- From: Fuqiao Xue <xfq@w3.org>
- Date: Fri, 18 Apr 2025 11:04:46 +0800
- To: www-international@w3.org
https://www.w3.org/2025/04/17-i18n-minutes.html
text version:
– DRAFT –
Internationalization Working Group Teleconference
17 April 2025
[2]Agenda. [3]IRC log.
[2] https://www.w3.org/events/meetings/b7edae68-f52c-4aab-a1a6-3c37459e0786/20250417T150000/
[3] https://www.w3.org/2025/04/17-i18n-irc
Attendees
Present
Addison, atsushi, Bert, Fuqiao, Richard
Regrets
JcK
Chair
Addison Phillips
Scribe
Bert
Contents
1. [4]Agenda Review
2. [5]Action Items
3. [6]Info Share
4. [7]Infoshare
5. [8]Review the Review RADAR
6. [9]Pending Issue Review
7. [10]WebNN review
8. [11]Google Local Fonts Proposal
9. [12]Encoding: relationship of windows-1252 and Latin1
10. [13]i18n-activity 2002: language detection and und
11. [14]AOB?
12. [15]Summary of action items
13. [16]Summary of resolutions
Meeting minutes
Agenda Review
Action Items
<addison> [17]https://github.com/w3c/i18n-actions/issues
[17] https://github.com/w3c/i18n-actions/issues
<addison> #166
<gb> [18]Action 166 reply to pointerevents 505 (on aphillips)
due 2025-04-17
[18] https://github.com/w3c/i18n-actions/issues/166
addison: I think I did that.
<addison> close #166
<gb> Closed [19]issue #166
[19] https://github.com/w3c/i18n-actions/issues/166
<addison> #165
<gb> [20]Action 165 add a conformance section to suppress the
respec warning to specdev (on aphillips) due 2025-04-10
[20] https://github.com/w3c/i18n-actions/issues/165
<addison> #164
<gb> [21]Action 164 add the google local fonts proposal to
future agenda (on aphillips) due 2025-04-10
[21] https://github.com/w3c/i18n-actions/issues/164
<addison> close #164
<gb> Closed [22]issue #164
[22] https://github.com/w3c/i18n-actions/issues/164
<addison> #162
<gb> [23]Action 162 poll I18N/CSS for new day/time (on
aphillips) due 2025-03-25
[23] https://github.com/w3c/i18n-actions/issues/162
addison: We got responses from Florian and Elika, but Elika not
currently available, so need a new schedule.
<addison> #159
<gb> [24]Action 159 write up proposal for specdev char-string
section, adding material that deals with the encoding interface
et al (on aphillips) due 2025-02-27
[24] https://github.com/w3c/i18n-actions/issues/159
<addison> #157
<gb> [25]Action 157 write glossary proposal identifying options
and next steps for those options (on aphillips) due 2025-02-20
[25] https://github.com/w3c/i18n-actions/issues/157
<addison> #136
<gb> [26]Issue 136 follow up on XML errata (by aphillips)
[task]
[26] https://github.com/w3c/i18n-actions/issues/136
addison: I followed up with plh about XML errata that we want
to file.
… He found that the XML list got closed and asked systems to
reopen.
<addison> #135
<gb> [27]Action 135 follow up on XR issue 1393 about locale in
session (on aphillips) due 2024-10-17
[27] https://github.com/w3c/i18n-actions/issues/135
<addison> #127
<gb> [28]Action 127 make a list of shared topics of interest
between TG2 and W3C-I18N (on aphillips) due 2024-09-30
[28] https://github.com/w3c/i18n-actions/issues/127
<addison> #89
<gb> [29]Action 89 update i18n specs to support dark mode (on
xfq) due 2024-04-18
[29] https://github.com/w3c/i18n-actions/issues/89
<addison> #33
<gb> [30]Action 33 Close issues marked `close?` or bring to WG
for further review (on aphillips)
[30] https://github.com/w3c/i18n-actions/issues/33
xfq: There is anew comment on respec thread, but not looked at
it in detail yet.
<addison> #7
<gb> [31]Action 7 Remind shepherds to tend to their awaiting
comment resolutions (Evergreen) (on aphillips, xfq, himorin,
r12a, bert-github) due 18 Jul 2023
[31] https://github.com/w3c/i18n-actions/issues/7
<addison> #4
<gb> [32]Action 4 Work with respec and bikeshed to provide the
character markup template as easy-to-use markup (on aphillips)
due 27 Jul 2023
[32] https://github.com/w3c/i18n-actions/issues/4
Info Share
<addison> [33]explainers-by-googlers/reduce-accept-language#10
[33] https://github.com/explainers-by-googlers/reduce-accept-language/issues/10
<gb> [34]Issue 10 I18N objections to reducing accept-language
(by aphillips)
[34] https://github.com/explainers-by-googlers/reduce-accept-language/issues/10
addison: There is some activty on ^^
… You may want to take a look.
… What we should reply.
… Next week we'll have visitors.
<addison> [35]w3c/wcag-act#572
[35] https://github.com/w3c/wcag-act/issues/572
<gb> [36]Issue 572 How are rule identifiers matched to one
another? (by bert-github) [i18n-needs-resolution]
[36] https://github.com/w3c/wcag-act/issues/572
addison: Because of ^^
… They don't quite understand our issue.
… They chose to use only lowercase ASCII in their spec,
… but haven't explained it to their users.
… They will come to our call next week and we can talk.
Bert: I will read through tthe issue.
Infoshare
<r12a> [37]https://typo.social/@ri
[37] https://typo.social/@ri
<r12a> first 4 items
r12a: Got Greek case conversion in all browsers now.
… Malayalam orthography notes now with 9000 terms (previously
4500)
xfq: Good dictionary
r12a: Not quite a dictionary.
… Wiktionary is a dictionary.
… But you can search through my list for a term to look up in
Wiktionary.
Review the Review RADAR
<addison> [38]https://github.com/orgs/w3c/projects/91/views/1
[38] https://github.com/orgs/w3c/projects/91/views/1
Pending Issue Review
<addison> [39]https://github.com/w3c/i18n-activity/
issues?q=is%3Aissue+is%3Aopen+label%3Apending
[39] https://github.com/w3c/i18n-activity/issues?q=is:issue+is:open+label:pending
WebNN review
<addison> [40]i18n-activity#1999
[40] https://github.com/w3c/i18n-activity/issues/1999
<gb> [41]Issue 1999 String metadata and localization for
operator labels (by xfq) [pending] [s:webnn]
[t:loc_localization]
[41] https://github.com/w3c/i18n-activity/issues/1999
xfq: I reviewed WebNN and found an issue.
… Core of the spec is for machine learning.
… They have labels, which are strings.
… I asked if it is a language-independent identifier or can
contain natural language strings.
… If the latter, it should have metadata.
addison: I expect they will say it is for programmers only and
not a language string.
… It is a way to propegate error messages,
… We generally think that language metadata is useful here.
<gb> [42]CLOSED Issue 1548 Label by text possibly with natural
language, without metadata attached (by himorin) [pending]
[s:webnn]
[42] https://github.com/w3c/i18n-activity/issues/1548
xfq: They don't clearly say if it is language-agnostic. I asked
they make it clearer.
addison: The label might be the name of the operator.
xfq: Yes, or whatever the developer wants.
RESOLUTION: file the issue
Google Local Fonts Proposal
<addison> [43]explainers-by-googlers/
limiting-local-fonts-access?tab=readme-ov-file
[43] https://github.com/explainers-by-googlers/limiting-local-fonts-access?tab=readme-ov-file
addison: I had some comments on ^^.
<addison> > ensure that the same list of fonts is used for all
users on a platform: why is this a goal?
addison: Their goal says ^^
addison: That isn't necessarily a goal. It is an outcome.
<addison> > This document does not propose solutions for
platforms that have varied or customizable font packages, such
as Linux, for which it is difficult to build an allowlist of
platform fonts due to its wide range of distributions that
tailor to varying user preferences
addison: And they have a non-goal ^^
… But Linux isn't the only platform that is customizabkle.
Windows and MacOS can also be customized, e.g., with language
packages.
xfq: We can't add labels on that repo :-(
r12a: We can make a tracker issue in our repo, though.
<r12a> --
<r12a> Recent CSSWG discussions have brought up the idea of
prescribing user agents to not expose user-installed fonts on
the web as a privacy protecting measure. These measures mirror
Safari’s approach of limiting local font availability by
restricting to fonts that are bundled with the operating system
by default.
<r12a> --
r12a: I agree with what you said, especially about ^^
… I'm not using Safari at the moment, because font support is
so bad.
… We have explained why system fonts alone is completely
inadequate.
addison: Who will file the issue?
xfq: One clreq member already filed an issue.
r12a: The explainer refers to MacOS Ventura as the base? That
is quite old.
<xfq> (about at least don't implement this until there is
systemic support for generic font families)
ACTION: addison: file issue(s) against
limiting-local-font-access
<gb> Created [44]action #167
[44] https://github.com/w3c/i18n-actions/issues/167
r12a: First remark should be that they haven't read the Safari
discussion.
… And they don't seem to acknowledge the issues that
system-font-only creates and they don't talk about the proposed
solutions, such as opt-in and local document use.
Encoding: relationship of windows-1252 and Latin1
<addison> [45]whatwg/encoding#345
[45] https://github.com/whatwg/encoding/pull/345
<gb> [46]Pull Request 345 Explain the relationship between
windows-1252, Latin1, and ASCII (by domenic) [i18n-tracker]
[clarification]
[46] https://github.com/whatwg/encoding/pull/345
addison: Issue says encoding in Windows-1252 includes Latin1.
… We haven't touched our encoding stuff recently.
… It is old.
… AnnevK mentions in the issue that there still problems with
the Latin and Windows-1252, in C0 range
… First 256 characters in Unicode not the same as windows-1252.
xfq: How much impact would this have on authors?
addison: If authors specify latin-1 to smuggle bytes in, they
will be surprised.
… But it shouldn't have impact on web authors.
i18n-activity 2002: language detection and und
<addison> [47]i18n-activity#2002
[47] https://github.com/w3c/i18n-activity/issues/2002
<gb> [48]Issue 2002 Language Detection: Should we allow results
that are less than "und" to be included? (by w3cbot) [tracker]
[t:lang_detection] [s:translation-api]
[48] https://github.com/w3c/i18n-activity/issues/2002
<addison> [49]webmachinelearning/translation-api#51
[49] https://github.com/webmachinelearning/translation-api/issues/51
<gb> [50]Issue 51 Language Detection: Should we allow results
that are less than "und" to be included? (by nathanmemmott)
[i18n-tracker]
[50] https://github.com/webmachinelearning/translation-api/issues/51
addison: If you do detection, you get a vector back with
confidence factors.
… Should UND be allowed in the middle? Or only at the end of
the list, even if confidence factor is higher than some others?
… I think language detectors can retiurn 0 confidence. Depends
on how it is implemented. E.g., some Chinese-like punctuation
could give a low signal for another languge.
… There may be better methods, but you can get small values.
… So if those factors are lower than the factor for UND, should
they still be included in the list?
addison: E.g., several language are written in devenagari, but
if you only look for Hindi, you may conclude a low confidence
factor.
r12a: You cannot detect the language from the script, exceprt
maybe for Japanese, where there is only one language that uses
the script.
addison: Yes, but script can help.
… My Hindi example: if you look for Hindi but it isn't, you
will probably still get a factor, though a low one.
atsushi: [can't hear]
… This is not dictionary-based, but machine learning. Could be
some sort of non-atural-language detection.
… Could be some use case.
addison: It may have no immediately practical use, but could be
useful to test the language detector.
Bert: I'm not clear what it means to detect "und".
addison: In my language detector I also had a factor to express
confidence that it is *not* a given language. E.g., if there
are multiple detecotrs each detecting one language.
… If a detector then can't detect a language with sufficient
confidence, it returns "und".
AOB?
addison: Don't forget next week we'll have WAI people to
discuss ACT rules.
Summary of action items
1. [51]addison: file issue(s) against
limiting-local-font-access
Summary of resolutions
1. [52]file the issue
Received on Friday, 18 April 2025 03:05:03 UTC