Rolandt

In May of this year, we saw a development path with first devices through certification by the end of the year in 2021. With the completion of several test events and forecasting, our members have updated the schedule to reflect a commitment to ensuring that the SDK, and related tools, are ready to meet the expectations of the market when launched and enable a large ecosystem of interoperable Matter products. Our refined plans include ongoing SDK and certification program development in 2H 2021, targeting a “pre-ballot” version of the technical spec available to members at year’s end. In the first half of 2022, we expect to see the SDK released, the first devices through certification, and our formal certification program opening.  

Tobin Richardson, President & CEO, Connectivity Standard Alliance

The smarthome market is a mess. Incompatible products, lots of proprietary gateways, different architectures, privacy issues. You will need to replace everything in the future, and we can only hope for better interoperability with a standard. I was hoping to see products towards the end of the year, but now Matter is pushed out further into the future.

This week I broke my streak of not sending out the Datasette newsletter, figured out how to use Sphinx for Python class documentation, worked out how to run Datasette on GitHub Codespaces, implemented Datasette column metadata and got tantalizingly close to a solution for an elusive Datasette feature.

API reference documentation for sqlite-utils using Sphinx

I've never been a big fan of Javadoc-style API documentation: I usually find that documentation structured around classes and methods fails to show me how to actually use those classes to solve real-world problems. I've tended to avoid it for my own projects.

My sqlite-utils Python library has a ton of functionality, but it mainly boils down to two classes: Database and Table. Since it already has pretty comprehesive narrative documentation explaining the different problems it can solve, I decided to try experimenting with the Sphinx autodoc module to produce some classic API reference documentation for it:

Since autodoc works from docstrings, this was also a great excuse to add more comprehensive docstrings and type hints to the library. This helps tools like Jupyter notebooks and VS Code display more useful inline help.

This proved to be time well spent! Here's what sqlite-utils looks like in VS Code now:

Running mypy against the type hints also helped me identify and fix a couple of obscure edge-case bugs in the existing methods, detailed in the 3.15.1 release notes. It's taken me a few years but I'm finally starting to come round to Python's optional typing as being worth the additional effort!

Figuring out how to use autodoc in Sphinx, and then how to get the documentation to build correctly on Read The Docs took some effort. I wrote up what I learned in this TIL.

Datasette on GitHub Codespaces

GitHub released their new Codespaces online development environments to general availability this week and I'm really excited about it. I ran a team at Eventbrite for a while resonsible for development environment tooling and it really was shocking how much time and money was lost to broken local development environments, even with a significant amount of engineering effort applied to the problem.

Codespaces promises a fresh, working development environment on-demand any time you need it. That's a very exciting premise! Their detailed write-up of how they convinced GitHub's own internal engineers to move to it is full of intriguing details - getting an existing application working with it is no small feat, but the pay-off looks very promising indeed.

So... I decided to try and get Datasette running on it. It works really well!

You can run Datasette in any Codespace environment using the following steps:

1. Open the terminal. Three-bar-menu-icon, View, Terminal does the trick.
2. In the terminal run pip install datasette datasette-x-forwarded-host (more on this in a moment).
3. Run datasette - Codespaces will automatically setup port forwarding and give you a link to "Open in Browser" - click the link and you're done!

You can pip install sqlite-utils and then use sqlite-utils insert to create SQLite databases to use with Datasette.

There was one catch: the first time I ran Datasette, clicking on any of the internal links within the web application took me to http://localhost/ pages that broke with a 404.

It turns out the Codespaces proxy sends a host: localhost header - which Datasette then uses to incorrectly construct internal URLs.

So I wrote a tiny ASGI plugin, datasette-x-forwarded-host, which takes the incoming X-Forwarded-Host provided by Codespaces and uses that as the Host header within Datasette itself. After that everything worked fine.

sqlite-utils insert --flatten

Early this week I finally figured out Cloud Run logging. It's actually really good! In doing so, I worked out a convoluted recipe for tailing the JSON logs locally and piping them into a SQLite database so that I could analyze them with Datasette.

Part of the reason it was convoluted is that Cloud Run logs feature nested JSON, but sqlite-utils insert only works against an array of flat JSON objects. I had to use this jq monstrosity to flatten the nested JSON into key/value pairs.

Since I've had to solve this problem a few times now I decided to improve sqlite-utils to have it do the work instead. You can now use the new --flatten option like so:

sqlite-utils insert logs.db logs log.json --flatten

To create a schema that flattens nested objects into a topkey_nextkey structure like so:

CREATE TABLE [logs] (
   [httpRequest_latency] TEXT,
   [httpRequest_requestMethod] TEXT,
   [httpRequest_requestSize] TEXT,
   [httpRequest_status] INTEGER,
   [insertId] TEXT,
   [labels_service] TEXT
);

Full documentation for --flatten.

Datasette column metadata

I've been wanting to add this for a while: Datasette's main branch now includes an implementation of column descriptions metadata for Datasette tables. This is best illustrated by a screenshot (of this live demo):

You can add the following to metadata.yml (or .json) to specify descriptions for the columns of a given table:

databases:
  fixtures:
    roadside_attractions:
      columns:
        name: The name of the attraction
        address: The street address for the attraction

Column descriptions will be shown in a <dl> at the top of the page, and will also be added to the menu that appears when you click on the cog icon at the top of a column.

Getting closer to query column metadata, too

Datasette lets you execute arbitrary SQL queries, like this one:

select
  roadside_attractions.name,
  roadside_attractions.address,
  attraction_characteristic.name
from
  roadside_attraction_characteristics
  join roadside_attractions on roadside_attractions.pk = roadside_attraction_characteristics.attraction_id
  join attraction_characteristic on attraction_characteristic.pk = roadside_attraction_characteristics.characteristic_id

You can try that here. It returns the following:

The columns it returns have names... but I've long wanted to do more with these results. If I could derive which source columns each of those output columns were, there are a bunch of interesting things I could do, most notably:

• If the output column is a known foreign key relationship, I could turn it into a hyperlink (as seen on this table page)
• If the original table column has the new column metadata, I could display that as additional documentation

The challenge is: given an abitrary SQL query, how can I figure out what the resulting columns are going to be and how to tie those back to the original tables?

Thanks to a hint from the SQLite forum I'm getting tantalizingly close to a solution.

The trick is to horribly abuse SQLite's explain output. Here's what it looks like for the example query above:

The magic is on line 12: ResultRow 3 3 means "return a result that spans three columns, starting at register 3" - so that's register 3, 4 and 5. Those three registers are populated by the Column operations on line 9, 10 and 11 (the register they write into is in the p3 column). Each Column operation specifies the table (as p1) and the column index within that table (p2). And those table references map back to the OpenRead lines at the start, where p1 is that table register (referered to by Column) and p1 is the root page of the table within the schema.

Running select rootpage, name from sqlite_master where rootpage in (45, 46, 47) produces the following:

Tie all of this together, and it may be possible to use explain to derive the original tables and columns for each of the outputs of an arbitrary query!

I was almost ready to declare victory, until I tried running it against a query with an order by column at the end... and the results no longer matched up.

You can follow my ongoing investigation here - the short version is that I think I'm going to have to learn to decode a whole bunch more opcodes before I can get this to work.

This is also a very risk way of attacking this problem. The SQLite documentation for the bytecode engine includes the following warning:

This document describes SQLite internals. The information provided here is not needed for routine application development using SQLite. This document is intended for people who want to delve more deeply into the internal operation of SQLite.

The bytecode engine is not an API of SQLite. Details about the bytecode engine change from one release of SQLite to the next. Applications that use SQLite should not depend on any of the details found in this document.

So it's pretty clear that this is a highly unsupported way of working with SQLite!

I'm still tempted to try it though. This feature is very much a nice-to-have: if it breaks and the additional column context stops displaying it's not a critical bug - and hopefully I'll be able to ship a Datasette update that takes into account those breaking SQLite changes relatively shortly afterwards.

If I can find another, more supported way to solve this I'll jump on it!

In the meantime, I did use this technque to solve a simpler problem. Datasette extracts :named parameters from arbitrary SQL queries and turns them into form fields - but since it uses a simple regular expression for this it could be confused by things like a literal 00:04:05 time string contained in a SQL query.

The explain output for that query includes the following:

So I wrote some code which uses explain to extract just the p4 operands from Variable columns and treats those as the extracted parameters! This feels a lot safer than the more complex ResultRow/Column logic - and it also falls back to the regular expression if it runs into any SQL errors. More in the issue.

TIL this week

• Tailing Google Cloud Run request logs and importing them into SQLite
• Find local variables in the traceback for an exception
• Adding Sphinx autodoc to a project, and configuring Read The Docs to build it

Releases this week

• datasette-x-forwarded-host: 0.1 - 2021-08-12
  Treat the X-Forwarded-Host header as the Host header
• sqlite-utils: 3.15.1 - (84 releases total) - 2021-08-10
  Python CLI utility and library for manipulating SQLite databases
• datasette-query-links: 0.1.2 - (3 releases total) - 2021-08-09
  Turn SELECT queries returned by a query into links to execute them
• datasette: 0.59a1 - (96 releases total) - 2021-08-09
  An open source multi-tool for exploring and publishing data
• datasette-pyinstrument: 0.1 - 2021-08-08
  Use pyinstrument to analyze Datasette page performance

The New York Times go with the angled arrows to show the shifts in racial population. The red-orange arrows show an increase in the share of white population, and the teal arrows show an increase in the share of people of color. Longer arrows mean a greater percentage point change.

Whereas The Washington Post focused more on the changes for each demographic individually, NYT focused more on how two broad groups compared.

Tags: census, New York Times, race

Vox grabbed some data from the Kaiser Family Foundation and did a story a few days ago on it, then posted a different visualization of it that attracted some attention:

I’m a pretty ardent donut detractor, but I have to also admit that they work pretty well for this use case, and we can reproduce the graphic in R quite nicely. The following code chunk:

• scrapes the datawrapper chart Vox embedded in their original article
• extracts the data from it using V8
• reformats it and uses some basic ggplot2 idioms for making the facets

library(V8) # we need to execute javascript to get the data
library(ggtext) # pretty ggtext
library(stringi) # some basic string ops
library(rvest) # scraping!
library(hrbragg) # remotes::install_github("hrbrmstr/hrbragg")
library(tidyverse) # duh

ctx <- v8() # init the V8 engine

pg <- read_html("https://datawrapper.dwcdn.net/jlEL9/6/") # get the data — this is embedded in the Vox article

# clean up the javascript so we can evaluate it
html_nodes(pg, xpath = ".//script[contains(., 'DW_SVEL')]") %>% 
  html_text() %>% 
  stri_replace_first_fixed("window.", "") %>% 
  stri_replace_all_regex("window.__.*", "", multiline = TRUE) %>% 
  ctx$eval()

# get the dat from the V8 engine and reshape it a bit, then join it with state abbreviations
ctx$get("__DW_SVELTE_PROPS__")$data %>% 
  read_tsv(
    col_names = c(
      "state", 
      "Reported cases among fully vaccinated", 
      "Reported cases among not fully vaccinated"
    )
  ) %>% 
  gather(measure, value, -state) %>% 
  left_join(
    tibble(
      state = state.name,
      abbr = state.abb
    ) %>% 
      add_row(
        state = "District of Columbia",
        abbr = "DC"
      ) 
  ) -> voxxed

# basic ggplot idiom for faceted donuts

ggplot() +
  geom_col(
    data = voxxed,
    aes(3, value, fill = measure), # play with "3" here and below to change width of the donut
    color = NA,
    position = position_stack()
  ) +
  ggtext::geom_richtext(
    data = voxxed %>% filter(measure == "Reported cases among not fully vaccinated"),
    # change 0.2 and 0 to see what they do
    aes(0.2, 0, label = sprintf("%s<br/><b><span style='color:#8a264a'>%s%%</span></b>", abbr, value)),
    size = 5, label.size = 0
  ) +
  scale_x_continuous(
    limits = c(0.2, 3 + 0.5) # this 3 links to the 3 above. tweak 0.2 and 0.5 to see what it does to the shape
  ) +
  scale_fill_manual(
    name = NULL,
    values = c(
      "Reported cases among fully vaccinated" = "#d59e67",
      "Reported cases among not fully vaccinated" = "#8a264a"
    )
  ) +
  coord_polar(theta = "y") +
  facet_wrap(~state, ncol = 6) +
  labs(
    x = NULL, y = NULL,
    title = "Breakthrough cases are not driving the US Covid-19 surge",
    caption = "Source: Kaiser Family Foundation\nNote: Case data in recent months, as of July\nOriginal chart by Vox <https://www.vox.com/22602039/breakthrough-cases-covid-19-delta-variant-masks-vaccines>"
  ) +
  theme_inter(grid="") +
  theme(
    axis.text.x = elb(),
    axis.text.x.top = elb(),
    axis.text.x.bottom = elb(),
    axis.text.y = elb(),
    axis.text.y.left = elb(),
    axis.text.y.right = elb()
  ) +
  theme(
    panel.spacing.x = unit(0, "npc"),
    panel.spacing.y = unit(0, "npc")
  ) + 
  theme(
    strip.text = elb(),
    strip.text.x = elb(),
    strip.text.y = elb()
  ) +
  theme(
    legend.position = "top",
    legend.text = elt(size = 12),
    legend.justification = "left"
  ) +
  theme(
    plot.caption = elt(hjust = 0)
  )

I once said that the Disciplinary Committee of the Democratic City Committee was a witch hunt. I regret that. The Disciplinary Committee has not been engaged on a witch hunt, but rather on something older and worse. To explain, let me tell you a bit about my family.

We were slaves. This is a story we all tell around the table in Spring. For my mother’s ancestors, it was literally and immediately true. Her people are Starrs and Starrels, but in the old country they were the Staroselskys, because that was the name of the fellow who owned the whole extended family. Time passed, hate increased, pogroms threatened them all. The whole village got up and ran off to New York — Yonkel, The Rabbi’s Son, my great-Uncle Max, and his tiny half-brother Joel, who would become my grandfather.

Years later, my uncle Timmy (Joel Jr.), who was then in high school, was searching the newspaper for a summer job. A Chicagoland family wanted a driver. He applied and got the job. His employer turned out to be the grandson of the fellow who had owned our whole family. It’s a small world.

But before they left for a new world, the Staroselskys in Spring surely sat around a table where they, too, told the tale that “we were slaves.” If you looked at them as they reclined around that meagre table, you might imagine that they had always been there.

But they had not always been there. They hadn’t always spoken Yiddish. There hadn’t been Yiddish. Almost certainly, their ancestors had lived in the region we now call Spain. But it wasn’t Spain back then: it was a patchwork of local governments that we remember as al-Andalus. It was a place where Arabs, Christians and Jews pretty much got along.

That ended with the Reconquista: the entire land was now a Christian Nation, and illegal aliens were to be deported at once. Jews were, by definition, illegal. And so my ancestors lost everything — houses, businesses, friends — and headed East. How far? Not far: Barcelona to Poznań is roughly the distance from Jackson, Mississippi to Harlem. It’s the distance from the Oklahoma State Penitentiary to Weedpatch Camp near Bakersfield, California.

Immigration enforcement agencies were tasked with rooting out Jewish illegals. Some might be hiding in attics or basements. Some might be pretending to be Christian. For example, even if the local Democratic City Committee always met on Shabbat in a private club that discriminated against Jews, Blacks, and women, you could tell people you didn’t like politics. You could say you didn’t care for jámon Íberico, and besides, who can afford it? You could get stays and delays, continuances; with everyone tossed out of the country, calling witnesses or subpoenaing enforcement officials could take some time. They were slippery and tricky, those bad hombres the Jews.

This family saga echoes Trump’s deportation machine, but aspects hit closer to my home.

The Disciplinary Committee insists that I am guilty without dispute. We’ve heard that before.

The Disciplinary Committee insists that I must recognize their authority and that I increase my guilt each time I remind them that they do not possess the power they claim. We’ve heard that before.

The Disciplinary Committee asserts that I may address only “my behavior” and my beliefs, all of which are obviously very wrong, and that I must not mention that I have always obeyed our bylaws and rules. I am clearly in the wrong: everyone says so. We’ve heard that before.

The Disciplinary Committee demands that I confine my answer to their specific allegations, but they do not disclose these allegations or the evidence supporting them. We’ve heard that before.

The Disciplinary Committee asserts that the crimes of which I am already known to be guilty are not offenses against the bylaws or rules, but rather are things “detrimental to the mission of the MDCC.” Anything might be detrimental to the mission of the MDCC in someone’s view, but I must bow my neck before their (improperly) constituted power and accept whatever sentence it pleases them to bestow. We’ve heard that before.

The Disciplinary Committee demands proof of my beliefs, even those incapable of proof. We’ve heard that before.

I apologize: this was not a witch hunt. It was a Jew hunt, an inquisition. ( They didn’t think they were looking for Jews, I expect; perhaps they told themselves they were looking for a way to rid themselves of an unfriendly, upsetting, loud and unpleasant fellow with a funny accent.)

Here’s a description of another such affray, from exactly 79 years ago.

In August 1942, Germany’s Police Battalion 101 was commanded by Capt. Wohlauf, a career police officer. Wohlauf had recently been married, and his wife joined him in Poland for the sort of honeymoon that was possible that year. They weren’t political, or prominent, or well connected, or special. They weren’t Nazis. If you were at their Hamburg wedding the previous May, you'd have thought the policeman had found himself a nice young lady. But on August 25, Wohlauf’s new wife found an unusual local amusement.

They entertained themselves by flogging Jews with whips. Not only was Frau Wohlauf a party to all of this, but so were the wives of some of the locally stationed Germans, as well as a group of German Red Cross nurses. Frau Wohlauf, if conforming to her usual practice, probably carried that symbol of domination, a riding whip, with her. That day, she and the other German women got to observe firsthand how their men were purging the world of the putative Jewish menace, by killing around one thousand and deporting ten thousand more to their deaths. This is how the pregnant Frau Wohlauf spent her honeymoon.

An old joke tells lawyers that, if the facts are on your side, you should pound the facts. If the law is on your side, you should pound the law. If neither is on your side, you should pound the table.

In this situation, the law, the facts, the Democratic Party, and the table are all on my side. All that is left for the Disciplinary Committee to pound, is me.

If, as I once believed and might still be persuaded, my critics were people of good will and sense, all this could easily have been avoided.

The Democratic Party is a party, not a social club. You may be a Democrat and hold beliefs that are wrong — even repellent beliefs, so long as you keep them private. You may secretly believe things that are racist, sexist, ageists, ableist, or wrong: vote with us, support our policies, keep your repellent beliefs to yourself, and you are welcome in the party. You may in your heart believe that Christ was made before he was born, or that Christ has always been and ever shall be, that he is or is not consubstantial with his Father. You may think P==NP, that AI is impossible, that Tolkien was not a great novelist or that Cobb would have caught it. In the Democratic Party, we do not inquire into or expound upon private belief.

We are not always friends, and after this affair some of us will never more be friends, but we can be Democrats. I have, throughout this mess, acted as a Democrat: it is time for the Committee to remember what it is.

If I have been abrasive, that was because that was the manner I thought best calculated to achieve my goals. There is no rule in the Democratic Party that says, Democrats must be well liked. That club meets down the street. Far from being grounds for expulsion, it was recently the policy of MDCC leadership to reduce membership by inviting members to resign. (One prominent activist bitterly protested these efforts, which she believed were directed at excluding gay people from the Democratic City Committee. Shortly before, a member of the Executive Committee had confided to another member that she supported the Pride Parade even though her religion considered homosexuality a sin. The speaker was surprised to learn that the woman to whom she addressed this remark was queer, and the radiant anger of her reply ought to have taught everyone present to take the greatest care of mixing private belief into political discussion.)

My opponents might have offered assistance, advice and support; they did not. They might have treated my proposals with scrupulous fairness, and treated me with respect; they did not. They might have sought common ground: that is what parties do. They might have sought to compromise. They might have traded horses. This they did not do.

Instead, my critics insist on this confrontation, and arrange this confrontation on grounds that are deeply offensive not only to me, but also to every Democrat and indeed to every American. In this, they harmed the party directly. Indirectly, they have created circumstances that have led members of the Party to reveal abhorrent prejudices that ought to have remained their private shame, and to say what never should have been spoken.

Photo: Denkmal für die ermordeten juden europas, Yoav Aziz on Unsplash.

Testing

Three dogs have been trained in Vancouver to sniff out COVID-19 with very high accuracy!

Transmission

Yow! This preprint says that 6.8% of COVID-19 patients in England caught it in the hospital.

Vaccines

Good news! Adolescents tolerated two doses of Moderna, and the vax arm had no cases vs. four in the placebo arm. (The numbers are probably too small to be entirely meaningful, but this gets us a step closer to have Moderna approved for teens.)

Variants

This study from Qatar looked at vaccine effectiveness against Delta.

This looks really strange to me. Pfizer worse after two doses?? The paper did say that there was a long time between doses, but still, this seems really anomalous.

This preprint from England says that Pfizer and AZ are still effective against Delta.

Notwithstanding the previous paper, there have been a lot of papers which suggest that immunity against Delta is, well, crap. This tweet has this summary graphic:

The Public Health England study is very good, from what I can tell, and it says that the effectiveness against Delta is fine.

The Israel study is confounded by the fact that the most recent outbreak happened to happen in a well-vaccinated area, and some more socially isolated/insular groups didn’t get hit until later.

The Mayo study’s vaccinated cohort was from Minnesota and the unvaxxed cohort was from Minnesota, Wisconsin, Arizona, Florida, and Iowa: areas with different case loads and prior infection possibility.

In the REACT study:

• They mailed testing swabs to random addresses in England and asked people to take swabs and return them. They might not have done it right. You can imagine them not sticking it in far enough, which would give a false negative.
• They asked people to self-report. When they used that data, they estimated a vaccine effectiveness of 62%. They asked the participants for permission to access their health records. For the ones who agreed, the vaccine effectiveness was 62% against asymptomatic infection, meaning either there was some self-selection, some underlying confounding variable, or I suppose people misremembering or lying.
• In England at that time, being vaccinated correlated strongly with age.
• This was the 13th such study. Over time, the percent of people actually responding has dropped from 30% to 11%. There are dangers that the sample is becoming less representative; It is also possible that mischief in reporting is more likely with the smaller sample return size.

The bottom line is… I don’t know. I would tend to trust the Public Health England numbers more, and those say that the vaccine effectiveness is still quite good. But it’s disquieting that there have been multiple studies which say the effectiveness has dropped.

NB: England and the US mostly had a short (3- or 4-week dosing interval), while Canada has mostly had a longer dosing interval (12- or 7-week). There isn’t a lot of data out on longer dosing intervals, but the data which is out plus everything they know from other vaccines is that the longer dose should give more durable/longer-lasting protection.

I have no explanation for the Qatar study and am out of time to dig deeper.

This preprint gives further evidence that breakthrough Delta cases carry as much Delta in their noses as unvaccinated people.

Recommended Reading

This article is on a radically different / cheaper / easier way to manufacture vaccines. (Paywalled)

This article talks about what the end game for the pandemic looks like.

This article talks about kids and Delta (USA focused).

Long-Term Care Homes

Today there was a press briefing where Dr. Henry/Minister Dix announced new orders for Long-Term Care Homes. They said that they were hoping that they could cope with unvaccinated workers by testing them frequently and having more stringent PPE requirements for them. Delta had other ideas, apparently, so they are now mandating that anybody who works in a nursing home has to get vaccinated by 12 October or they get fired. Volunteers and personal service providers (I think like people who cut hair in the nursing homes?) must get vaccinated effective immediately.

I think this is the right decision but I worry that it will lead to short-term staff shortages, ulp.

Vaccinations

The Weekly Data Summary for 5 Aug has information on infection/hospitalization/death by vax status:

Transmission

Also from the Weekly Data Summary, R is now 2.1 overall, although they caution that in areas with low case numbers it shows clusters and might not represent the community at large.

Statistics

+513 new cases, +1 death, +4,792 first doses, +22,154 second doses.

Currently 81 in hospital / 33 in ICU, 3,834 active cases, 148,702 recovered.

We have 601,904 doses in fridges; we’ll use it up in 23.9 days at last week’s rate. We’ve given more doses than we’d received by 23 days ago.

We have 557,827 mRNA doses in fridges; we’ll use it up in 22.2 days at last week’s rate. We’ve given more doses than we’d received by 17 days ago.

94.1% of the people who had gotten their first dose by 7 weeks ago have gotten their second dose.

90.1% of the people who had gotten their first dose by 4 weeks ago have gotten their second dose.

86.9% of all people who have gotten their first dose have gotten their second dose.

Charts

I tried out the Franklin Food Lab in my #EastVan neighbourhood this week: frozen ramen!

I bought two & had the spicy tonkatsu for lunch, the yuzu shio in my freezer for another time. Very tasty, looking forward to trying all the flavours!

This exhibit at the Leiden national museum of antiquities about Doggerland seems worth a visit. The Guardian has a nice article about how many archeological finds from what is now the North Sea but was inhabited land eight thousand years ago, have been collected on the beaches (because Dutch beaches are replenished with sand brought in from further out at sea) by people walking on the beach.

Stephen Baxter wrote several books about a speculative history in which the inhabitants of Doggerland successfully defended against the rising sea and mesolithic tsunami that in our timeline created the southern North Sea. In his books Doggerland became an early and powerful centre of civilisation, taking on the role that ancient Rome had in our actual history. I enjoyed reading those books, and I think it might be fun to ‘backfill’ my memory of reading them with how the actual archeological finds connected to Baxter’s starting point look.

Stephen Baxter wrote Stone Spring, Bronze Summer and Iron Winter, aka the Northland trilogy, on this alternative history.

A 1929 German plan to turn the North Sea into a giant polder, basically recreating Doggerland, with dams from Norfolk to Jutland and across the Channel (shunting the Rhine and Thames southwards to the Atlantic). Image source Dutch National Archive, no known copyright restrictions.

The intersection of tasks and notes poses an interesting problem. Often, a task requires notes for context and details that can’t be captured with a single line of text. Likewise, notes very often spawn tasks of their own. The difficulty is how to harmonize the two coherently.

If you’re a Club MacStories member, you know this is something that has bedeviled Federico’s annual iOS and iPadOS review for years. He solved the problem by combining Obsidian with Todoist’s web API linking the two apps together in a way that complements the way he writes.

Federico’s approach takes advantage of the web technologies underlying those apps. It’s a powerful solution, but it’s not a fully native approach technically or from a design standpoint. The technique is also less suited for someone who isn’t writing thousands of words most days, and instead, just needs to flesh out their tasks with context than the single line of text many apps offer. Fortunately, there are many alternative approaches to the task and note-taking conundrum, including a new one out today from Cultured Code, the maker of Things that I like a lot.

If you write in Markdown, Things is fully capable for a drafting a story like this one.

Instead of injecting tasks into notes, Things brings a full-featured note-taking solution into version 3.14 of Things. Adding a note to a task isn’t new to Things, but the latest update expands the feature significantly. Using Markdown syntax, you can now create headings, make text bold or italic, and add bulleted and numbered lists, links, code blocks, and highlight text. The formatting is rendered inline, providing a sense of structure and style to notes. For anyone unfamiliar with Markdown syntax, Cultured Code has also created a handy guide.

Things’ bulleted lists support multiple levels of indentation based on the number of spaces that precede the bullet. Everything is neatly lined up and orderly, although I do have one quibble. I’m used to using the tab key to indent and Shift + Tab to outdent bulleted lists, which is common to most text editors and note-taking apps. Unfortunately, because the tab key is used to move the focus between UI elements in Things, to increase the level of indentation, creating a nested list, you’ll need to back up, add a space, and then move back to where the text of your note goes. I do appreciate, however, how you can cut and paste a bulleted item from one spot to another in a list without winding up with a duplicated bullet at the beginning of the item that you have to delete.

Bulleted lists are easy to reorganize in Things.

Because notes attached to tasks can be full-blown documents now, Things has also added the ability to search inside a note. On the iPhone and iPad, tap the More button and select Find in Text. On the Mac, you’ll find the option in the Edit menu, or you can use the keyboard shortcut ⌘⇧F, which also works on iOS and iPadOS. Things offers the option to Find and Replace text too. Finally, Cultured Code has improved its sync engine, making the syncing of notes more efficient and faster, which should benefit anyone who uses it to take extended notes.

I wish every developer that offered notes functionality in their app would put as much care and attention into them as Cultured Code. Few apps provide formatting, let alone what is effectively a mini Markdown text editor just for notes. It’s the sort of flexibility that sets Things apart from other task managers. I expect the new notes functionality will be perfect for anyone who has felt constrained by the typical one-liner plain text notes found in most alternatives.

Things is is sold separately for the iPhone, iPad, and Mac for $9.99, $19.99, and $49.99 respectively.

Support MacStories Directly

Club MacStories offers exclusive access to extra MacStories content, delivered every week; it’s also a way to support us directly.

Club MacStories will help you discover the best apps for your devices and get the most out of your iPhone, iPad, and Mac. Plus, it’s made in Italy.

GitHub’s Engineering Team has moved to Codespaces

My absolute dream development environment is one where I can spin up a new, working development environment in seconds - to try something new on a branch, or because I broke something and don't want to spend time figuring out how to fix it. This article from GitHub explains how they got there: from a half-day setup to a 45 minute bootstrap in a codespace, then to five minutes through shallow cloning and a nightly pre-built Docker image and finally to 10 seconds be setting up "pools of codespaces, fully cloned and bootstrapped, waiting to be connected with a developer who wants to get to work".

Electron and Rust may be cool (and also entirely the wrong things to tout in this context), but according to pretty much everyone who’s tried it the overall experience is vastly degraded and a far cry from the previous versions–to the point where the picture below is making the rounds on Twitter:

This shift away from fully native apps and the fact that they are removing iCloud support from version 8 in order to enforce the use of their cloud sync service (in an obvious lock-in ploy) was the last straw, so I just downloaded Secrets, paid for the Premium version ($19.99 for each platform) and imported all my 1Password data into it.

Full disclosure: I know the developer of Secrets, and I am glad to support him in building a fully native, well-designed and secure app.

The only thing missing for me (which I use dozens of times a day when on Windows, and is going to be a major pain to try to do without) is a watch app with TOTP tokens and some credit card info, but I hope enough people switch to Secrets to help fund that kind of development (WatchKit is a nightmare to debug, so it would be a significant time investment).

And yes, I know their company valuation is through the roof–that is just added incentive to grow a captive customer base even further.

Laundry day doesn’t always have to be a major chore. With a few simple tweaks to your routine, you can transform a tedious activity into a satisfying ritual. The best part: Once you start doing your laundry correctly, you’ll likely do a lot less of it. Here’s how to get it right.

I shared updates about my cancer treatments with my friends on Facebook and it helped to get encouragement. But something else also happened on my timeline: Facebook’s advertising algorithms began targeting me for cancer ads from scammers selling phony treatments. These companies promised that I could cure my cancer “naturally without toxic chemotherapy or surgery” using vitamin IV therapy that allegedly had “the same mechanism as chemotherapy.” … I reported the ads to Facebook in the hope the platform would remove them (it didn’t). I also wrote about it, joining the legion of voices raising the alarm about mis- and disinformation on social media.

A year later, not much has changed on Facebook. While the mega-corporation has made promises to try to contain false news about COVID, it remains a massive problem on the platform, along with fake news on cancer. Although the Center for Countering Digital Hate (CCDH) identified 12 health disinformation “superspreader” accounts on Facebook in March, 59 per cent of the content still had not been removed by July.

12 disinformation superspreaders. Twelve!

It’s very grim situation. Facebook helps scammers sell phony treatments that will eventually lead to dead people. If you ever wondered why anybody would not want to be vaccinated, you found the money trail leading to the source of disinformation.

More >

Having used 1Password since its very beginning, I grew increasingly distrustful of their product management and roadmap (the key point for me being that I will not subscribe to their cloud syncing service), so this is an attempt at putting together a systematic list of decent alternatives for my own use.

The features I need are:

• iCloud or OneDrive/Dropbox cloud sync across Mac and iOS (Windows and Linux are secondary).
• TOTP support.
• Having a subset of data (TOTP and credit card PIN codes) quickly available on the Apple Watch.

This will be turned into a proper table later, but here are the ones I’ve used and tested, as well as a few others of note:

Secrets is my current choice. It does not have a Watch app, but:

• It has great import features–it can import from multiple other password managers, and imported my 1Password vault without a hitch, including TOTP tokens, notes and other metadata.
• It does native iCloud syncing between iOS and macOS.
• The UX is very smooth (on the phone, you can navigate pretty much anywhere with just your thumb).

It also has a number of “creature comforts” like displaying passwords in large type or spelling them out, tag support, and a browser extension for Safari (as well as TouchID/FaceID and Shortcuts support, as well as other native features).

Only thing I’m missing on the Mac is getting to my favorites to copy a password/TOTP straight from the menu item, which I could do immediately on 1Password Mini.

Full disclosure: I know the Secrets developer personally (as we both worked at Portugal Telecom), and I believe the security design to be very sound.

The password management space is something that I see as being ripe for Sherlocking in various ways (at least for browser logins), so its worth keeping in mind that there are already some pretty usable options out there that do 80-90% of what most people need:

• Keychain Access (and iCloud keychain) work for browser logins and are getting TOTP in upcoming releases of iOS/macOS, but have no support for arbitrary secure fields, notes, etc. (and no Watch support for TOTP either). Interestingly enough, Apple now has a Windows app for managing passwords as well.
• Microsoft Authenticator provides mobile access to the Edge browser keychain and has TOTP support (and partial Watch support, since you can use it for MFA with Microsoft personal and corporate accounts) but also doesn’t go beyond that.

The KeePass ecosystem seems like the best long-term option (given its maturity, stability, features and cross-platform support), but clients are kludgy and lack creature comforts.

Still, they might be the best solution for those of you who need full cross-platform, “local” vaults, and a great one if you want a third-party complement to the first-party options above.

And, of course, if you need something that works in Android (which I don’t these days) or Linux (which I might need in the future), this seems like the best way to go.

• Keepassium, an Open Source, KeePass-compatible app that can use any iOS cloud provider (works OK in cursory testing with iCloud, but am not sure how reliable it will be in the long run). They also have a beta (Catalyst) Mac app that mostly works.
• MacPass, a polished macOS counterpart that seems to have some issues under BigSur (which are getting fixed).
• Strongbox also has macOS and iOS apps that can sync in various ways, but I haven’t tested them yet.
• KeePassXC is ugly as sin but great for converting your 1Password vaults, as well as installable via brew install --cask keepassxc

• PwSafe
• PasswordWallet
• QtPass, a front-end for pass
• Passbolt, an intriguing option for teams
• vaultwarden, a self-hostable sync server for Bitwarden, another little ecosystem I’m aware of but which seems too complex.

This article was originally published August 12, 2021.

For lower-stakes COVID-19 checks, rapid antigen tests offer a quick and simple way to screen for SARS-CoV-2 at home. They can be particularly helpful if you know or suspect you’ve been exposed to the coronavirus but can’t access professional testing (or wait for the results).

httpbin is a developer utility web application that receives a request and spits out the request data as a response. httpbin is for testing your API requests, webhook requests, etc.

I usually use it when developing client-side applications to watch what I am sending and what the server is receiving. For example, if I am making a simple curl POST to the remote server and if I want to know exactly what the server is receiving, then I can do

curl -X POST "https://httpbin.org/post" -H  "accept: application/json"

and it returns

{
  "args": {}, 
  "data": "", 
  "files": {}, 
  "form": {}, 
  "headers": {
    "Accept": "application/json", 
    "Host": "httpbin.org", 
    "User-Agent": "curl/7.58.0", 
    "X-Amzn-Trace-Id": "Root=1-6115338b-5a9fad6351af744531493045"
  }, 
  "json": null, 
  "origin": "49.207.207.139", 
  "url": "https://httpbin.org/post"
}

httpbin is very useful if you try to see what kind of headers or payload your application sends. You might be able to debug better. You could use the public instance at https://httpbin.org.

I know that, at least according to the code, httpbin doesn't do anything scary with your data posted. But if you want to be extra careful, you can build and run your instance.  

Recently I did install it on Google's CloudRun so that I could use it for my personal use. There are two ways to do it. You can build a Docker image using the Dockerfile provided by the project. Just one thing make sure you expose $PORT instead of 80. And start the gunicorn command on $PORT. This is required because by default CloudRun listens on port 8080 or the value defined by $PORT.

The container must listen for requests on 0.0.0.0 on the port to which requests are sent. By default, requests are sent to 8080, but you can configure Cloud Run to send requests to the port of your choice. Cloud Run injects the PORT environment variable into the container. Inside Cloud Run container instances, the value of the PORT environment variable always reflects the port to which requests are sent. It defaults to 8080.

Listening for requests on the correct port

If you plan to use the existing Docker image, then you can deploy using the command.

gcloud run deploy my-httpbin --allow-unauthenticated --platform managed --image mirror.gcr.io/kennethreitz/httpbin

This command will deploy the public image from the DockerHub. But after deployment, it fails to start. Because CloudRun is exposing only 8080. To fix this issue, log into your Google Cloud Console. Go to CloudRun and your httpbin instance. Edit the container general details. Set its value to 80. This change should start the service.

Now you have your own instance of httpbin. Of course, it's open to the world once people get to know the URL. That is because if we did --allow-unauthenticated as part of our deployment. You can play with a different type of authentication if required.

You can subscribe to my blog using RSS Feed. But if you are the person who loves getting email, then you can join my readers by signing up.

In The Premonition, author Michael Lewis describes how a mixed group of experts would participate in weekly (and sometimes daily) email exchanges/zoom calls to discuss the spread of Coronavirus and the response to it.

Over time, the group became the best source of up-to-date knowledge and expertise in pandemic planning and response.

Consider the four factors that made this work:

1. Speed. The group was created quickly to respond to an urgent need.
2. ​​Privacy. Even members didn’t know who else was there.
3. Exclusivity. There was no promotion of the group. Existing members simply invited others they felt should join.
4. Convenience. The creators used Zoom and email (tools that were most convenient for members to use).

These four factors make it easy for experts in any field to jump in and participate at scheduled times each week.

It’s sad these kinds of groups, the ones which deliver the most value, are precisely those most organisations would struggle to create.

Most organisations can’t move fast enough, select convenient tools, or keep it exclusive (often because they’re measured by engagement).

Perhaps one solution is to do this within a community. You might not be able to do the above for the community as a whole, but once you have a community up and running there’s nothing stopping you from providing all of the above to members within the community.

The post Speed, Privacy, Exclusivity, And Convenience first appeared on FeverBee.

In episode 2 I mentioned three aspects of pl/python that are reasons to use it instead of pl/pgsql: access to Python modules, metaprogramming, and introspection.

Although this episode focuses on metaprogramming — by which I mean using Python to dynamically compose and run SQL queries — my favorite example combines all three aspects.

The context for the example is an analytics dashboard with a dozen panels, each driven by a pl/plython function that’s parameterized by the id of a school or a course. So, for example, the Questions and Answers panel on the course dashboard is driven by a function, questions_and_answers_for_group(group_id), which wraps a SQL query that:

– calls another pl/python function, questions_for_group(group_id), to find notes in the group that contain question marks

– finds the replies to those notes

– builds a table that summarizes the question/answer pairs

Here’s the SQL wrapped by the questions_and_answers_for_group(group_id) function.

sql = f"""
  with questions as ( 
    select *
    from questions_for_group('{_group_id}')
    ),
  ids_and_refs as (
    select 
      id,
      unnest ("references") as ref
      from annotation
      where groupid = '{_group_id}'
  ),
  combined as (
    select
      q.*,
      array_agg(ir.id) as reply_ids
    from ids_and_refs ir
    inner join questions q on q.id = ir.ref
    group by q.id, q.url, q.title, q.questioner, q.question, q.quote
  ),
  unnested as (
    select
      c.url,
      c.title,
      c.quote,
      c.questioner,
      c.question,
      unnest(reply_ids) as reply_id
    from combined c
  )
  select distinct
    course_for_group('{_group_id}') as course,
    teacher_for_group('{_group_id}') as teacher,
    clean_url(u.url) as url,
    u.title,
    u.quote,
    u.questioner,
    (regexp_matches(u.question, '.+\?'))[1] as question,
    display_name_from_anno(u.reply_id) as answerer,
    text_from_anno(u.reply_id) as answer,
    app_host() || '/course/render_questions_and_answers/{_group_id}' as viewer
  from unnested u
  order by course, teacher, url, title, questioner, question

This isn’t yet what I mean by pl/python metaprogramming. You could as easily wrap this SQL code in a pl/pgsql function. More easily, in fact, because in pl/pgsql you could just write _group_id instead of '{_group_id}'.

To get where we’re going, let’s zoom out and look at the whole questions_and_and_answer_for_group(group_id) function.

questions_and_answers_for_group(_group_id text)
  returns setof question_and_answer_for_group as $$
  from plpython_helpers import (
    exists_group_view,
    get_caller_name,
    memoize_view_name
  )
  base_view_name = get_caller_name()
  view_name = f'{base_view_name}_{_group_id}'
  if exists_group_view(plpy, view_name):
    sql = f""" select * from {view_name} """
  else:
    sql = f"""
    <SEE ABOVE>
    """
    memoize_view_name(sql, view_name)
    sql = f""" select * from {view_name} """
  return plpy.execute(sql)
$$ language plpython3u;

This still isn’t what I mean by metaprogramming. It introduces introspection — this is a pl/python function that discovers its own name and works with an eponymous materialized view — but that’s for a later episode.

It also introduces the use of Python modules by pl/python functions. A key thing to note here is that this is an example of what I call a memoizing function. When called it looks for a materialized view that captures the results of the SQL query shown above. If yes, it only needs to use a simple SELECT to return the cached result. If no, it calls memoize_view_name to run the underlying query and cache it in a materialized view that the next call to questions_and_answers_for_group(group_id) will use in a simple SELECT. Note that memoize_view_name is a special function that isn’t defined in Postgres using CREATE FUNCTION foo() like a normal pl/python function. Instead it’s defined using def foo() in a Python module called plpython_helpers. The functions there can do things — like create materialized views — that pl/python functions can’t. More about that in another episode.

The focus in this episode is metaprogramming, which is used in this example to roll up the results of multiple calls to questions_and_answers_for_group(group_id). That happens when the group_id refers to a course that has sections. If you’re teaching the course and you’ve got students in a dozen sections, you don’t want to look at a dozen dashboards; you’d much rather see everything on the primary course dashboard.

Here’s the function that does that consolidation.

create function consolidated_questions_and_answers_for_group(group_id text)
  returns setof question_and_answer_for_group as $$
  from plpython_helpers import (
    get_caller_name,
    sql_for_consolidated_and_memoized_function_for_group
  )
  base_view_name = get_caller_name()
  sql = sql_for_consolidated_and_memoized_function_for_group(
    plpy, base_view_name, 'questions_and_answers_for_group', _group_id)
  sql += ' order by course, url, title, questioner, answerer'
  return plpy.execute(sql)
$$ language plpython3u;

This pl/python function not only memoizes its results as above, it also consolidates results for all sections of a course. The memoization happens here.

def sql_for_consolidated_and_memoized_function_for_group(plpy, 
    base_view_name, function, group_id):
  view_name = f'{base_view_name}_{group_id}'
  sql = f""" select exists_view('{view_name}') as exists """
  exists = row_zero_value_for_colname(plpy, sql, 'exists')
  if exists:
    sql = f""" select * from {view_name} """
  else:
    sql = consolidator_for_group_as_sql(plpy, group_id, function)
    memoize_view_name(sql, view_name)
    sql = f""" select * from {view_name} """
  return sql

The consolidation happens here, and this is finally what I think of as classical metaprogramming: using Python to compose SQL.

def consolidator_for_group_as_sql(plpy, _group_id, _function):
  sql = f"select type_for_group('{_group_id}') as type"
  type = row_zero_value_for_colname(plpy, sql, 'type')
  if type == 'section_group' or type == 'none':
    sql = f"select * from {_function}('{_group_id}')"
  if type == 'course_group' or type == 'course':
    sql = f"select has_sections('{_group_id}')"
    has_sections = row_zero_value_for_colname(plpy, sql, 'has_sections')
    if has_sections:
      sql = f"""
        select array_agg(group_id) as group_ids 
        from sections_for_course('{_group_id}')
      """
      group_ids = row_zero_value_for_colname(plpy, sql, 'group_ids')
      selects = [f"select * from {_function}('{_group_id}') "]
      for group_id in group_ids:
        selects.append(f"select * from {_function}('{group_id}')")
      sql = ' union '.join(selects)
    else:
      sql = f"select * from {_function}('{_group_id}')"
  return sql

If the inbound _groupid is p1mqaeep, the inbound _function is questions_and_answers_for_group, and the group has no sections, the SQL will just be select * from questions_and_answers_for_group('p1mqaeep').

If the group does have sections, then the SQL will instead look like this:

select * from questions_and_answers_for_group('p1mqaeep')
union 
select * from questions_and_answers_for_group('x7fe93ba')
union
select * from questions_and_answers_for_group('qz9a4b3d')

This is a very long-winded way of saying that pl/python is an effective way to compose and run arbitarily complex SQL code. In theory you could do the same thing using pl/pgsql, in practice it would be insane to try. I’ve entangled the example with other aspects — modules, introspection — because that’s the real situation. pl/python’s maximal power emerges from the interplay of all three aspects. That said, it’s a fantastic way to extend Postgres with user-defined functions that compose and run SQL code.

—
1 https://blog.jonudell.net/2021/07/21/a-virtuous-cycle-for-analytics/
2 https://blog.jonudell.net/2021/07/24/pl-pgsql-versus-pl-python-heres-why-im-using-both-to-write-postgres-functions/
3 https://blog.jonudell.net/2021/07/27/working-with-postgres-types/
4 https://blog.jonudell.net/2021/08/05/the-tao-of-unicode-sparklines/
5 https://blog.jonudell.net/2021/08/13/pl-python-metaprogramming/
6 https://blog.jonudell.net/2021/08/15/postgres-and-json-finding-document-hotspots-part-1/
7 https://blog.jonudell.net/2021/08/19/postgres-set-returning-functions-that-self-memoize-as-materialized-views/
8 https://blog.jonudell.net/2021/08/21/postgres-functional-style/
9 https://blog.jonudell.net/2021/08/26/working-in-a-hybrid-metabase-postgres-code-base/
10 https://blog.jonudell.net/2021/08/28/working-with-interdependent-postgres-functions-and-materialized-views/
11 https://blog.jonudell.net/2021/09/05/metabase-as-a-lightweight-app-server/
12 https://blog.jonudell.net/2021/09/07/the-postgres-repl/

The newest stable release of Thunderbird, version 91, is available for download on our website now. Existing Thunderbird users will be updated to the newest version in the coming weeks.

Thunderbird 91 is our biggest release in years with a ton of new features, bug fixes and polish across the app. This past year had its challenges for the Thunderbird team, our community and our users. But in the midst of a global pandemic, the important role that email plays in our lives became even more obvious. Our team was blown away by the support we received in terms of donations and open source contributions and we extend a big thanks to everyone who helped out Thunderbird in the lead up to this release.

There are a ton of changes in the new Thunderbird, you can see them all in the release notes. In this post we’ll focus on the most notable and visible ones.

Multi-Process Support (Faster Thunderbird)

Thunderbird has gotten faster with multi-process support. The new multi-process Thunderbird takes better advantage of the processor in your computer by splitting up the application into multiple smaller processes instead of running as one large one. That’s a lot of geekspeak to say that Thunderbird 91 will feel like it got a speed boost.

New Account Setup

One of the most noticeable changes for Thunderbird 91 is the new account setup wizard. The new wizard not only features a better look, but does auto-discovery of calendars and address books and allows most users to set them up with just a click. After setting up an account, the wizard also points users at additional (optional) things to do – such as adding a signature or setting up end-to-end encryption.

The New Account Setup Wizard

Attachments Pane + Drag-and-Drop Overlay

The attachments pane has been moved to the bottom of the compose window for better visibility of filenames as well as being able to see many at once. We’ve also added an overlay that appears when you drag-and-drop a file into the compose window asking how you would like to handle the file in that email (such as putting a picture in-line in your message or simply attaching it to the email).

Compose window with bottom attachment pane.

The new attachment drag-and-drop overlay.

PDF Viewer

Thunderbird now has a built-in PDF viewer, which means you can read and even do some editing on PDFs sent to you as attachments. You can do all this without ever leaving Thunderbird, allowing you to return to your inbox without missing a beat.

The PDF Viewer in Thunderbird 91

UI Density Control

Depending on how you use Thunderbird and whether you are using it on a large desktop monitor or a small laptop touchscreen, you may want the icons and text of the interface to be larger and more spread out or very compact. In Thunderbird 91 under the View -> Density in the menu, you can select the UI density for the entire application. Three options are available: compact – which puts everything closer together, normal – the experience you are accustomed to in Thunderbird, and touch – that makes icons bigger and separates elements.

Play around with this new level of control and find what works best for you!

UI density control options

Calendar Sidebar Improvements

Managing multiple calendars has been made easier with the calendar sidebar improvements in this release. There is a quick enable button for disabled calendars, as well as a show/hide icon for easily toggling what calendars are visible. There is also a lock indicator for read-only calendars. Additionally, although not a sidebar improvement, there are now better color accents to highlight the current day in the calendar.

Improved Calendar sidebar

Better Dark Theme

Thunderbird’s Dark Theme got even better in this release. In the past some windows and dialogues looked a bit out of place if you had Thunderbird’s dark theme selected. Now almost every dialogue and window in Thunderbird is fully styled to respect the user’s color scheme preferences.

Dark Theme

Other Notable Mentions

You really have to scroll through the release notes as there are a lot of little changes that make Thunderbird 91 feel really polished. Some other notable mentions are:

• Native support for macOS devices built with Apple Silicon CPU
• CardDAV Address Book Support (and CardDAV auto-detection support)
• Printing UI updated
• Ability to change order of accounts in UI
• Support opening of .ics files by double-click in calendar

A Bright Future

The new Thunderbird is a release that the team and community are proud of. Much of what was improved in this newest version was built upon the foundation we’ve been laying for the last three years. The next release will feature a new address book, further end-to-end encrypted email improvements, better Import/Export tools, a ton of calendar improvements and lots of amazing stuff that we’re not ready to talk about yet.

If you love Thunderbird and you want to see your favorite Email client get even better, we are supported by donations from our users and volunteer contributions from our community. Please consider helping us out with your expertise or with a donation if you appreciate the work that we are doing.

Click Here to Donate to Thunderbird

Click Here to Get Involved in Thunderbird’s Development

The post Thunderbird 91 Available Now appeared first on The Thunderbird Blog.

Imagine an Internet of Snitches, each scanning whatever data they have access to for evidence of crime. Beyond the OS itself, individual phone apps could start looking for contraband. Personal computers would follow their lead. Home network file servers could pore through photos, videos and file backups for CSAM and maybe even evidence of copyright […]

The post Internet of Snitches appeared first on Purism.

I can't remember where I heard about Houses Slide but I'm fascinated by it.

It's a work of contemporary music/art by Laura Bowler which "describes one woman's intimate psychological journey to figure out her response to the climate crisis, from an initial depressing realisation of the gravity of the issue, through to her refusal to be overwhelmed and decision to take positive action."

And "With a text created by Cordelia Lynn using submissions from members of the public, this concert presentation conceived and directed by Katie Mitchell is an industry-first, being powered entirely by bicycles."

That's the bit that really struck me.

'A concert presentation powered entirely by bicycles.'

On the stage apparently. And at the back. Obviously, it's slightly gimmicky and perhaps obvious, the first thing you'd think of. But there's something powerful and clear about it.

Increasingly, we need to think about power and work and energy and seeing it embodied in some cyclists is useful. It takes a bunch of people, cycling for quite a while, to provide sound and light for a concert. (A review and pictures) We need to understand the scale of that kind of thing.

The number of drive in movie theatres on PEI has doubled with the arrival of The Runway Drive In in Summerside.

It is, quite literally, almost right on the runway at Slemon Park (née CFB Summerside). The programming appears to be a mixture of films new and old: this weekend it will be Footloose, Grease, and The Secrets We Keep.

I still haven’t quite gotten over the closure of the Princess Pat Drive In in Cascumpec, and it seemed unlikely that the Island would see another in the west. I hope they do well.