Rolandt

In einer Woche ist Weihnachten und wer noch schnell ein Last-Minute-Geschenk für einen Nerd braucht, findet hier bestimmt etwas. Alle Produkte benutze ich selbst!

• 65W-Ladegerät für alles, was einen USB-C Anschluss zum Laden hat. Wichtig:: Nur in GaN-Technik kaufen. Klein und effizient.
• Ladekabel mit Anzeige dazu. Achtung, das kann nur USB 2 mit 480 MBit, taugt also wenig für die Datenübertragung.
• Schnelles Datenkabel, bis zu 80x schneller als das Ladekabel. Beide Kabel sind mit Nylon umsponnen und haltbarer als die Standardware.
• Für iPhone 12 und neuer, zwei MagSafe-Accessories: Anker 622 Kabellose Powerbank oder das MOFT Wallet. Beide halten magnetisch an der Rückseite des iPhones und sind gleichzeitig ein Aufsteller.
• Die weltbesten Winterhandschuhe.
• Ein Wera Toolcheck Plus für den Bastelwastel. Gutes Werkzeug geht immer!
• Ein gutes Taschenmesser braucht jeder, am besten mit Schere: Fein und vielfältig von Victorinox oder grob und stabil von Leatherman.
• Jabra Elite 3, die billigsten zu empfehlenden Earbuds, und die Elite 5, die besten von Jabra.
• Eine Maus, die bei Schulter und Ellenbogen-Problemen Abhilfe schafft: Logitech Lift für kleine und mittlere Hände und Logitech MX Vertical für eher lange Finger. Lift gibt es auch für Linkshänder. Erfahrung: Nach einer Woche war der Tennisarm beschwerdefrei.
• Poly Sync 20 für alle, die bei Telcos keinen Kopfhörer tragen wollen.

I’d like the next best-seller to be “The Compassionate Programmer: Programming as if People Mattered” because I think being pragmatic hasn’t played out so well.

– me on Mastodon earlier today

The idea for this book first came to me while I was supervising undergrads’ senior projects in computer science at the University of Toronto. I routinely pointed students at The Pragmatic Programmer, but the more often I did, the less satisfied I was:

1. It had a lot to say about software design, but like most books on the subject, it made very few references to specific systems. (Beautiful Code and The Architecture of Open Source Applications were inspired by this frustration.)

2. Few of its pronouncements were backed up by empirical evidence. After reading Glass’s Facts and Fallacies of Software Engineering, I realized that (a) we actually knew a lot about how programs and programming actually work and (b) many of the things I believed were either wrong or unproven. (Making Software and It Will Never Work in Theory were attempts to address this.)

3. It left out things that mattered to me as a working programmer. Well into my thirties I chose not to notice how our industry excludes people who aren’t affluent, white or Asian, straight, physically able, and male. Once I acknowledged that, it became impossible to read anything about teamwork or management in tech without asking what it said about fairness or employees’ rights.

4. While it had a lot to say about software design, it left out security, privacy, accessibility, and how to not amplify hate online. These can’t be sprinkled onto a program after Version 1 ships, so I think they need to be taught just as early as design patterns or the SOLID principles.

The twentieth anniversary edition of PP didn’t fix these shortcomings, but neither have other books of its type. For example, I have eight recent books on how to be a software engineering manager on my tablet right now; only two of them mention tech’s history of discrimination and exclusion, and none of them talk about employment rights or what to do if your boss asks you to add an option to a wayfinding application that will avoid police checkpoints. (Which user persona did you think of first: a drunk driver in Manitoba or someone on their way home from a political protest in Tunisia?)

Hence, The Compassionate Programmer (or possibly Building Tech Together—I dither over titles). What would we teach young programmers about how to be better at their craft if our starting point was the belief that people matter? Extensible, testable designs? Absolutely: tidying up the kitchen so that the next person who wants to cook doesn’t have to work around your dirty dishes is a sign of a compassionate roommate. Version control and issue triage? Ditto. Debugging strategies? Absolutely—showing a young programmer how to use a breakpointing debugger effectively is one of the kindest things you can do for them.

But the title says “programmer” not “programming”, so I’d include just as much on intellectual property (because a lot of companies say or do things that aren’t actually legally enforceable), how to run a meeting so that everyone has a fair chance to be heard (slides, video), danger signs in interviews, and what to do if you’re unfairly fired. I’d explain why “radical candor” is bullshit so that people wouldn’t blame themselves for toxic management. And I’d reference every relevant empirical study I could find because we know a lot more than most people realize, and because the more working programmers know about research, the more likely they are to ask researchers to start tackling problems that practitioners actually care about.

There’s no way all of that could fit in a single book, but every book leaves out more than it includes, which means every book embodies its author’s priorities. When we say, “If you’re going to read one thing about programming, read this,” and the book talks about the Visitor pattern but not about algorithmic bias, it’s telling people the former is more important than the latter, and so are we.

I don’t want to do that any more because I no longer believe it’s possible to build good things well unless we think about how they’re going to be used and abused. I’ve tried to write this book twice and given up both times. I think that doing it properly would require a year of full-time work; so far I haven’t found a company or foundation willing to support that, but even if I could, I think it’s time for fresh eyes on the problem and fresh voices on the airwaves. If you ever take this on, please let me know.

mkalus shared this story from Myles Power.

When I started my foray into Holocaust Denial 5-years ago, I began by stating that no event of any significance in the world takes place without generating a flutter of conspiracy speculations. As unpleasant and potentially dangerous a belief in one or more of these conspiracies can be, they do, for the most part, come from a natural human desire to find an explanation for what appears to be unexplainable. However some historical events are so well documented that a belief in certain conspiracies comes from a much darker place where people feel compelled to bend and warp reality to justify their dislike or even hatred towards a certain group of people – and there is no better example of this than those who distort or flat-out deny the facts of the holocaust. 

Over the years I have countered what I consider to be the strongest arguments made by Holocaust denialists with reactive ease, but have not yet refuted Germar Rudolf’s book, The Chemistry of Auschwitz, leading some online to believe that I am unable to do so and therefore, somehow, the Nazis did not systematically kill millions of people in their death camps. The truth however is that I simply don’t like being told what to do and rather childishly, have been watching with great joy as revisionists claim victory online. Now that things have died down a little, I have decided it time to tackle this rather bland and unconvincing book and bring an end to this project.

The Chemistry of Auschwitz

The Chemistry of Auschwitz is an excruciatingly difficult read due to the multiple revisions made by its author. These revisions in this revisionist book are marked by jarring tonal changes in writing styles indicating that Rudolf no longer believes in many of the arguments he made when the book was first published. However, rather than removing these arguments from the book in their entirety, Rudolf seems to have simply edited out his endorsement leaving them for the most part intact. Perhaps he was worried that his girthy magnum opus would be reduced to a mere pamphlet if he were to remove all the mistakes he would openly admit to, or maybe he simply wanted to dog-whistle to his intended audience. Regardless, these edits resulted in a confusing and spineless book which is an echo of its former self. For example in an early addition, Rudolf regurgitates an argument from a fellow revisionist and former roommate, Gerald Fredrick Töben, who stated that one of the doors found in Crematorium I was not gas tight and therefore the room where it was situated could not have functioned as a gas chamber. Rudolf states that “This door is neither of sturdy construction, nor is it air-tight (note the keyhole). It is partly glazed and opens inwards i.e., into the room, where corpses were allegedly piling up”. However soon after the publication Rudolf discovered that the flimsy wooden door was not the same door used when the building was retrofitted into a gas chamber. Not wanting to remove this easily debunked keystone of the revisionist movement, Rudolf frames the argument as if Leuchter is the one who made it when he visited the site in the 1980s. 

The book also suffers from an identity crisis as Rudolf struggles to pin down who his target audience actually is. There are sections written in a dry scientific manner, which I imagine the average revisionist would struggle with, punctuated by sections written in the style of a teenager’s MySpace page in the early 2000s. For example, why in a book that supposedly disproves the holocaust would its author think it was a good idea to randomly start talking about his birthday? It feels like it has been written by someone who lacks a filter and who believes it would be a crime not to publish every thought that tumbled through their head, giving the book a bloated feel. When combined with the revisions it becomes as I said, excruciatingly difficult to read. 

If you were to distil down Rudolf’s book, his reasoning and methods for disbelieving in the industrial mass murder of people the Nazis deemed undesirables is identical to that found the Leuchter report. Rudolf illegally obtained samples from Auschwitz and covertly had them tested for the presence of a specific family of hydrogen ferrocyanides, commonly known as Prussion blues. Unable to find their presence in the samples he stole, he concluded that no gassing took place. Like with Leuchter, Rudolf’s reasoning and conclusion are wrong because Prussian blues would not necessarily be formed in the conditions found in homicidal gas chambers. This is due to the fact that their synthesis is sensitive to concentration, temperature, the amount of carbon dioxide present (from humans exhaling), presence of water, and the presence of Fe (III) that is already complexed with cyanide. In order to prove his thesis, Rudolf needed to demonstrate that these compounds can form in the conditions found in the homicidal gas chambers – something which he was unable to do, rendering his conclusion bassless. Rudolf, like Leuchter also erroneously believes Prussian blues to be the major form of cyanide residue discounting all other compounds. 

Additional fundamental flaws have been pointed out by many others. Gilles Karmasyan pointed out that Rudolf based the quantities of hydrogen cyanide released on a single figure which he deceitfully used to base an evaporation rate curve from before concluding that the rate of release of the deadly gas is too slow to kill anyone in a few minutes. To come to his conclusion, he ignores the importance of temperature and the simple and dark fact that human body temperature is way above the boiling point of hydrogen cyanide. Others have taken issue with Rudolf’s baseless claim that no forensic analysis was conducted on facilities at Auschwitz-Birkenau after the camp’s liberation. In reality toxicological analysis was carried out as early as 1945 by the Krakow Forensic Institute. They collected samples from, among other places, the ruins of Krematorium II and discovered the presence of cyanide compounds. Traces of hydrogen cyanide were also found in metal objects found in the hair or victims, such as pins, clasps, and gold-plated glasses holders. Even Rudolf claims about the lack of holes in the roof of the gas chambers has been thoroughly debunked. 

If, like me, you are able to see through the scientific jargon, The Chemistry of Auschwitz is a carbon copy of the Leuchter report, only written by someone who is competent in their field. At their cores, both share the same fundamental flaw regarding the synthesis and analysis of cyanide based residues rendering them both invalid. This is further proof that the people who promoted this book to me either have not read it or they don’t understand what is actually being discussed. Regardless, the book was not the great challenge that I or the likes of the TDS postcast hoped it to be and I have found myself struggling to find things to talk about that have not already been discussed. 

When I previously brought attention to Rudolf’s lewd convictions it was not just because it is hilarious, but because it is a perfect example of how he and his fellow revisionists construct their arguments. They like to get bogged down in the minutiae whilst purposely ignoring the bigger picture. What I am basically saying is that cyanide-based residue is not the only evidence we have of homicidal gas chambers, and by allowing ourselves to get bogged down on the insignificant, we miss the bigger picture. We have warehouses of personal effects that once belonged to the victims, pictures from the camps when they were functional, and witness testimonies not only from people who survived the camps but from those who worked there. 

Oskar Gröning was a German SS Unterscharführer whose responsibilities included counting and sorting the money taken from prisoners at the Auschwitz concentration camp. After the war, he returned to Germany where he led a normal life, reluctant to talk about his time at the death camp for more than 40 years later until learning about Holocaust denial. He obtained a pamphlet by the Holocaust denier Thies Christophersen which he then mailed back to Christophersen having written his own commentary on it condemning Holocaust denial which included the following. 

“I saw everything. The gas chambers, the cremations, the selection process. One and a half million Jews were murdered in Auschwitz. I was there.” – Oskar Gröning

Gröning’s actions drew the attention of a neo-Nazi magazine and he was soon receiving anonymous calls from people trying to convince him what he saw with his own eyes was wrong.  

“People who tried to prove that what I had seen with my own eyes, what I had experienced in Auschwitz was a big, big mistake, a big hallucination on my part because it hadn’t happened” – Oskar Gröning

Following his interactions with Holocaust deniers, Gröning decided to speak openly about his experiences as an SS soldier working at the infamous death camp and publicly denounce those who maintain the events he witnessed never happened. 

“I would like you to believe me. I saw the gas chambers. I saw the crematoria. I saw the open fires. I would like you to believe that these atrocities happened, because I was there.” – Oskar Gröning

His self-incriminating statements exposed his life to public scrutiny; in particular his confession to stealing jewelry and money from gas chamber victims for his personal benefit. As Gröning did not physically kill anyone one himself he did not consider himself guilty of any crime, but felt immense guilt after hearing the screams. However after the legal precedent set in 2011 by the conviction of the former Sobibor extermination camp guard John Demjanjuk in 2015, Gröning was found guilty of being an accessory to the murder of at least 300,000 people. 

Gröning’s testimony is a drop in the ocean of evidence demonstrably proving the mass extermination of the people the Nazis considered to be untermensch. The evidence is in fact so robust and plentiful that those who don’t believe the holocaust happened are choosing to do so! There is no doubt in my mind that Rudolf intentionally set out to disprove the use of Zyklon-B when he traveled to Poland after being commissioned to do so by Hajo Hermann on behalf of Otto Ernst Remer. Unfortunately for Rudolf, scientific discoveries do not happen in a vacuum and if you purposely ignore the mountain of contradictory evidence to reach your distasteful conclusion as he did you are not doing science – you are attempting to excuse away your bigotry. 

When I published my first blog post on Holocaust denial 5-years ago, I had no idea how big of a project it would turn into. This was due in part to revisionists moving the goalposts at every opportunity, but now I feel it’s time to bring this project to a close. Nothing I ever publish will be good enough for them because, as I said they believe what they want to believe because this was never about historical accuracy or science.

I write this on December 16, 2022, as the landscape shifts convulsively, as Twitter becomes the main character on Twitter. It’s reasonable to wonder what comes next.

Anybody who says they can predict what’s going to happen is wrong. But we can peer down one or two of the infinitely-forking paths into the future and maybe see something useful.

However many the paths, there are only two interesting outcomes: First, and most likely, Twitter no longer matters. Second, and not impossible, it finds a route back to relevance.

Is my investigation pursued in an even-handed spirit? Not in the slightest. While Twitter has been a remarkable and in my experience mostly good part of the big picture, I’m hoping for it to fade away and open a better path forward, based on standards-driven federation, open to everyone, and at least partly free of the clutches of Big Tech.

And that’s what I really want to write about; I have some of that down, but it feels like the future-of-Twitter issue needs discussing first. So here we go.

Scenario: The walking dead

This supposes that things go on about as now, with Elon doing crazedly offensive things; John Gruber captures the flavor well.

The heads of state and policymakers and CEOs and Nobel-prize winners melt away, alt-rightists get progressive voices silenced, and what journalists remain are only there to cover the Twitter-disaster story. There’s a tide of litigation from justifiably annoyed European regulators, and a background roar of bankers demanding interest payments on the $13B Elon convinced them to lend. Interesting individuals leave the stream, trolls mass at the margins…

It could go on for a long time. Because the Eurocrats are slow, and the banks really don’t want to write those loans off. But Twitter would become a shambling travesty, looking for brains to eat.

Ejecting Elon

Let’s look at the paths that lead to good outcomes, which all rely on the assumption that Elon Musk gets out of the way. I can see two different ways that could happen.

Maybe one day pretty soon Tesla’s Board of Directors sends him a nastygram saying “Come back to work or you’re fired.” Elon realizes he’s not having fun, finds someone plausible to run the Twitter shop, and mostly leaves them alone while he goes back to his other jobs.

Alternatively, he doesn’t, but the banks run out of patience. This would be a little surprising, because they’d like to postpone recognizing their losses as long as possible. But, maybe not. If zombie irrelevancy looms large enough, they might foreclose.

They have options. One popular way to deal with a distressed debtor is to find a buyer, who’ll snap up the company for a derisory price but (crucially) assume the debt, and give the banks a chance of getting more of their money back.

Who might pick up Twitter on the cheap? Bezos, maybe via the Washington Post? Meta, in yet another effort to compensate for Facebook’s declining appeal? One of the big progressive foundations, seeing a need for a somewhat-sane public square? Larry Ellison, looking for another pool to piss in? A PE firm with more imagination than most? Some outfit out of Cyprus with opaque ownership who want Twitter to be nicer to Russia?

This path implies wiping out Elon’s equity and he’d probably fight it down to the last ditch, which could lead to the lenders forcing bankruptcy, which presumably lets them kick Elon to the curb and sell it to whoever.

And after that…

One way or another, after Elon’s gone, let’s suppose his replacement is sane and somewhat competent (otherwise we’re on the zombie road). This person is relentlessly focused on advertising and sales, so they deputize someone else to rebuild Engineering; maybe do-able in today’s tech labor market.

Thus the new boss can pour 100% of their energy into figuring out what it would take to lure the advertisers back and puts that right at the top of everyone’s priority list.

Who knows, it might work. Maybe Twitter’s Community Notes feature will be a breath of fresh air, because Truth, after all, is powerful medicine.

With luck, and good execution, Twitter might find its way back, might become again the place where the conversations that matter happen.

My bet

They’re not gonna make it. Big tech success relies on being able to recruit talented employees, and the landscape is always competitive for those people. In how many of these scenarios will people want to work at Twitter? I just can’t see it.

It makes me partly unhappy, because Twitter has been one of history’s lucky accidents. It’s improved my life.

Still, I believe that there’s a decent chance we can build something better, based on the Fediverse. Not a sure thing at all, but let’s try.

It somehow seems appropriate to begin the usual round of “year in review” posts with some of the stuff I enjoyed during 2022.

This was, I think, the year I watched the most video, ever–something I will definitely try to correct next year, since it is too big of a rabbit hole and has gotten in the way of my actually accomplishing stuff.

I blame it partially on my work schedule, which has shifted towards late mornings and evenings, in turn leaving less energy (and time) to pursue other things but passive watching.

But let’s do this again, this time before the year is over…

Severance still takes the crown in terms of sheer impact. I wrote about it earlier and still stand by everything there, since this is just one of those shows that makes an imprint–to the point I had flashbacks when this season of Mythic Quest (which is still quite fun, by the way) depicted brand new white walled office corridors.

After that, The Peripheral was likely the series I enjoyed the most. Despite plenty of criticism, I think I liked it because it took William Gibson’s concept and just ran with it off into a tangent–or, if you prefer sticking to its canon, a “stub”, in fact. Having read the books it was surprising, sure, but definitely entertaining.

It is also very much a production of our age: dystopian future, female hero, anti-hero, villain and ally, a visually pervasive sense of timeless futurism, and, of course, the shadow of The Jackpot, which I strongly believe has already started in our timeline…

Next up, Andor was a close third. Like many people, I think Rogue One was one of the best Star Wars movies ever, and Andor plays to that feeling of a lived in, gritty universe that goes a step beyond what The Mandalorian aimed for before it devolved into futility with Boba Fett and Kenobi, neither of which I remember finishing watching.

And, of course, I quite enjoyed Stranger Things – it has remarkable staying power, and this season had a pretty entertaining plot.

Other than these, I’ve fallen into the habit of having TV evenings with the kids where we watch some of my (occasionally ancient) DVD archives. They’ve quite enjoyed things like Allo Allo, Samurai Jack and, amazingly enough, Yes Minister¹, among other stuff.

I still have a sizable backlog of movies from last year, since going to the theatre is now very much a thing of the “before times”. So, before you ask, I didn’t watch Top Gun. Heck, I haven’t even watched The Matrix Ressurections yet…

But we did watch a few things here and there. Lightyear easily takes the cake here, because it was so crammed with nice little Easter Eggs (and a major plot twist) that it made its little side trip into Pixar’s increasingly grown up world so much fun, and seems well worthy of becoming a new franchise. I can’t wait for Zurg’s comeback.

Everything Everywhere All at Once came out of left field and was pretty amazing, but somehow I enjoyed Three Thousand Years of Longing a bit more–it’s flawed, but original in its own way.

As to the usual superhero fare Dr. Strange was infinitely better than the inanity of Thor. I honestly don’t get why they let Taika Waititi get away with this (not that I expected the movie to be… serious in any way, but it was just too dumb to be fun). I have Black Adam queued up on HBO to watch over the holidays and I completely expect it to top either, which given its ratings is quite telling.

Yes, well, it’s become a habit too, despite my stated intent of weaning myself out of it back in January.

Like many geeks on the planet I watch LTT and a variety of PC hardware/gaming channels, but the ones I actively seek out these days are mostly related to 3D printing and other Maker activities:

• In the Maker space, Zack Freedman is the kind of quirky, quintessentially geeky hardware hacker that pulls the kind of oddball creative stunts I love (and the creator of Gridfinity, which I use extensively).
• Ivan Miranda is probably the next most entertaining, due to the sheer scale of his projects and his enthusiasm.
• Thomas Sanladerer and CNC Kitchen rise above the dozens of 3D printer channels by going into the techniques and rationale at length (I also follow lots of 3D printer channels that focus almost exclusively on reviews and modding, but I’ll spare you those).
• Jeff Geerling and Andreas Spiess sort of book-end the spectrum of single-board computers and electronics I’m particularly interested in (most of the IoT and home automation channels are… boring and shallow by comparison).

Then there’s music, game development, and a choice bit of gadgetry:

• Floyd Steinberg and Charles Cornell are also good examples of the spectrum of music-related stuff I’m interested in (from custom synths to music theory with actual fun examples).
• Stimming’s latest series is a great example of why I’ll watch or listen to anything he does.
• Benn Jordan is surprisingly great–I’ll just leave it at that, you go watch and let it sink in.
• Jakob Haq is definitely the one to go for if you care about music on iOS in any way.
• Loopop is the gold standard for music gadgetry reviews, and I have to avoid watching it to avoid gear acquisition syndrome.
• The Linux Experiment turned out to be a quite entertaining way to expose my kids to the ecosystem, and Nick has a good sense of humor.
• Sebastian Lague posts amazing stuff about game development that me and my kids quite enjoy, and David R.B. has been posting live updates on his development of Arcadian Rift that provide a lot of background on what game development is really like these days (also something we all enjoy here).
• Michael Fisher takes me back to when I was a mobile product manager, and his reviews have amazing production value. My hat’s off to you, sir.

And, finally, noclip runs very, very good documentaries about the history of some gaming franchises (some harking back to the 8-bit days).

In fact, one of the best reasons I spend a fair chunk of time on YouTube turns out to be indie documentaries, which have been steadily increasing in quality and production value over the years–so it’s definitely not all influencer fodder.

I probably game a couple of hours or so a week at most, and I must say it’s become mostly casual. Besides some classic arcade emulation, I have been enjoying Dead Cells, and pop into xCloud every few days to quite literally sample the catalog–I just don’t have the time for prolonged gaming sessions.

That time goes up and down as I mix up some indoors exercise with Beat Saber, Pistol Whip and similar games on the Quest, but not by much since my work schedule has been making it hard do do any form of exercise during the day, and I don’t really think of those as gaming–they are definitely something I enjoy doing and wish I could do more, but VR is still very much a separate state of mind for me.

As far as triple-A gaming goes, I’d say that other while on vacation it is a rainy day occupation, at best. I haven’t finished Control or Halo Infinite yet. In fact, as I noted recently, I find Halo Infinite’s final section to be somewhat gratuituous and utterly boring. It was ultimately so frustrating I just went and started replaying the open world section again.

Since I now work in telco again, I spent a lot of time reading various technical tomes regarding 5G “advancements”.

Which, to be blunt, turned out to be mostly rationalizations of why earlier 3GPP releases were dismal failures at adopting mainstream technology, and a lot of over-engineered explanations as to how virtualization, Kubernetes and IP SDNs are eating traditional vendors’ lunches.

The only really interesting bits where I learned completely new stuff were, as usual, about the radio layer–everything else is just obfuscation of things that anyone in the ISP and cloud-native spaces have taken for granted for a few years now.

Oh, and the Programming Rust book, for which unfortunately I have no immediate professional use.

Sadly, none of the non-technical books I read this year really stuck with me. I did make a stab at going through Charles Stross’ and John Scalzi’s latest oeuvres (and I found The Kaiju Preservation Society to be quite entertaining), but my search for brainier books sort of fizzed out and I ended up wasting time on pop sci-fi instead to take my mind off things.

Again, I intend to make up for that next year. We’ll see how that goes…

I wanted to write a bash script that would start a Datasette server running, run a request against it using curl, then stop the server again.

It should then return an exit code indicating if the curl request was succesful or not.

Research notes in this issue - here's the final script I came up with:

#!/bin/bash

# Generate certificates
python -m trustme
# This creates server.pem, server.key, client.pem

# Start the server in the background
datasette --memory \
    --ssl-keyfile=server.key \
    --ssl-certfile=server.pem \
    -p 8152 &

# Store the background process ID in a variable
server_pid=$!

# Wait for the server to start
sleep 2

# Make a test request using curl
curl -f --cacert client.pem 'https://localhost:8152/_memory.json'

# Save curl's exit code (-f option causes it to return one on HTTP errors)
curl_exit_code=$?

# Shut down the server
kill $server_pid
sleep 1

# Clean up the certificates
rm server.pem server.key client.pem

echo $curl_exit_code
exit $curl_exit_code

There are a few additional tricks in this - it's running python -m trustme to create self-signed certificates in the current directory which are used for the test - but the key parts are these:

• datasette ... & - starts Datasette running as a background process
• server_pid=$! - records the PID of the server we just started so we can shut it down later
• curl -f ... - makes a curl request, but returns an exit code that indicates if the request succeeded or was a 404 or 500 error or similar
• curl_exit_code=$? - records that exit code for later
• kill $server_id - causes the server to exit - I then did a sleep 1 to provide time for it to output its shutdown text to the terminal
• exit $curl_exit_code - the exit code for the Bash script is now the same as the exit code returned by curl

I'm running this script in CI in GitHub Actions.

An improved version by Jan Lehnardt

Jan Lehnardt submitted a pull request with this improved version:

#!/bin/bash

# Generate certificates
python -m trustme
# This creates server.pem, server.key, client.pem

cleanup () {
    rm server.pem server.key client.pem
}

# Start the server in the background
datasette --memory \
    --ssl-keyfile=server.key \
    --ssl-certfile=server.pem \
    -p 8152 &

# Store the background process ID in a variable
server_pid=$!

test_url='https://localhost:8152/_memory.json'

# Wait for the server to start

# h/t https://github.com/pouchdb/pouchdb/blob/25db22fb0ff025b8d2c698da30c6c409066baa0c/bin/run-test.sh#L102-L113
waiting=0
until $(curl --output /dev/null --silent --insecure --head --fail --max-time 2 $test_url); do
    if [ $waiting -eq 4 ]; then
        echo "$test_url can not be reached, server failed to start"
        cleanup
        exit 1
    fi
    let waiting=waiting+1
    sleep 1
done

# Make a test request using curl
curl -f --cacert client.pem $test_url

# Save curl's exit code (-f option causes it to return one on HTTP errors)
curl_exit_code=$?

# Shut down the server
kill $server_pid
waiting=0
#         show all pids
#         |       find just the $server_pid
#         |       |                  don’t match on the previous grep
#         |       |                  |            we don’t need the output
#         |       |                  |            |
until ( ! ps ax | grep $server_pid | grep -v grep > /dev/null ); do
    if [ $waiting -eq 4 ]; then
        echo "$server_pid does still exist, server failed to stop"
        cleanup
        exit 1
    fi
    let waiting=waiting+1
    sleep 1
done

# Clean up the certificates
cleanup

echo $curl_exit_code
exit $curl_exit_code

There's not much extra commentary I can add to this - Jan's inline comments are fantastic!

I really like the waiting=0 pattern here for retrying up to 4 times.

Worth breaking down the curl command here a bit:

curl --output /dev/null --silent --insecure --head --fail --max-time 2 $test_url

It's avoiding any output at all with a combination of writing output to /dev/null and using --silent to turn off logging.

It uses --insecure because our server is running with a self-signed certificate, which will produce errors without this option - and here we just want to poll until the server is up and running.

--max-time ensures each poll waits a maximum of two seconds.

And as before, --fail causes curl to return an exit code that indicates if the request was successful or not.

Jan used this as the impetus to start a new library of shell utility functions.

My first delivery of Chefs Plate meals arrived today.¹

My company sent everyone SnackMagic boxes with treats for the holidays, and the box also came with a discount code for Chefs Plate, a meal delivery service.²

I decided to try it even though I’m not really the target market for it. We live close by to grocery stores and I enjoy cooking things from raw ingredients.

But, it was steeply discounted, and I’ve always been curious about these services. I don’t know how long this one has been available or what others there are in Canada / are live in Vancouver.

I picked the meals ahead of time with Rachael. This order included:

• Herby Panko-Crusted Chicken
• Creamy Butternut Squash Orzo
• Loaded Beef Burrito Bowls

Usual price $60CAD, discounted price $27CAD. For 2 people for 3 meals that’s pretty good.

Unboxing

You can pick different days of the week for delivery and they leave it at your door.

There is a glossy double sided print out of meal prep and cooking instructions. My first reaction was that this is a lot of packaging and paper and inks and it’s definitely wrecking the planet.

The kits are in a big cardboard box, each meal’s ingredients in a bag. The bags are compostable so you can use them for your bucket liner (as it says in the explainer).

There’s a compartment in the bottom with an ice pack and things that need to be kept cold. This week, chicken breasts and some ground beef. There was also a loose whole bulb of garlic.

The kit bags aren’t full to the brim, so the tops could be folded down. But I still had to find room in my apartment sized fridge to stick them.

Making Herby Panko-Crusted Chicken

We decided to make the chicken for dinner.

Here’s everything from the kit laid out on the counter. The sticker on the bag has a full ingredient list for everything like the spice blend and then mayo, plus the nutrition info for the whole meal.

The double sided instructions are good. Tells you the equipment and other stuff you need (basics like salt, pepper, oil, honey). And the instructions are laid out in time it takes to cook. So preheat the oven right away and get the potato wedges started.

And now we’ll teleport all the way to the plated meal. Chicken breast covered in mayo and panko-crusted, potato wedges baked in the oven with Montreal Steak Spice, and honey glazed carrots.

It was good! We added our own salt and pepper to everything as instructed and it was nice and flavourful.

It called for parchment paper for both the potato wedges and the final bake of the chicken. That was a nice tip for me to save on clean up and keep things crisp.

It definitely reminded me that potato wedges are something we can make any time!

The carrots are pretty much how I make them anyway - finished with a little butter and honey. Dill if you like, or I’ll add cumin seeds sometimes.

The mayo-instead-of-egg was another nice tip that I’ll be using.

All in all, a good, tasty first experience.

On Meal Kits Generally

The discounts to get me on a subscription is very much an investment driven business.

The packaging feels like a lot. But: no plastic other than on the meat.

There is an app, so maybe I can opt out of the printed meal instructions.

Are the ingredients local? Could one check a box and have 80% local ingredients?

Could one apply this to community supported agriculture?

This meal was ~$9CAD with discount, ~$20CAD without. For 2 people, that’s reasonable. The chicken breasts alone would cost me $5-6. Everything else low cost: 2 carrots, 3 potatoes, mayo, steak spice. Maybe another $1.

We used to do zero take out. And then the pandemic. Now we’re at 1-2 take out meals per week. This would be cheaper.

I’m interested in perhaps doing some food sharing with people, where everyone cooks a larger quantity and then shares with the group. Don’t know if I’ll get to it as a habit this year, but it’s the sort of thing I’d like to see more of.

And bulk buying!

I have done zero research on this in Vancouver. If you have thoughts on meal kits or on fun community food locally, let me know!

Tonight’s dinner was a trial run of #ChefsPlate, a meal kit service that I’m doing a trial run of.

Full write up on the blog blog.bmannconsulting.com/2022/12/17/trial-run-of.html

Here’s a quick one for my friends still on Twitter. There is a new policy against promotion of alternative social platforms . If you have not moved to Mastodon yet, do it now. It will be much easier to use tools to make the move now, while many in your Twitter social graph are still reachable. Also: get a blog or personal web site under your own domain ASAP and link it from your Twitter profile.

I’m a big fan of kickstands on my city bikes. My winter bike has not had one for a while, so I finally broke down and bought one. This particular bike didn’t look like it would be compatible with a mounting position just ahead of the rear wheel so I elected to get one that mounts off of the chainstay. In addition, the one that I chose also has an additional clamp that attaches to the seat stay for greater rigidity.

It looks like this model will fit a wide variety of bikes. However, in this particular case, since my bike has something coming down from the chain stay to mount the cable adjuster for the rear internal hub, there was a bit of interference with the kickstand mount.

A little early morning work with the Dremel.

Much better.

One of the things that I didn’t fully appreciate is that with a single sided kickstand, it is easy to roll the bike even with the kickstand down. Additionally, the rear mount eliminates any interference with the pedals.

This is in contrast to my plescher double legged stand on the pink bike, which is great, but once the stand is down, the bike cannot roll.

Mounted a flag for today’s ride with Santa. I haven’t used this flag since 2015.

For those joint us for the Cycling Good Cheer ride, I’ll see you either at North York Centre, or at the meet up at Bloor and Yonge.

INTRODUCTION: [This trend], which everyone knows about, and [that trend], which is so incredibly arcane that you probably didn’t know about it until just now, and [this other trend over here] which might seem, at first blush, to be completely unrelated, when all taken together, lead us to the (proprietary, secret, heavily patented, trademarked, and NDAed) insight that we could increase shareholder value by [doing stuff]. We will need $ [a large number] and after [not too long] we will be able to realize an increase in value to $ [an even larger number], unless [hell freezes over in midsummer].

Neal Stephenson, Cryptonomicon, 1999, pg. 238.

Teaching and learning programming is much harder than actually programming: the mental models required are more complex (1) and the tools we need don’t exist (2, 3). Following Stephenson’s guideline quoted above, though, I think we could combine a couple of existing technologies in a new way to make this process easier, faster, and more enjoyable.

Step 1: Replay

Let’s start with asciinema, which allows you to record and play back a terminal session as text rather than as a video. The difference is important: “text” in a video is just pixels that happen to be arranged in the shapes of letters, while asciinema records and replays actual text with timings. If you pause an asciinema recording you can select and copy the text, and search engines and accessibility tools can “see” the text as well.

Replay isn’t limited to text. Tools like Selenium can capture mouse events as well as keystrokes and save them with timings so that you can replay your interactions with a browser. This is obviously useful for testing, but most people don’t take advantage of three corollary benefits:

1. The sequence of events (“move the mouse here, right click, press J, press U”) is a searchable, editable artifact. You can, for example, find all the interaction records that involve scrolling or stitch together two existing interactions to see what happens if they run consecutively without having to re-do either.

2. The browser is still in its final state when replay finishes, so you can do “what if?” experiments.

3. You can single-step or interrupt a sequence during playback, which allows “what if?” experiments in the middle of replay.

The first property is very useful for teachers, but the second and third are potentially game changers for learners. As an analogy, people can learn to fly by reading books that have photographs of cockpit controls, but interactive flight simulators are more effective and more engaging.

Step 2: Voiceover

Replay-with-branching has two shortcomings, the first of which is, where is the explanation? Wasim Lorgat has written a brief tutorial on how to turn terminal demos into animated SVGs that can be embedded in a page beside an explanation, but reading the latter while watching the former is a heavy cognitive load. What we want is a voiceover that’s synchronized with the event stream that’s driving the animation, which is what Scrimba provides. Their “scrims” aren’t videos: instead, each one is a narrated replay of a programming session in an in-browser coding sandbox. If you pause at any point, you can search or copy the text from the animated coding window; if you scroll backward or forward you see that text change just as you would using “undo” and “redo” in an editor.

Note: I wanted to embed a demo scrim in this blog post, but it looks like their creators have gone into the coding bootcamp business and the tools to make and replay scrims are now closed-source. If I’m wrong, please correct me and I’ll update this post.

Step 3: Branching

I don’t know if anyone has married voiceover with a browser testing tool, but as we discovered during the Browsercast project, modern browser APIs make it pretty easy to synchronize media with browser events. (You may have to dismiss some warnings or press the space bar to start that demo: the code is ten years old and I haven’t had time to update it.) However, that still leaves us with our second problem: exploration. Learners don’t want to watch programming—they want to change things and play “what if?” and then come back back to the lesson and pick up where they left off.

But if learners change things, the lesson might stop working. For example, a learner might delete a column from a dataframe that the rest of the lesson depended on. The instructor can try to accommodate this by creating checkpoints that learners can restart from, but I’ve always hated it when game designers tell me I’m only allowed to save the game at certain points. Checkpointing also puts a heavy burden on instructors: if they change something early in the lesson, they often have to go through and manually re-create all of the subsequent checkpoints. In practice, lesson maintenance is as hard as software maintenance (see the note above about Browsercast), and any tool that makes it harder is unlikely to be popular.

Luckily, there’s a solution. As far back as 2009, projects like Snowflock proved that cloning a virtual machine can be almost as fast as forking a process. The technical details are fascinating, but in brief, the trick is that the new VM shares pages of virtual memory with its parent until one or the other tries to modify something, at which point the child VM gets its own copy of the affected page to scribble on. This “copy on write” strategy means that the child VM can start running almost immediately, but it also means that the parent can be cloned at any time, which is exactly what we need for teaching.

Step 4: The Stephenson Moment

So let’s pull this together. What I want is a tool that:

1. Records my interactions with my desktop as a stream of logical events rather than as video.

2. Simultaneously, records what I’m saying.

3. Allows me to edit the combination in the way that Camtasia or OBS lets me edit screencasts.

4. Lets a learner replay the result in a sandbox VM on their computer or in the cloud.

5. Lets the learner stop the replay at any point and fork a new VM to play around in.

6. When she’s done, lets her throw away that VM and resume the lesson where she left off.

Let’s call recordings like this “sessioncasts”. They’re better than screencasts, computational notebooks, and coding sandboxes for all the reasons given in the previous sections, but they’re more general and therefore more authentic. if your lesson switches between VSCode, a terminal, the browser, and back to VSCode, you can record it all in one sessioncast so that learners can see and do real-life programming. I don’t know of anything else that can offer this.

What would it take to build a sessioncasting tool? What’s the “large number” after the dollar sign in the Cryptonomicon quote? I honestly don’t know, but all the pieces exist—we just need to assemble them (for some admittedly large value of “just”). I doubt vulture capitalists would fund a tool to make teaching and learning easier, but sessioncasts would be a boon for tech support as well: if I can’t reproduce your bug, you can record a sessioncast and send it to me so I can poke around. There are obvious security and privacy concerns, but there have been times when I would have traded minor but useful body parts for this. If you’re already working on this or something better, please give me a shout.

After taking vows of celibacy and abstinence and foregoing all of our material possessions for homespun robes, we (viz. appended resumes) will move into a modest complex of scavenged refrigerator boxes in the central Gobi Desert, where real estate is so cheap that we are actually being paid to occupy it, thereby enhancing shareholder value even before we have actually done anything. On a daily ration consisting of a handful of uncooked rice and a ladleful of water, we will [begin to do stuff].

Neal Stephenson, Cryptonomicon, 1999, pg. 240.

The streak of crazy weeks got extended by one more. Whatever happens next week is the final one, after that it’s AFK and pen down for two weeks. Much needed too.

This week I

• Did lots of bookkeeping on Monday, and this entire Sunday, while E and Y went ice skating, until 22:00. Finally finished up my taxes for E and my personal companies, and for my holding company. Early next week I need to finalise taxes for the final entity, and then what I had intended to do early November is behind me, over 5 weeks late.
• Took Y gift hunting for a birthday party she attended this week
• Created and gave a short presentation about ActivityPub and Mastodon during an open meet-up of the Dutch Forum for Standardisation (which sets the government’s mandatory and recommended open standards). A fun meet-up and good conversation, triggered ofcourse by the chaos at Twitter, but the conversation quickly moved passed that and focussed on ActivityPub and governance.
• Facilitated an evening long unconference, as final session of the yearly Masterclass Network Politics organised by Internet Society Netherlands. I’ve done this before, and it’s fun. During the final session all current participants and a number of participants of earlier editions decide on a few topics to revisit or discuss more. In three rounds we addressed almost a dozen topics. The atmosphere was great. I enjoy it when everyone is eagerly contributing substance to conversations.
• Continuing on from last week’s process memo and discussion with decision makers, this week saw a second round of it. Meaning I spent a long evening writing up a decision document, which then went through a few iterations with the client, to ensure a decision was taken this week. All mostly to ensure that the administrative reality reflects the reality on the ground, because the steps it decides on would have been taken anyway. It did contain some budget provisions for next year already, so that’s progress.
• Had the weekly client meetings
• Discussed a Green Deal statistical dashboard with the Dutch Statistics Office, to get a sense of what they’re planning to do with it
• Prepared a session with the Ministry for the Interior for early February, as kick-off of the implementation process of the EU High Value Data list
• Had a planning session of a visit by the Dutch minister for digitisation which will take place in January
• Visited a client’s Christmas party
• Made to presentations about the impact of EU data regulation for a province and for a combined session with speakers from the Minsitry for the Interior and the Cadastral Office. The latter was a good session.
• Participated in a workshop about GAIA-X and the EU data space(s), which was useful to get a view about the latest steps taken
• Discussed a potential project with the National Police, and received a RFP from a client for most of my time for next year.
• Spent a lovely evening over dinner with my team, at the home of my business partner F. Celebrating Christmas and looking back at this year. Next week we’ll take a full day to look at next year. The final week of the year we’re giving everyone off from work, so they won’t have to take paid leave for it. It was a busy year, and it’s time to take a rest.
• Had conversations with a range of people we invited to join the interprovincial ethics committee. By mid next week we’ll have finalised the members list, so we can start mid January.
• Had a haircut
• Went into the city center with E and Y Saturday morning for coffee, some shopping and poffertjes, steaming hot in the sub-zero temperatures this week.
• Finished the week with a power outage, during dinner tonight the entire street went dark. Took some 90 minutes or so.

The roof of the early 17th century Theatrum Anatomicum in the 15th century Waag building in Amsterdam. In this space I hosted my unconference for the Masterclass Network Politics Monday evening.

Today was the third annual ride down Yonge St with Santa, organized by the Toronto Community Bikeways Coalition to promote cycling infrastructure along Yonge St.

We are gathered at North York Centre. Here Rudolph is interviewing Santa for the media.

The gang’s all here so we can get ready to go.

I’ll note that Santa has a new ride this year, courtesy of Happy Fiets. His trailer has a trick set up that detects when you are slowing down, and then applies the brakes on the trailer so that you are not overrun.

Off we go.

Just after the most dangerous part of the entire ride: the 401 underpass.

Rudolph leads the way up from Hogg’s Hollow.

Santa is all smiles going up the hill this year.

Up the other incline from Lawrence.

We pause at Davisville where another group is waiting at the northern end of the Yonge cycle track.

Good to see Joe and Kay.

Also we are joined by Frosty the Snowman.

Starting up again from Davisville.

Here’s another one of those bus stop platforms. I was looking for a ramp, but of course in this case, they are not necessary since there is an adjacent cycle track.

Passing by the Kartik Saini ghost bike.

Another pause at the Summerhill LCBO.

Some refreshments and hot chocolate are on offer.

Alison from CycleTO is on message.

Dave of NRBI, and Arianne.

Frosty is revealed to be one of the Richardsons.

We turn west on Bloor. Here Santa is greeting the shoppers along this stretch.

We’ve turned south at the ROM.

Santa takes the lane.

Getting ready to cross at Armoury.

Headed to City Hall past the courthouse.

Here we are at Nathan Phillips Square.

Councillor Diane Saxe reminds us that there are votes coming up in the spring having to do with the bike lanes on Yonge.

Quite the cast of characters.

Santa and Rudolph.

Finally, Santa with all his reindeer.

Thanks once again to Albert and TCBC for organizing today’s ride, to Happy Fiets for providing rides for Santa, Rudolph and Frosty, to the marshalls who helped keep everyone safe, to Councillor Saxe who rode with us.

Thanks also to everyone who showed up, especially due to the overlap with the World Cup final. Have a Happy Holiday, and here’s hoping for a safe 2023 for everyone.

And as always, Bromptoning has a wonderful video of the whole ride.

Does anyone know how to get to the source photo from the #iOS “Photo Shuffle”?

I have mine set to rotate hourly and it’s great to see them, but then I’d like to perhaps post that photo or remember where / when it was by going to it in my photo library.

You press and hold to get to this customize screen (see screenshot)

I left Twitter some time ago. Today I learned that Elon Musk has blocked Mastodon on Twitter. So much for his claimed support for free speech.

I would ask everyone who reads this WordPress post to let their contacts know that joining Mastodon is not actually very difficult – as some misinformers would have you think.

You can find me at Mastodon

There is also a Wikipedia page (of course)

https://en.wikipedia.org/wiki/Mastodon_(social_network)

I got my personal site imported into #LogSeq and have a workflow where I can edit on my phone or desktop and commit / push to GitHub where it auto builds to https://logseq.bmannconsulting.com

The graph view is always fun to play with even if it’s not useful ;)

Datasette 1.0a2: Upserts and finely grained permissions describes the new upsert API and much improved permissions capabilities introduced in the latest Datasette 1.0a2 alpha release.

I have prostate cancer and will begin treatment in January. I wanted to share my experience for several reasons. First, if you have a prostate, you might benefit from hearing somebody else’s experience — since you may find yourself in the same boat I’m now in. Second, I made my decisions in consultation with my … Continued

The post My cancer appeared first on without bullshit.

Wastewater

Jeff’s spreadsheet charts showing COVID-19 levels in wastewater (using MetroVan data):

Non-COVID Respiratory Illnesses

If you haven’t been living under a rock, you’ve probably heard about the health care system getting slammed by respiratory illnesses, but it looks like the levels peaked between ~26 Nov and 3 Dec!

From the BC CDC Pathogen Characterization page:

Adult non-COVID respiratory viruses:

Pediatric non-COVID Van/North Shore/Richmond:

Children’s Hospital (Vancouver) only:

US Pacific Northwest flu positivity also looks like it peaked around 26 Nov. From the US CDC FluView Interactive:

This article talks about an adenovirus outbreak in Wisconsin.

This article from the USA talks about invasive strep throat. This article from the UK says that at least 19 children have died from strep A.

This is a nice article from the USA which talks about the onslaught of infections this season.

This article talks about the difference between cold, flu, COVID-19, and RSV, and has this very nice chart: