@mwichary The segmented displays on the Philadelphia L line are different from all of these. Would you like me to go take photos?

https://aresluna.org/segmented-type/

This character is amazing. Here's the glyph for it on my Android phone. I spent a fun fifteen or thirty minutes figuring out how it was composed. “Bismillah”, it transpires, is a contraction of “bismi Allah” (“in the name of God”), and in this one the "bismi" starts at farthest right and runs all the way through the tall part of the "ar-" in "ar-raḥman", with the ‘m’ part over on the left, and the mim's hanging final tail joining up with the ‘n’ of raḥman.

The "Allah" part is atop that, in the middle.

The Egyptian multiplication algorithm, as I understand it, multiplies 9 × 1/99 and gets a result of 1/22 + 1/33 + 1/99 + 1/198. This is numerically equal to 1/11, but it's not clear to me how the Egyptians would have recognized this.

Does anyone happen to know anything about this? I understand how Egyptian fractions can be added and even multiplied, but not how they can be compared for equality or to see which is less or greater.

#egyptianFractions #arithmetic #algorithms

As I recall,, in the TeX hyphenation algorithm worked out by Knuth and one of his students, they didn't simply store the preferred hyphenation for every word in the dictionary. Rather, to save memory, they stored a list of letter clusters and the relative badness of a hyphen placed in each position in each cluster. (There was also a very short list of exceptions words.)

Does TeX still do this, now that it's quite practical to simply have an explicit table of the way to hyphenate every word?

This article on Arabic script typesetting and justification is turning out even better than I had hoped.

“My favourite resident of the block, and one of my favourite characters in all of Unicode, is U+FDFD, ﷽ : four-word invocation, bismillāh ar-raḥmān ar-raḥīm, as a single codepoint. A monument from the era when rendering was baked into the encoding because nobody trusted the renderer to do anything, preserved forever, like a fly in amber that recites.”

https://lr0.org/blog/p/arabic/

I generally understand “evil” to mean when someone pursues their own goals heedless of any loss or suffering it brings to others.

I think this is at least an excellent first approximation. Most people committing evil acts aren't sadistic. They are just pursuing their own goals, and choosing to ignore or disregard the costs to others.

On this definition, large corporations are evil *by design*. They are constructed specifically to pursue their own corporate goals, heedless of any loss or suffering it brings to others.

Smaller corporations can be to some extent restrained in this evil by the actions of their employees. Larger ones cannot. The larger a corporation is, the eviller.

In the 1990s programmers used to get a lot of repetitive strain injuries like carpal tunnel syndrome. I don't hear much about that any more. Is that because people are more generally aware of the risks and are more care not to develop these sorts of injuries?

“The reply took and the closure of the ticket took half an hour or so. The reasons behind it took five hundred years to pile up, and they involve a twice-mutilated vizier, a Qurʾān that vanished for four centuries, a Beirut newspaperman with a deadline, and an Egyptian physician who taught himself font engineering for fun (or that what I imagine about him). Walking through these, ended up to be the most enjoyable couple of weeks in that job, and I want to go through it here too.”

I haven't read this yet, but how could I not?

https://lr0.org/blog/p/arabic/

RE: https://mastodon.online/@mwichary/116715106443531447

This is really clever. It's a web font with ligature rules defined so that you can type plain text ` {l:10,50,30,80,20}` and get a little inline sparkline, or `{b:10,50,30,80,20}` and get a little inline bar chart.

https://franktisellano.github.io/datatype/

Of course the converse is that if you're delegating your authentication to Google, then Google can authenticate anyone they feel like, whether or not that person is you.

Does it sometimes happen to people that they have been using Google Auth everywhere, and then one day Google revokes their account and refuses to vouch for them, and they're locked out of everything?

I'm not just being paranoid, am I?

What if instead of “1 failed, 12 passed” it said “1 problem detected, 12 tests failed to detect anything”?

This reminds me of how novice programmers often have to work hard to learn that compiler error messages are not something to be ashamed of, not something to suppress, they are something to engage with.

I wonder if the terminology “test failure” is putting us in the wrong mindset. It's not a failure, it's a victory! A problem was caught early, before it did any damage and when it was still cheap to fix. The time and effort spent writing the test is being paid back!

Instead of "successful test" we have the value-neutral "passing test". What would be a similarly value-neutral term for a test that doesn't pass?

Second best part of my programming routine: When I run the test suite and it all passes.

Best part: When it fails.

This reminds me a bit of how there were several very good arguments against heliocentrism. Here's one: if the earth is moving relative to the stars, we ought to see the stars move in the sky over the course of the year. But we don't.

Copernicus might reply “well, maybe they're so far away that the motion is undetectable”. But you can do the calculation and find out that for that to be true, the stars would have to be ridiculously far away, like tens of thousands of times as far away as the sun. It seems implausible. And ad-hoc.

Stellar parallax wasn't observed until the 1830s.

I _am_ old enough, barely, to remember the end of the "structured programming vs GOTO" wars.

Most people don't remember how serious and mainstream the GOTO people were. They had some good arguments, lots of “okay, but”s.

In fact, it wasn't even until 1964 that it wasn proved that it was even possible to do without GOTO. It was a big step forward when it was shown formally that ‘if’ and ‘while’ were sufficient to compute anything that unstructured programs could.

(That's https://en.wikipedia.org/wiki/Structured_program_theorem )

Then of course the arguments changed to “well, maybe it's _possible_, but it's going to be slower or clumsier in many situations”. Which is true!

The GOTO arguments hung on for a long time, well into the 1980s. But today, the GOTO people are dead and buried.

(Related, here's my all-time favorite computer science paper, Donald Knuth's “Structured Programming with GO TO Statements”. https://pic.plover.com/knuth-GOTO.pdf )

I am not _quite_ old enough to remember the assembler-vs-high-level-language holy wars of the past. I picked up hints of them from reading computer magazines in the early 1980s but they were long over by the time I arrived on the scene.

But the “Story of Mel, a Real Programmer”, is out of that time:

“‘You never know where it's going to put things’,
he explained, ‘so you'd have to use separate constants’.”

“It was a long time before I understood that remark.”

(If you haven't read the Story of Mel, you should. It is an important piece of programmer ur-culture.)

https://users.cs.utah.edu/~elb/folklore/mel.html

“[Richard] Hamming writes in the _art and science of engineering_ about the programmers of his past, confident wizards of assembly, who refused out of stubborn pride to adopt the compiler. they could do the work better themselves, they asserted. and they were right. in the days before modern optimizing compilers, it was easy for a capable programmer to produce assembly of substantially better quality than whatever the machine would generate.”

(Alice Maz)

https://substack.com/@alicemaz/p-135931706

This is what I think of when people ask if I'm worried about “deskilling”. I was, long ago, a competent writer of passable System/370 assembly language, and a little less long ago I was an unusually skilled writer of C. I do not regret the loss of those skills.

The real skills, the ones worth holding to, are to be able to evaluate tools, choose the right tools at the right time, and then use the tools skillfully. That has never changed.

I was fascinated by this note from Alice Maz about how she tried to build a game of court intrigue where an AI simulates the personalities and behaviors of the various courtiers. I'd been hoping that good-quality LLMs would enable totally new genres of games, and I was excited to find out more about this one.

But, just as fascinating, she said it didn't work. “the personality and memory can give them varied flavor and make them coherent over 10+ compactions with a fairly humanlike specific near-term and general long-term awareness of everything that happened. but they have no spark. they don’t do anything interesting.”

https://substack.com/@alicemaz/note/c-265000587

There are Wikipedia articles written about Ramon Llull's book _Ars Generalis Ultima_ in German, Spanish, Catalan, French, Italian, Basque (!), Italian, Polish, and Slovak.

But not English.

Isn't that odd?

https://de.wikipedia.org/wiki/Ars_generalis_ultima