W3C

– DRAFT –
I18N ⇔ CSS

21 November 2023

Attendees

Present
atsushi, florian, r12a, xfq
Regrets
-
Chair
-
Scribe
xfq

Meeting minutes

Agenda Review

florian: the CSSWG need to come up with a structure
… a list of things that need to be described for each one we might want to add
… the conclusion from the last meeting was that fantasai and I had to prepare a list
… maybe the font name, maybe a sample of the fonts, maybe a description of what it's for
… and maybe a description of why it matters

<r12a> https://github.com/w3c/i18n-discuss/wiki/Generic-font-families

r12a: Chris has started to implement bits of it in the css-fonts spec
… in Geek Week I made a list of generic font categories
… see above
… including Windows, macOS, Noto, and SIL fonts
… you've got a list of what the categories would be
… you've got a description of why you'd need it
… and you've got links to the actual fonts

florian: is it complete for now, even though we might add some more in the future?

r12a: it's an initial list based on an exhaustive survey of the pre-installed fonts on macOS and Windows
… somebody responded on the thread on the CSS thread that there may be more

florian: those are not necessarily available through pre-installed or commonly available fonts

r12a: this list is for discussion
… what's also useful about it is it's not very long

florian: some are very specific to some writing systems
… loopless is meant for Thai, but it @@

xfq: it only lists Fangsong for SC
… but there are also TC and Japanese Fangsong fonts

r12a: and heavy right down at the bottom

florian: is that different from bold?

r12a: these are much heavier than bold
… it seemed to be a common thing that many scripts would have extremely heavy font faces
… the same goes for slanted
… we had Hebrew in there and I took it out

r12a: Syriac @@
… it is to persuade people that there needs to be an open interface so that people @@
… in the case of Syriac, we have three fonts
… I think Noto needs two fonts for Lanna

r12a: there are 2 scenarios
… one is where you're using the font to style the text for things like headings
… it's very common
… and the other scenario is like nastaliq

florian: I am convinced that those need supporting
… I'm not convinced that those needs supporting through generic font families
… because language is also a factor in selecting fonts
… in terms of UI to pick up fonts, absolutely
… that's also useful for systems which are less clear-cut than Windows or Mac about what's a system font and what's not
… like Debian GNU/Linux
… is there within that list anything which like nastaliq?

r12a: syriac and lanna
… it's not "what you can do", it's actually "what you must not do"
… this is the key factor
… you must not fall back to this wrong font

florian: I will certainly take some time offline after this meeting to read through the list and organise my thoughts

<r12a> https://typo.social/@ri/111368995519804048

florian: but probably I'm not the only person who should be caring about this

https://www.w3.org/2023/11/17-clreq-minutes.html#t01

xfq: Fangsong fonts can also have Latin glyphs, which might contain Latin letters of similar style to the Chinese Fangsong font

florian: the motivation for having those is script specific, but there is no restriction on the unicode range covered, so if it contains glyphs for latin (for example), that'll be used too

florian: at least I'll take an action to review it
… but I suspect that's not enough

<r12a> this is a mechanism to select fonts - it's not influenced by unicode character ranges - so latin characters in a fangsong font is not an issue

florian: I expect my availability to improve starting from next week

r12a: working on ruby!

florian: yes, I hope to

Ruby

florian: a reasonable next topic I think is the question of how it is supposed to be read
… Muraka-san has written ruby-t2s-req
… it currently only mentions Japanese
… does it apply to Chinese and Zaima?

r12a: every other week somebody pops up and says we could use ruby for this sort of thing

florian: I'm thinking about the next steps
… not to do them, but to discuss them a little bit

r12a: do you think it would be an HTML attribute?
… I would have thought it would be a CSS property

florian: it's a semantic difference
… sometimes it's phonetic, sometimes it's not
… that's not for immediate solution
… but it should be somewhere on our radar
… I wouldn't want to miss it

w3c/ruby-t2s-req#42

xfq: I have two comments
… @@1
… Zaima PR ^
… Warichu tts, copy/paste, string-searching etc.

Minutes manually created (not a transcript), formatted by scribe.perl version 221 (Fri Jul 21 14:01:30 2023 UTC).

Diagnostics

Succeeded: s/which contain/which might contain

Succeeded: s/@@/the motivation for having those is script specific, but there is no restriction on the unicode range covered, so if it contains glyphs for latin (for example), that'll be used to

Succeeded: s/used to/used too

All speakers: florian, r12a, xfq

Active on IRC: r12a, xfq