Daring FireballModify

2025-12-13 02:37:47

Youki Terada, writing for Edutopia in 2022 (via Jens Kutílek):

Under close scrutiny, the evidence for dyslexia-friendly fonts falls apart. In a 2017 study, for example, researchers tested whether OpenDyslexic, a popular font with thicker lines near the bottom of the letters, could improve the reading rate and accuracy for young children with dyslexia. According to the developers of the font, which is open-source and free of charge, the “heaviness” of the letters prevented them from turning upside down for readers with dyslexia, which they claimed would improve reading accuracy and speed.

Researchers put the font to the test, comparing it with two other popular fonts designed for legibility — Arial and Times New Roman — and discovered that the purportedly dyslexia-friendly font actually reduced reading speed and accuracy. In addition, none of the students preferred to read material in OpenDyslexic, a surprising rebuke for a font specifically designed for the task.

In a separate 2018 study, researchers compared another popular dyslexia font — Dyslexie, which charges a fee for usage — with Arial and Times New Roman and found no benefit to reading accuracy and speed. As with the previous dyslexia font, children expressed a preference for the mainstream fonts. “All in all, the font Dyslexie, developed to facilitate the reading of dyslexic people, does not have the desired effect,” the researchers concluded. “Children with dyslexia do not read better when text is printed in the font Dyslexie than when text is printed in Arial or Times New Roman.”

Quoting from the abstract of the first study cited above:

Results from this alternating treatment experiment show no improvement in reading rate or accuracy for individual students with dyslexia, as well as the group as a whole. While some students commented that [OpenDyslexic] was “new” or “different”, none of the participants reported preferring to read material presented in that font. These results indicate there may be no benefit for translating print materials to this font.

The problem isn’t dyslexia if you don’t notice that OpenDyslexic is “different”.

Quoting from the second cited study:

Words written in Dyslexie font were not read faster or more accurately. Moreover, participants showed a preference for the fonts Arial and Times New Roman rather than Dyslexie, and again, preference was not related to reading performance. These experiments clearly justify the conclusion that the Dyslexie font neither benefits nor impedes the reading process of children with and without dyslexia.

OpenDyslexic’s website has a “related research” page but of the four articles they link to, three are 404s, and the other one only studied “extra-large letter spacing”. I chased down the correct link to one of the other articles they cite, and the only fonts it studied were Verdana, Arial, Georgia, and Times.

Some people claim to prefer reading text set with OpenDyslexic. Some people like Comic Sans, too. But I was unaware that these typefaces that purport to be designed specifically to benefit people with dyslexia are based on misguided beliefs that dyslexia is a visual problem, and that actual research shows they do not provide the benefits they claim to. They’re just ugly fonts.

2025-12-13 01:29:02

Small follow-up point re: my post this week on iMessage’s delivery architecture being built atop the Apple Push Notification service:

APNs can only relay messages up to 4 or 16 KB in size, depending on the iOS or iPadOS version. If the message text is too long or if an attachment such as a photo is included, the attachment is encrypted using AES in CTR mode with a randomly generated 256-bit key and uploaded to iCloud. The AES key for the attachment, its Uniform Resource Identifier (URI), and an SHA-1 hash of its encrypted form are then sent to the recipient as the contents of an iMessage, with their confidentiality and integrity protected through normal iMessage encryption, as shown in the following diagram.

This explains why you can often text, but not send or receive images, with iMessage over in-flight Wi-Fi. (Thanks to Adam Shostack for flagging this detail.)

2025-12-12 23:53:32

OpenAI:

In ChatGPT, GPT‑5.2 Instant, Thinking, and Pro will begin rolling out today, starting with paid plans. In the API, they are available now to all developers.

Overall, GPT‑5.2 brings significant improvements in general intelligence, long-context understanding, agentic tool-calling, and vision — making it better at executing complex, real-world tasks end-to-end than any previous model.

5.1 was released just one month ago, but 5.2 delivers a slew of measurable improvements across the board. Where 5.1 was seemingly more about the feel of responses, the personality, 5.2 was clearly focused on tangible and benchmarkable gains.

2025-12-11 10:25:42

Before anyone starts patting the Trump administration on its back for one good typographic decision, take a gander at the hard-to-believe-this-is-real new signage at (and alas, on) the White House. This is the sort of signage that typically spells “Business Center” across from the check-in desk at a Courtyard Marriott. The Biden State Department replacing Times New Roman with Calibri was a typographic misdemeanor. Festooning the White House with signage set in gold-plated Shelley Script ought to land Trump in The Hague.

(The idea that the Oval Office ought to be explicitly labeled “The Oval Office” — whatever the typeface or signage style — brings to mind this classic Far Side cartoon, which I think aptly illustrates the president’s mental faculties.)

2025-12-11 10:01:56

The fifth of five rules in Matthew Butterick’s “Typography in Ten Minutes”:

And finally, font choice. The fastest, easiest, and most visible improvement you can make to your typography is to ignore the fonts already loaded on your computer (known as system fonts) and the free fonts that inundate the internet. Instead, buy a professional font (like those found in font recommendations). A professional font gives you the benefit of a professional designer’s skills without having to hire one.

If that’s impossible, you can still make good typography with system fonts. But choose wisely. And never choose Times New Roman or Arial, as those fonts are favored only by the apathetic and sloppy. Not by typographers. Not by you.

(Via Antonio Cavedoni.)

2025-12-11 07:53:51

I’m a big believer in reading original source material. For example, when Apple provided me, alongside only a handful of other outlets, with a statement regarding their decision to delay the “more personalized Siri” back in March, I ran the full statement, verbatim. I added my own commentary, but I wanted to let Apple’s own statement speak for itself first. It drives me nuts when news sites in possession of a statement or original document do not make the full original text available, even if only in a link at the bottom, and choose only to quote short excerpts.

With regard to today’s news regarding Marco Rubio’s directive re-establishing Times New Roman as the default font for U.S. State Department documents (rescinding the Biden administration’s 2023 change to Calibri), I very much wanted to read the original. The New York Times broke the news, stated that they had obtained the memo, and quoted phrases and words from it, but they did not provide a copy of the original.

The State Department has not made this document publicly available, and to my knowledge, no one else has published it. I have obtained a copy from a source, and have made it available here in plain text format. The only change I’ve made is to replace non-breaking spaces (U+00A0) with regular spaces.¹

Please do read it yourself, and do so with an open mind.

It seems clear to me that The New York Times did Rubio dirty in their characterization of the directive. The Times story, credited to reporters Michael Crowley and Hamed Aleaziz, ran under the headline “At State Dept., a Typeface Falls Victim in the War Against Woke”, and opens thus:

Secretary of State Marco Rubio waded into the surprisingly fraught politics of typefaces on Tuesday with an order halting the State Department’s official use of Calibri, reversing a 2023 Biden-era directive that Mr. Rubio called a “wasteful” sop to diversity.

While mostly framed as a matter of clarity and formality in presentation, Mr. Rubio’s directive to all diplomatic posts around the world blamed “radical” diversity, equity, inclusion and accessibility programs for what he said was a misguided and ineffective switch from the serif typeface Times New Roman to sans serif Calibri in official department paperwork.

Rubio’s memo ran about 950 words. Here are the full quotes the Times pulled from it, consisting of just 56 words, aside from the memo’s subject line (“Return to Tradition: Times New Roman 14-Point Font Required for All Department Paper”):

“wasteful”

“radical”

“restore decorum and professionalism to the department’s written work.”

“informal”

“clashes”

“was not among the department’s most illegal, immoral, radical or wasteful instances of D.E.I.A.”

“accessibility-based document remediation cases”

“Switching to Calibri achieved nothing except the degradation of the department’s official correspondence.”

“generally perceived to connote tradition, formality and ceremony”

Rubio’s memo wasn’t merely “mostly framed as a matter of clarity and formality in presentation”. That’s entirely what the memo is about. Serif typefaces like Times New Roman are more formal. It was the Biden administration and then-Secretary of State Antony Blinken who categorized the 2023 change to Calibri as driven by accessibility. I do not have access to Blinken’s memo making that change (under the cringe-inducing subject line “The Times (New Roman) are a-Changin”), but it was first reported by John Hudson and Annabelle Timsit at The Washington Post, where they wrote:

The secretary’s decision was motivated by accessibility issues and not aesthetics, said a senior State Department official familiar with the change.

Rubio’s memo makes the argument — correctly — that aesthetics matter, and that the argument that Calibri was in any way more accessible than Times New Roman was bogus. Rubio’s memo does not lash out against accessibility as a concern or goal. He simply makes the argument that Blinken’s order mandating Calibri in the name of accessibility was an empty gesture. Purely performative, at the cost of aesthetics. Going back to that 2023 story at the Post, they quote from Blinken’s memo thus:

In its cable, the State Department said it was choosing to shift to 14-point Calibri font because serif fonts like Times New Roman “can introduce accessibility issues for individuals with disabilities who use Optical Character Recognition technology or screen readers. It can also cause visual recognition issues for individuals with learning disabilities,” it said.

The bit here about OCR is utter nonsense, a voodoo belief. No OCR or screen-reader software in use today has any problem whatsoever with Times New Roman. That’s just made-up nonsense, and I’d like to see sources for the claim about “visual recognition issues for individuals with learning disabilities”. I don’t think it’s true, and citing it alongside a provably wrong claim about OCR software makes me even more skeptical.

Rubio brings actual numbers to make his case, which is more than can be said for anyone I’ve found arguing that Calibri is somehow more accessible than Times New Roman. Rubio’s argument is alluded to in the Times’s article thus:

But Mr. Rubio called it a failure by its own standards, saying that “accessibility-based document remediation cases” at the department had not declined.

Here’s the full passage from Rubio’s memo:

And although switching to Calibri was not among the Department’s most illegal, immoral, radical, or wasteful instances of DEIA (see, e.g., Executive Orders 14151, 14173, 14281, and Memorandum on Removing Discrimination and Discriminatory Equity Ideology From the Foreign Service (DCPD202500375)) it was nonetheless cosmetic: the switch was promised to mitigate “accessibility issues for individuals with disabilities,” and employees were promised, “Your adoption supports the Department’s commitment to create a more accessible workplace,” but these promises were false. In fact, the number of accessibility-based document remediation cases at the Department of State was the same in the year after adopting Calibri as in the year before (1,192 cases in FY2024 versus 1,193 cases in FY2022). And the costs of remediation actually increased by $145,000 in that period — nearly a 20% jump. Switching to Calibri achieved nothing except the degradation of the Department’s official correspondence.

2024 was a Biden year, not a Trump year, so there’s no reason to think the remediation numbers were counted differently. The change to Calibri was the worst kind of accessibility effort: one that was founded on nothing more than feel-good performance. It was a change everyone could see and notice, but one that had no practical benefit whatsoever. Good on Rubio for rescinding a bad decision, and even better for doing so with a fair and informative explanation.² (His memo even explains, “Fonts are specific variations of a typeface.... Through common use, the word font has come to mean both typeface and font.”)