Meeting minutes
<addison> trackbot, prepare teleconference
Agenda Review
Action Items
<addison> https://
addison: i've been working on the explainer
… thanks r12a for the suggestion
r12a: glossary is in progress
xfq: ACTION-1248, still pending
<addison> w3c/i18n-glossary#22?
<ghurlbot> Issue 22 Items from lreq that could be added to Glossary (r12a)
r12a: ACTION-1251, in progress
Info Share
addison: DST change
<addison> Fuqiao: w3c released a beta version of a new website
<addison> ... feedback is welcome
https://
<addison> ... a few i18n related issues tracked here ^^
JcK: maybe late next week
RADAR Review
https://
Bert: started reading css-text-3
… from top to bottom
… you have to read the text from top to bottom to understand
atsushi: WoT Profile in discussion
… can move it to Awaiting comment resolution
Group Meeting at TPAC
addison: TPAC 2023 will be held from 11 to 15 September 2023, with a main in-person hub in Sevilla, Spain.
r12a: I'd like to go, depending on the covid situation
Bert: would like to go
xfq_: would like to go, not 100% sure yet
atsushi: me too
addison: it seems we will have one to half a dozen participants
ACTION: addison: request physical meeting at tpac
<trackbot> Created ACTION-1252 - Request physical meeting at tpac [on Addison Phillips - due 2023-03-16].
SecurePaymentConfirmation update
addison: I had an interchange with Ian Jacobs about our open issue with them regarding their note in the text
… "we don't have metadata yet, see this issue"
… I made some text proposals to him
… additions to the i18n considerations section
https://
addison: any thoughts?
xfq_: they talk about both localization and string metadata
addison: right
… and linked to qa-bidi-unicode-controls
Bert: maybe there's a way to rewrite the last paragraph to make it clearer
r12a: it's quite difficult to parse that sentence
… "If displayable strings contain or might contain bidirectional text"
addison: suppose you have a field, if they have metadata, you have bidi isolation, and all will be well
… from the point of view of spillover effect
This specification does not (yet) include mechanisms for callers to
associate language or direction metadata with the displayable strings
they provide as input to the API (e.g., {{PaymentCredentialInstrument/displayName}}).
addison: this talks about the base direction ^
r12a: I didn't know what you mean by "bidirectional text"
… it wasn't entirely sure what "the application" is
addison: it's the receiver, the consumer
r12a: "the consuming application needs to ensure that they correctly process bidi text"
addison: two parts
… I will request these changes
… and update our guidance in specdev to match
r12a: I'll finish the revision of qa-bidi-unicode-controls
ACTION: addison: propose changes to the bidi recommendation text and update spc pr 232 with that
<ghurlbot> Sorry, I don't know what repository to use.
<trackbot> Created ACTION-1253 - Propose changes to the bidi recommendation text and update spc pr 232 with that [on Addison Phillips - due 2023-03-16].
addison: thank you
Review proposed changes to accept-language
https://
r12a: basically, Google is saying all this stuff after the first entry of the AL header is not necessary
… which sounds drastic to me
xfq_: they won't send more than one language in the header
r12a: they're talking about reducing Navigator.languages as well
… I assume this is security related
addison: it's fingerprinting
r12a: it's also possible not to get anything out of the web that you would actually want to get out of it
<r12a> https://
<r12a> "Our proposal is to only send the user’s most preferred language (for previous example is en-US) in the Accept-Language header instead of sending all languages in the Accept-Language header."
xfq_: we should think about how we strike a balance between privacy and features users need
<Bert> I wonder how this proposal differs from the existing (experimental) RFC 2295
atsushi: the number of people using more than two languages might be small
addison: the numbers are probably reasonably small
… most browsers use the system locale as AL
… minority language speakers may not run their OS in that language
… it's probably a small number of people relative to the total web population of the web
… these are the people who are disadvantaged the most
… they might be tracked
atsushi: make it behind a permission prompt?
addison: maybe they should have a warning or something
… as a WG, it is not our responsibility to champion security stuff, it is our responsibility to champion global users
… should we take a position say that "don't forget these use cases"
https://
ACTION: addison: add issue to the accept-language repo with sense of WG
<ghurlbot> Sorry, I don't know what repository to use.
<trackbot> Created ACTION-1254 - Add issue to the accept-language repo with sense of wg [on Addison Phillips - due 2023-03-16].
addison: should I take an action to start a conversation with Victor Tan?
xfq_: I think so
I18N Glossary references
<ghurlbot> Pull Request 21 Address #19 (instructions for using) (aphillips)
https://
addison: instructions for use
https://
addison: the Bikeshed section is super new
… they're based on conversation with @@ I had
r12a: "the definitions in the glossary are defined to be informative rather than normative", and may change from time to time
… what I've been wondering is to what extent we expect people to refer to the document
… I think our consensus was to take our definition and maintain it in their spec
addison: two use cases here
… associates with comments we made on some spec
… an example in SPC, we talked about locale
… rather than adding one sentence in their spec
… they can just refer to ours
r12a: that's informative reference
… as opposed to normative reference
… "we're defining locale, with these parameters"
addison: if css defines code point/unit, grapheme etc.
… you want to be specific on what you mean
… we have a lot of jargons
r12a: I would think that it might be useful to extend your second paragraph a bit more
… "if you want normative definition, define it in your document ,if you want informative def, refer to our doc"
addison: what about our relationship with Infra
… are we happy with that relationship?
r12a: expectation is that we're happy with it until we find something we're not happy with
xfq_: in Infra they define string as a sequence of unsigned 16-bit integers
… which might not apply to all W3C specs
… if we define string, we might want to define abstract string
addison: we don't define string
… but we have recommendation in specdev
r12a: I'm rewriting the glossary
export
r12a: when other specs import it, do they import the text, or a link?
addison: link
addison: PNG issue ^
https://
"Decoders running on systems with a non-Latin-1 legacy character encoding should remap character codes so that Latin-1 characters are displayed correctly."
addison: the use I recommended in PNG, is it appropriate?
r12a: Chris asked it because it's in a normative statement
addison: that normative statement contains a link to our informative document
… is that the right thing for us to recommend to people?
… if that normative statement requires a normative def, I would say you should import it in the spec
r12a: import the link?
addison: import the text
r12a: we should make that clear
… when going through the entries
… we don't need any ids because respec adds it automatically
… I can't find lint-ignore in the latest respec docs
addison: I think you can make it a global config
… instead of adding it to all of our dfns
r12a: we don't need plural forms because respec recognizes that
… data-lt captures the alternatives
data-lt="localization|localized|L10N|Localization" can be reduced to data-lt="localized|L10N"
addison: let's test it
… when we're sure I'll clean it
r12a: Unicode code point is not in the right section
… wonder we should post a version it part way through
addison: if you make a PR you can continue working on it
addison: I usually use my fork
r12a: I've done half of these
… is there a benefit in making a PR?
… the way I usually do a PR is to do it on a website
… capture all in dreamweaver and past all on github
addison: if you use github branch you can sync your branch
r12a: when I edit it in github it creates a branch for you
… no need to maintain a branch in my own computer
… I have too many repos
AOB?
r12a: I'm rewriting pretty much the whole glossary
… watch out for conflicts
close action-1249
<trackbot> Closed action-1249.
ghurlbot, bye