Khaled Hosny

@nprofile…wguy @nprofile…uvqd RTL bottons usually control paragraph (sometimes called base) direction which influences neutral or characters with weak directionality. Characters with strong directionality can be forced to take a different direction using directional overrides https://www.unicode.org/reports/tr9/#Explicit_Directional_Overrides

2026-04-02 03:03:05 UTC

- reply

@nprofile…05la @nprofile…uvqd It is not something fonts can control, but it is something fonts can support. Unicode has control characters to force text to be RTL or LTR despite its default direction. Fonts can include ltrm/ltra or rtlm/rtla features to substitute glyphs when text direction is LTR or RTL. The input text will need to use these characters (or an equivalent solution, like HTML’s <bdo> element).

2026-01-29 18:25:33 UTC

- reply

@nprofile…8clk It has “public.unicodeVariationSequences” lib key, see https://unifiedfontobject.org/versions/ufo3/lib.plist/#publicunicodevariationsequences

2026-01-14 19:05:34 UTC

- reply

@nprofile…8clk @nprofile…zk9d That is how emoji presentation is implemented in most text stacks. If a symbol has emoji presentation (either by default or by the use of emoji presentation selector) then a color font that supports this symbol (or a font from a hard-coded list of color emoji fonts) will be used. The font does not need to handle the emoji presentation selector themselves (in fact almost all color emoji fonts have no special handling for it).

2025-12-14 04:46:05 UTC

- reply

@nprofile…ukmn that what I would expect from an application that does not directly support VF. It got the metrics from DirectWrite which makes named instances look like static fonts to the application, but when embedded font in PDF it embeds the glyf table itself and it doesn't do sny instantiation, so you get the outlines of default instance.

2025-12-14 04:43:23 UTC

- reply

@nprofile…ukmn If I were to guess, PDF export uses application-specific code which in this case is not VF-aware, while printing uses system facilities that are. There are tools to extract fonts from PDF to examine them (mutool command line utility from MuPDF has exatract subcommand). @nprofile…n60g @nprofile…qppg

2025-12-14 04:32:02 UTC

- reply

@nprofile…ukmn PDF knows nothing about variable fonts. The PDF producer has to instantiate the font when embedding it in PDF files.

2025-08-23 10:11:39 UTC

- reply

@nprofile…mjt5 Seems to be foundry type, so my guess was half correct.

2025-08-20 05:39:02 UTC

- reply

@nprofile…mjt5 from the call gor proposals page: “Speakers are responsible for their own conference registration, travel, and accommodation expenses.”

2025-08-19 23:57:04 UTC

I considered submitting a talk for ATypI in Dubai since I’d at least not need to go through the humiliating experience of applying for a visa to the places it is typical held in.
But having speakers pay for their (already expensive) tickets is borderline insulting, so no ATypI for me.

2025-07-24 09:07:17 UTC

- reply

@nprofile…mjt5 Looks like the kind of typeface cut in Europe, but I don’t immediately recognize it (not that I’m very familiar with European Arabic typefaces).
That rounded ر (and و to lesser extent), look peculiar though. Done to avoid kerning perhaps? Might be a clue that it is not foundry type.

2025-06-04 20:15:11 UTC

- reply

@nprofile…mjt5 I can’t really tell if it is a TeX code issue or a font issue without seeing both.

2025-06-04 18:42:51 UTC

- reply

@nprofile…mjt5 Depends on how the symbol is used. If the integral is used as a big operator then it would be the advance width of the integral. If it is used as a regular symbol then it would be like any another symbol (advance width + italic correction + math cut ins). This is determined by the macro (TeX command) that the symbol.

2025-06-02 21:24:01 UTC

- reply

@nprofile…gdrs FontForge has complete support for editing MATH table. The UI is clumsy at places, but no more clumsier than FontForge UI in general. I used it to build several math fonts, and I think a few open source font projects still use it. @nprofile…mjt5

2025-06-02 17:15:12 UTC

- reply

@nprofile…mjt5 It might depend on how the symbol is used (it might be explicitly asking to not stretch the anchor), but the engine has no restrictions in this area AFAIK @nprofile…gdrs

2025-05-31 16:45:45 UTC

- reply

@nprofile…mjt5 No worries. The original cut-ins implementation was indeed buggy (partly my fault, and partly the spec’s). David M. Jones did a great job fixing it, including inferring the actual intent of the spec. @nprofile…gdrs

2025-05-31 14:40:36 UTC

- reply

@nprofile…mjt5 I wrote some of the MATH support in XeTeX when I was the maintainer, but most of the MATH support (like the rest of XeTeX) was written by Jonathan Kew. Some of the recent fixes are by David M. Jones. @nprofile…gdrs

2025-05-30 13:16:02 UTC

- reply

@nprofile…mjt5 as for the subscript j moving down, it might be caused by subscriptTopMax or subscriptBaselineDropMin constants (https://github.com/notofonts/math/blob/main/documentation/building-math-fonts/index.md#subscriptshiftdown-subscripttopmax-subscriptbaselinedropmin), you can confirm by observing if other glyphs with ascenders move down too @nprofile…gdrs

2025-05-30 13:04:36 UTC

- reply

@nprofile…mjt5 I’m afraid neither LuaTeX, XeTeX, or Word apply math cut-ins for large operators. @nprofile…gdrs

2025-02-20 00:27:24 UTC

عملي المشارك في الدورة التاسعة من ملتقى القاهرة لفن الخط العربي، في أكتوبر الماضي.
سورة الإخلاص والمعوذتين بالخط الكوفي المصحفي. العمل مقسم إلي أربع صفحات تحاكي تنسيق المصاحف المبكرة المكتوبة بالخط الكوفي، وكتبت أسماء السور باللون الذهبي كما هي العادة في تلك المصاحف.
العمل مصمم رقميا ومطبوع، ويستخدم الخط الحاسوبي «رق» من تصميمي:
https://aliftype.com/raqq
https://typo.social/system/media_attachments/files/114/033/370/518/061/413/original/2c4e16f695fe5d8c.jpg

2023-10-04 21:30:57 UTC

- reply

@npub1xqv…230r Some one is angry about íj́ again?

2023-09-12 05:02:39 UTC

أعمل حاليا على دعم المعادلات الرياضياتية العربية في ليبرأوفيس، النتائج مبشرة حتى الآن.
#LibreOffice #Arabic #math
https://typo.social/system/media_attachments/files/111/050/418/894/116/694/original/58576d9fb1dc1526.png