Internationalization Working Group Teleconference

13 January 2022


addison, Atsushi, Bert, Erik Nordin, Fuqiao, JcK, Richard
Addison Phillips

Meeting minutes

<addison> trackbot, prepare teleconference

Agenda Review

richard: web authentication, use gettext from GNU for localization

Action Items

<addison> https://www.w3.org/International/track/actions/open

addison: 1094, on ecma-402
… start that conversation, and working on explainer
… backoff next week

addison: 1097, no time on this

addison: 1098, on lang-bidi-use-cases

richard: not yet. (first week after off)

addison: 1100, get this shortly

addison: 1101, did and they replied that they are welcome
… will create items

ACTION: addison: create naur review item

<trackbot> Created ACTION-1108 - Create naur review item [on Addison Phillips - due 2022-01-20].

<addison> close action-1101

<trackbot> Closed action-1101.

addison: 1103, the same explainer as ecma-402, not yet done

addison: 1104

addison: 1106, still need?

<addison> close action-1106

<trackbot> Closed action-1106.

jck: still need

<addison> https://github.com/WICG/handwriting-recognition/issues/1

addison: 1107, did do, and close this

<addison> zxx-x-shape

addison: they decided to use zxx

<addison> close action-1107

<trackbot> Closed action-1107.

Info Share

addison: welcome back, happy gregorian new year
… and welcome new member, Nordin joined yesterday

erik: internationalization team at Mozilla, joined this year with my colleagues
… one is in trainig of company and not here with me today

<addison> (other new person from Mozilla is Eemeli)

erik: internationalized domain, ICU, i18n toolings
… joining to the WG, involved to message format 2.0
… interested in DOM localization
… we hope to be more involved and to get world more internationalized

<xfq> https://github.com/w3c/documentreview/issues/28

xfq: new status called registry, this needs wide review as well
… added an issue to documentreview on this
… we can add some correspondence template into our repository

addison: registration strategy or value?

xfq: basically strategy and documentation

<JcK> Oh, in case that got lost over the holidays, the magic BCP link is https://www.rfc-editor.org/info/bcpNNN

xfq: no published registry document yet
… we will see whether they will really need review or not

addison: we could need different tag or status?
… who want to think about these?

richard: inclined to wait until we need to look into

xfq: same for me, and this is just a heads up

addison: chicken and egg
… wait until one comes
… for now, at least one in our pipeline?

xfq: webcodecs is creating registry, and coulbe in the pipeline

addison: keep it open

richard: monotype is interested in looking Arabic justification automatically
… especially on window stretching
… 6 people together and look into this area
… should be useful if we can find some solution
… could be plugged into CSS at some point
… they also seek to extend from Arabic to others

RADAR Review

addison: so, taking Arabic as one of majors

<addison> https://github.com/w3c/i18n-request/projects/1

addison: it might be useful to bring thease into alreq repository once some secured text created

addison: no incoming review request
… any status update?

richard: a month ago, tried to raise gap-analysis pipeline, on rewritting for testing
… started raising bugs, and Mozilla is working on these bugs
… continue to do that

<r12a> https://github.com/w3c/i18n-activity/projects/3

webauth recharter and gettext

richard: charter states to use gettext for localization facility
… read gettext, long documentation

addison: not hopeless, but why W3C spec pickes up this localization technology

richard: this is also kind of internationalization technology

<r12a> https://www.gnu.org/software/gettext/manual/gettext.html

<xfq> VC WG charter

richard: wondering how bidi is handled

<r12a> white-space

<r12a> # translator-comments

<r12a> #. extracted-comments

<r12a> #: reference…

<r12a> #, flag…

<r12a> #| msgid previous-untranslated-string

<r12a> msgid untranslated-string

<r12a> msgstr translated-string

<r12a> eg.

<r12a> #: lib/error.c:116

<r12a> msgid "Unknown system error"

<r12a> msgstr "Error desconegut del sistema"

richard: one line for original as id, one line for translation
… just a text file

addison: you may want to escape bidi, and make text as human readable
… but don't remember of escaping

bert: using gettext for some of my software, simple text file
… don't think any standard on markup nor escape
… usually tags are not for translation

addison: back to my meta question, whey would W3C points these specific thing

<xfq> https://www.w3.org/TR/mlw-metadata-us-impl/#Using_ITS_2.0_with_GNU_gettext_utilities.2FPO_files

addison: why not other internationalization or localization scheme?

xfq: our document has some lines for using GNU gettext
… gettext is stream comversion

addison: does this WG use it for ones like angle brackets
… and be careful of bidi controls

richard: gettext uses language code, two char with subtags

addison: Linux systems are using gettext for long time, and original source

richard: several advices on prurals also

addison: any other action?

richard: will make some comments

<xfq> VC WG charter

example of targets

How to order items in a multilingual list

<r12a> https://github.com/w3c/i18n-discuss/issues/20

richard: this thread is wondering also of direction which has clear guidelines
… discussing what order list is shown in a page
… and second point, getting item to the top based on users' preference
… conclusion could be different when you have long list or not

bert: wikipedia uses order of language code

richard: it is predictable
… same number of translations to be shown in pages of translation
… you will get one solid idea between pages on this order

richard: even picking some languages, we need to order lists

addison: randomly ordered list is possible...

richard: any preference?
… over whole list if there is no preferred language

<r12a> https://www.w3.org/International/articlelist

addison: it works ok
… if you are on long list, you need to know which language is in which code
… when we have completed list of cldr language list, many unknowns

<r12a> https://r12a.github.io/app-charuse/

richard: just typing and finding could work most of time
… above page has all in English and can filter and jump by typing

richard: starting talking on ways of getting information
… what order is preferred

addison: my preference is ordered by lang code, and pulling by typing
… for example, currency selection has the same issue

richard: that is the preference on W3C site redesign
… I think we should provide customized list, rather than providing ordered list

addison: if the list is large, people could miss the group at the top

bert: found pros/cons for that grouping

richard: if you only have three translation, could be picked up by browser preference
… I suspect not selecter
… in our article, like 40 translations, ordered list is better

addison: including English also seems better
… and also regionaly sorted groups

richard: to find a language quicker, having items at the top from preference could be better

addison: haven't done, keens to test that
… draw back is people could remember around where item was

richard: number of languages could differ per page
… if you have 6 languages for a page, you can look without checking selections by scrolling

addison: just to have lists in the same order as language tag, not to cause flame

eric: mozilla is always aware of identifying information than just ordering

addison: language preference is sent to all web server as header


addison: around top, any other?

<Bert> (But need to take into account that the lang pref may be wrong on purpose, precisely because people care about protecting their personal data.)

Summary of action items

  1. addison: create naur review item
Minutes manually created (not a transcript), formatted by scribe.perl version 185 (Thu Dec 2 18:51:55 2021 UTC).


Succeeded: s/yera/year/

Succeeded: s/DOM internationalization/DOM localization/

Succeeded: s/chickin and egg/chicken and egg/

Maybe present: eric, erik, xfq