear unaffected and the cause thermal damage when re-connected to mains power or additional fires later.

My only advice is one is none and two is one. On-site back up great start. But you have to go the extra mile to something analogous to back blaze to hedge against major damage to your physical plant. I only say it because it’s what saved my bacon.

Afterwords, the fire inspector said there was no clear culprit device. One of the lithium power banks…(as Gogol says, “what happened next is obscured by a think fog”.) 3 months + out of of the flat. Virtually everything from spoons to sofa requires cleaning. All that from a fire that was only about 7 minutes if the final time stamps from my synology box are to be believed.

Cost replacement insurance if you can get it. Schedule anything crazy valuable. And lastly email receipts and scanned receipts have been life savers.

Thanks for letting me chime in on Lithium

FA April 30, 2021 3:29 AM

@Clive Robinson

The first point is wrong, “interupts” were designed way back in the past for sporadic but frequently rare events, such as input from a teletype at not much more than 50baud (either 6.6 or 5 char/sec depending on if the transmirtion code was 7.5 or 10 symbols per char).

Can be, but that doesn’t prove anything related to the first point.

… (irrelevant digression about PICs, DMA, whatnot)

Lets say you have two indipendent streams both using XTAL control to better than 1ppm that is divided down to give two data streams transfering 32bit words at a rate of 44kHz. Both are very predictable, but because they are independent of each other their delta F thus drift rate with respect to each other is not predictable unless you can acurately measure the data clock rate to some infinite precision.

So the two interrupts are asynchronous w.r.t. each other, and every now and then they will be very close together. That’s no problem. Let T be the minimum of the two interrupt periods, and T1, T2 the times needed to service each one, including scheduling latency. Each interrupt needs to be serviced before the next one (on the same stream) arrives. As long as T1 + T2 is less than T, there is no problem at all. Even if servicing one of the streams is interrupted by the other, both will still finish in time. And if the condition above is not satisfied, the system is overloaded anyway, and even the most perfect super-hard-RT system you could dream up would fail. Which makes your

I could go on, but the simple fact is none of the inbuilt patches FA describes will solve these sorts of Hard RT problems.

rather irrelevant. Whether or not such a system will fail or work does depends only on the condition stated above.

And before you say that I’m dreaming, exactly such things are constantly running on a number of Linux audio procesing system that have been in constant use for years. Some of them handle up to eight such near-synchronous streams at a time. [1].

What will for audio systems as any longterm studio engineer knows is you get rid of all asynchronous inputs, that is you give them all the same master clock to work from and the XTAL’s in the source devices get “slaved” to that.

Using a master clock is indeed standard practice, but it works only for local sources. Anything arriving from the outside will be asynchronous and need to be resampled. It’s done all the time in broadcasting.

[1] All but one are resampled to the ‘master’ sample rate. The resampling ratio is continuously adjusted (very slowly) by a feedback loop that keeps the average number of samples in a buffer constant, basically a DLL in software. On a modern PC the CPU load for resampling, even at very high quality, is trivial.

Winter April 30, 2021 3:39 AM

“As long as T1 + T2 is less than T, there is no problem at all.”

Indeed, but we know how this works. T>T1+T2 is “inefficient”, especially if it is T1…Tn, so the hardware will be specced to be T > max(T1,T2) which is more efficient and will fail eventually.

FA April 30, 2021 3:53 AM


If you really need some kind of RealTime process, you better dedicate a device to that task that will not have to deal with an interrupt storm.

Certainly true in case failure would have dire consequences, like an aircraft going down or a chemical plant exploding. But that is not always the case.

On the Boeing Dreamliner, some functions that used to be handled by dedicated processors before are now ‘applications’ running on a central computer (with redundancy of course). I always wondered if this is a good idea. But then Boeing have done some other things that turned out to be ill-advised…

FA April 30, 2021 3:57 AM


Indeed, but we know how this works. T>T1+T2 is “inefficient”, especially if it is T1…Tn, so the hardware will be specced to be T > max(T1,T2) which is more efficient and will fail eventually.

🙂 not just ‘eventually’, but sooner than one would expect…

That’s bad design by cutting corners. Sadly it happens.

Winter April 30, 2021 4:01 AM

“Rather, instead you should be asking yourself why doesn’t a disinformation outlet want me to take a vaccine for a virulent virus?”

There are two ways to win a sprinting match:
1) Train to become the best
2) Kneecap the other contenders

Now think about countries. Which countries want to become the top dog by beating the other economically and militarily, and which countries by kneecapping the other contenders.

And how would these countries respond to a pandemic.

JonKnowsNothing April 30, 2021 9:25 AM

@Clive, Winter, SpaceLifeForm, All

re: Russian Sputnik V COVID Vaccine profile

MSM report on why the Brazil Authorities have rejected using the Russian COVID-19 vaccine had an interesting section on the details of their concern.

It was not because the vaccine was ineffective, it had to do with the wrapper adenovirus used to carry the genetic payload that targets the spike protein.

  • The Sputnik V vaccine is a 2 shot series.
  • It uses 2 different adenovirus as vectors
  • Shot 1 uses adenovirus type 26 (Ad26) isolated from children
  • Shot 2 uses adenovirus type 5 (Ad5), which causes colds and has been used extensively in biomedical research.

The issue seems to be with Shot 2 AD 5.

Because adenovirus are common and humans may have been previously exposed to many of them, the vaccine has to have enough vectors to deliver their payload before the existing adenovirus antibodies destroy the shell. If the antibody response is strong, the genetic portion does not make it to the targeted cells.

To make sure enough of the vectors survive long enough to deliver their payload is a difficulty with the adenovirus family.

A portion of the shell-vector viruses used in vaccines may be able to replicate themselves. It can be both good and undesirable. The replication-competent adenovirus (RCA) in Shot 2 is where the Brazil’s Health Regulatory Agency, Anvisa found something that concerned them.

The number of RCA allowed differs from the Russian stated standard and the USA standards.

  • Sputnik V developers set the limit of 1×10^3 [1,000] RCA per dose of 1×10^11 [100 billion] viral particles,
  • The US Food and Drug Administration’s guidance for gene therapies states that the agency recommends “a maximum level of 1 RCA in 3×10^10 [30 billion] viral particles.”
  • USA 3.33 RCA in 100 billion
    Russian 1,000 RCA in 100 billion.

It is not clear what the real problem is behind the rejection. It maybe undesirable to give people a vaccine for COVID-19 that also can give them the common cold. Given Ad5 is used for biomedical research, people may have already had Ad5 exposure so would mount an antibody response to it. That antibody response might impair or complicate the desired response to the COVID-19 genetic payload.

It is also undesirable to die from COVID-19.


ht tps://

(url fractured to prevent autorun)

SpaceLifeForm April 30, 2021 5:49 PM

@ JonKnowsNothing

Interesting. I wonder if the jabs were reversed that it would be more effective.

I still have not had a jab. I wanted J&J but some bad actors threw a spanner in the works.

I also am wondering…

Are the bad results (ex blood clots), a side effect due to someone having been previously infected?

But they did not know because they were asymptomatic?

Should one get an antibody test first?

I have learned of a med person that tested positive, got antibody test, then got vaccinated, and antibody levels did go up.

Lucky? Or just what should be expected?

This virus is very tricky.

JonKnowsNothing April 30, 2021 8:46 PM


re J&J jabs

It took jumping through hoops but I am scheduled to get a J&J jab Monday.

It’s all rather confusing how things are rolling out here: The County does most of the jabs, the HMO and Hospitals do more jabs, you can get jabs at some rolling clinics, you can get jabs at stores, grocery markets that have pharmacies and Big Box stores that have pharmacies.

What is hard to find out is who has which jabs. Some sites have all 3 types and some 2 and others just 1 type. It shifts day by day, maybe hour by hour.

My HMO has the 2 mRNAs and no JJ. I called the County COVID-19 Vaccination number to see if they knew who had JJ jabs.(1)

I am scheduled with a new County Vaccination Program for Home Bound persons and the County will send someone to punch me at home or rather on the sidewalk in front.

Give your local county a call and see if they can direct you the right jab site.

note: Providing I don’t drop dead right off, I’ll post updates on the aftermath.


1, Currently in California the US Government pays for the vaccination costs and residents in a County are generally eligible for COVID-19 County services. If this changes in the future, going outside of the HMO would be too costly as it would be treated as Uninsured Health Care and billed accordingly.

HM0 = health maintenance organization (HMO) is a medical insurance group that provides health services for a fixed annual fee.

ht tps://

(url fractured to prevent autorun)

Rachel April 30, 2021 10:38 PM

Clive Robinson, Nik

On the post last week about morse code. Thanks for your feedback.
Clive you focused on the need to learn before the age of 5. As we can safely guess you are highly competent in the use of Morse , if you learnt Morse Code after age 5 then you contradict your line of reasoning.

This point is important because of the methodology on learning languages per the book ‘Fluent Forever’ by Gabriel Wyner. Easily available across languages in mainstream bookstores

The method applies equally to programming, language, morse, or something else.

Entirely supported by up to date research on neurology, the text is respected internationally. One thing it seeks to do is burst through popular myths about language acquisition.

A big one is ‘need to learn as a child or the opportunity is lost forever’. As well explained by Gabriel Wyner, again fully substantiated by referenced research : adults are mcuh better at learning languages than children. Their brains are fully developed. The difference? Conventional methodology is terrible. It’s the method that is the problem.

The other reason I want to comment about this is because the book points to using SRS – Space Repetition Systems. Based on the idea

‘ you remember something best if you learn it just before you are about to forget it’.

the brain retains things for a short time, but repeat exposure at the correct interval convinces brain to deposit data into long term memory.

The FOSS software Anki can be used for creating flashcards, containg text, audio, images, video, implemented via SRS, for effectively learning any subject.

So Nik perhaps you can give Morse Code another opportunity

Winter May 1, 2021 4:34 AM

“As long as T1 + T2 is less than T, there is no problem at all.”

Indeed, but we know how this works. T>T1+T2 is “inefficient”, especially if it is T1…Tn, so the hardware will be specced to be T > max(T1,T2) which is more efficient and will fail eventually.

Winter May 1, 2021 4:36 AM

“Rather, instead you should be asking yourself why doesn’t a disinformation outlet want me to take a vaccine for a virulent virus?”

There are two ways to win a sprinting match:
1) Train to become the best
2) Kneecap the other contenders

Now think about countries. Which countries want to become the top dog by beating the other economically and militarily, and which countries by kneecapping the other contenders.

And how would these countries respond to a pandemic.

Clive Robinson May 1, 2021 6:57 AM

@ Rachel,

Clive you focused on the need to learn before the age of 5. As we can safely guess you are highly competent in the use of Morse , if you learnt Morse Code after age 5 then you contradict your line of reasoning.

I indicated that certain parts of the process which is the auditary sensing was best learnt as early as possible. It’s what is assumed[1] to be the reason behind why some people can never learn certain parts of a language as they do not have the ability with certain phonemes (the C and R in my name) yet have many many times the number of “pitch perfect” people in their population.

The ability to learn morse easily was found during WWII to be quite age related and they found those under 18 learned faster and better than those over 18.

With regards my ability with morse no I’m not profitient in any way any longer I just don’t use it and have not for decades.

I started learning morse code back in the early 1980’s not to use it, but to get the HF provison of the license. So I was never going to put the hours in any way. But as often happens about when I was sufficiently profficient various changes happened and I did not take the test (it’s why I’m not a G4). Also my reason to get the licence was for “experimenting” not “contesting” or “rag chewing”. I did almost all my operating over the years calling “testing testing testing” not “CQ CQ CQ”. But could send my call sign in morse when required.

There is a saying of “If you learn it well you remember it well”…

Let’s just say I never got into the rhythm and certainly never got to the point of hearing the sound of words in morse. It takes a lot of practice and their is a brick wall effect around 15WPM where you can nolonger grab a letter at a time but have to grab the words (if you can call them that, most morse is so full of codes it might as well be random letters to anyone who has not learnt them).

My intetest in morse kind of stopped when I’d finished developing a program for my Apple ][ to not just send but importantly for the time receive it. I then went on unsuccessfully trying to write a program to recognise phonemes thus speech (technically known as a discrete utterance speech recognition engine[2] it was not a success due to lack of computing power).

So my knowledge on what does and does not happen in humans and when is based on researching others research when trying to get a computer to replace the human.

Which brings us to,

A big one is ‘need to learn as a child or the opportunity is lost forever’.

Is true, when you actually know what it is that is being learnt. Learning “new skills” rather than “new things” has over many years been shown to have an age related component. Learning a new latin language is not learning a new skill, but learning a new thing for which you have already aquired the skills.

There are languages that have a “tonal or pitch” content that if you are “tone deaf” you are never ever going to learn. As I’ve mentioned oriental languages have a tonal component, with some to Western ears sound realy unpleasent as we tend to regard the tone as an “emmotional side channel” with a deep voice indicating being in control and with a tonal flatness as being not emotionaly perturbed. Whilst high frequency as panicky and animated tonal content indicating very high emotion such as anger or being scared or both.

But if you speak one of most Germanic/latin languaged you already have the skill set, you just need to learn new things which again you are often more than half way there with. It’s known for instance that those who speak German usually pick up other languages fairly easily because it requires a broader range of skills than speaking english.

But you see a problem with skill sets and the switch from gutteral languages from Northern Europe to more sibilitic languages such as the Romance languages.

Sibilants are made by the use of the tongue and forced air preasure and do not require the use of the vocal cords or throat (try saying “shhhh”). They sound notably louder than their non-sibilant guteral or frictive counterparts due to most of their acoustic energy occuring at higher frequencies. Importantly though skill with sibilants requires fine motor control of the front of the mouth thus face muscle control (as does tunefull whisteling). It’s why people who speak romance languages appear to have younger thus more attractive faces because their cheek bones look higher due to the improved musculature that gutteral languages do not need.

Thus learning a language with a high content of sibilants from a language that does not, requires both the learning of new skills and the learning of new things.

Which brings us to the difference between “skill” and “knowledge” many see them as being compleatly different that is it is almost as “physical” skill or it is “mental” knowledge. Which kind of misses the point, after a few moments thought most sgould realise that it is a spectrum, or more correctly a hierarchy of levels which is why I have used “things” not “knowledge” upto this point because most of what you actually have to learn when learning a new language is those bits in between and you have to build them up just like you do your face muscles.

The problem is we have no real idea how people learn. Take a fine motor skill such as pen manship it is assumed that you “pick it up with time and use” actually more recent findings suggest that’s about the worst way to do it. One interesting way to do it which has been shown to work is “thoughtfulness” that is you do not “do” but “think” your way through the action. You see this best with certain types of marshal art, where you go through the movment very slowley thinking where each part of your body should be. This slow highly thoughtfull process builds the skill into the lower levels of the hierarchy thus when you make the movment at full speed it is precise, efficient, and controled. There is some research going on into teaching people to learn to drive this way but I’ve yet to see the results.

Which leads into,

‘you remember something best if you learn it just before you are about to forget it’.

Whilst that may be true for “knowledge” how about “skills” or the “things” inbetween?

It kind of flies in the face of the old saying,

“Once you learn to ride a bike you never forget”.

It’s safest to say we know way less than we think about learning. Even the old “people think in three ways”,

1, Those who use words.
2, Those who see graphs/pictures.
3, Those who see formulars/equations.

With the obvious implication the order of efficiency is using words the least and equations the most. I try to do all three but I still have troubles going from the symbols of equations to the meaning without going through the picturing of graphs and flows.

But if you want to try it, imagine that learning an individal fascit is an effort process that is more or less linear. However your ability as your skills build is fairly obviously a percentage growth or more accurately an exponential one, the limit is your ability to form the chemicals and neurons to retain what you learn. We have reason to believe that neuron development has an inverse relationship with age, and what nasty chemicalls we poison our bodies with. Sadly good nutrition and clean living only takes you up to your supposed inate limits.

Anyway back to morse and my non use, my health being somewhat on the decrepit side, and having had a mis spent youth wearing the green my hearing is realy quite bad these days[3]. And whilst I can sort of lip read, COVID and masks has made me appear deafer than Beethoven… Very much to others anoyance my hearing appears to come and go, but speaking on any kind of phone is both difficult and embarrasing even with people I know. So I’m “mic shy” with radio as well, even though headphones and parametric equalisers on each channel do help a lot. Thus my interests are more in the non audio data modes of ameture radio these days. But whilst I do “listen in” to morse from time to time it’s more “watch in” as even idi-bidie $1 microcontrolers have the power to decode morse code if you filter it correctly and that is something I do know how to do with both analog and digital filtering.

Thus you might think I was a natural fit for social media but trust me on many levels when I say I’m not 😉

Anyway now I’m in “me time” I’ve a minor project to shuffle along[4].

[1] It’s assumed, as experiments to prove it would be quite unethical. However there were experimrnts done with bringing kittens up in a world of only vertical stripes then after a period of time putting them in a world of only horizontal stripes and observing the differences (and another group the other way around). However experiments with adult humans where they wear glasses that turns the world upside down shows it takes about 3-7days to adapt and slightly less to unadapt. Other tests have shown that removing one sense such as sight has an enhanced effect on other senses souch as hearing and touch.

[2] Whilst I could get the Apple ][ to do waveform matching it was way to much load on, 1MHz 8bit chip to do much more than that. Whilst I did start devekoping external hardware, it was not going to be a practical thus marketable product. It was around a decade and a bit later at the end of the 1990’s with 50MHz and faster 486’s being common on peoples desk tops that others finally got things working enough to be commercial products. But by then my interests had moved on to doing interesting things with video and encryption, which again suffered from lack of hardware grunt. My interests moved on and I started looking into making a properly secure computing system, you can read about bits of it here on this blog if you search for Castles v Prisons and C-v-P or as @Wael ended up calling it CvP. And yes it has a hardware issue as well though as pointed out to me by @Thoth after we had discused on and of using smart cards, somebody from Cambridge Computer labs started marketing a design based on what had been said on this blog… I’ve no idea if they made a commercial success of it or not.

[3] It’s actually difficult to explain the effects and probably why some maliciously call tinitus “selective hearing” as though sufferes do it deliberately. In my case it’s,probably due in part to the use of weapons and similar and I’m also told probably related to several boughts of labyrinthitis (which is what is believed caused Beethoven to go deaf). But trust me when I say which ever way does not matter, you realy realy don’t want the whistles, buzzing, squeaks and gurgles 24×7. For instance I can talk to people on a reasonably busy street, but go into a Mucky D’s and the noise inside just triggers it and placing an order is very very difficult if not imposible, which is why the cheap vouchers came in handy you just handed the cashier the voucher and the money…

[4] It’s a bit of fun, as you might know “stack based” OS’s and languages have a lot of advantages over more traditional ones in such things as you do not need MMU’s so high end microcontrolers can multitask without problems. However in general stack based systems are “type limited” that is values are put on the stack and that puts a limit on the size of unsigned integer to be that of the CPU data bus (usually but not always the maximum size of pointers as well). Which means all other types have to be compound which is messy on stacks or less than the range of the pointer or long unsigned int. So what I’m doing is building a stack based system that uses pointers for everything on the stack, these point to structures that then hold type information and methods. So far it’s working OK with 1kByte int’s and floats and complex numbers. I need to get matrix types working and then get onto the more interesting “crypto types”.

- May 4, 2021 1:29 PM

@ Bruce Schneier,

On going through the mess that is the first squid in May you will find,

1, #comment-376494

Where a certain person makes their intent quite clear, as others have indicated, the attack is against you, your reputation, and your blog.

Whilst they claim it’s about censorship against them, it is infact the opposite, they do not like their fabrications reviled for what they are. But underneath that there is something else, which is a desire to damage you personally through your blog.

I would be carefull what response you make as I would assume that there is more behind the attacks than “ego food” for a disgruntled individual.

Leave a comment


Allowed HTML <a href="URL"> • <em> <cite> <i> • <strong> <b> • <sub> <sup> • <ul> <ol> <li> • <blockquote> <pre> Markdown Extra syntax via

Sidebar photo of Bruce Schneier by Joe MacInnis.