i18n core WG

previous minutes

Francois: approved

actions

<scribe> ACTION: everybody to look for issues in http://www.w3.org/TR/timezone (ONGOING) [recorded in http://www.w3.org/2006/10/24-i18ncore-minutes.html#action01]

input from Felix at http://lists.w3.org/Archives/Member/member-i18n-core/2006Oct/0014.html

Francois: not clear what the difference between "UTC offset" and "zone offset"

Felix: it's the same

Francois: maybe Mark has something to say on this
... let's wait a week or so

<scribe> ACTION: Felix to look for possible participants for the ws i18n work (ONGOING) [recorded in http://www.w3.org/2006/10/24-i18ncore-minutes.html#action02]

<scribe> ACTION: Felix to write a mail about possibility for SVG tiny specific IRI tests to martin and the i18n core list (ONGOING) [recorded in http://www.w3.org/2006/10/24-i18ncore-minutes.html#action03]

mail about the talk with Chris http://lists.w3.org/Archives/Public/public-i18n-core/2006OctDec/0010.html

<scribe> ACTION: Francois to give input to wiki for the LTLI summary (ONGOING) [recorded in http://www.w3.org/2006/10/24-i18ncore-minutes.html#action04]

<scribe> ACTION: Francois to reply to Schema WG on normalization issue. (PENDING) [recorded in http://www.w3.org/2006/10/24-i18ncore-minutes.html#action05]

<scribe> ACTION: Francois to build a current issues list on charmod norm (ONGOING) [recorded in http://www.w3.org/2006/10/24-i18ncore-minutes.html#action06]

<scribe> ACTION: Felix to update the review radar (DONE) [recorded in http://www.w3.org/2006/10/24-i18ncore-minutes.html#action07]

LTLI discussion

editor's copy at http://www.w3.org/International/core/langtags/

<fyergeau> http://www.w3.org/International/core/langtags

francois: language identification part is fine
... needs updating for the RFCs

mark: the pointer to BCP 47 , you have to be careful
... you only get to a part of the BCP. Send a note to the RFC-editor that there should be a pointer for both RFCs

<scribe> ACTION: Felix to send a note to the RFC editor on the reference to the location of BCP 47 - RFCs [recorded in http://www.w3.org/2006/10/24-i18ncore-minutes.html#action08]

<mark> http://www.rfc-editor.org/rfc/bcp/bcp47.txt

mark: this link leads not to the matching part. in the LTLI draft, there is a different link ftp://ftp.rfc-editor.org/in-notes/bcp/bcp47.txt

francois: which one is the offical one?

mark: not sure

francois: we can update the entry in the LTLI draft

felix: I'll do

francois: the parts on language identification is fine
... we have more work as we work at locales

felix: is it still o.k. to concentrate on best practices for "language versus locale", and not define the field values like CLDR?

mark: yes

francois: LTLI should give recommendations how to specify locales on the web, kind of "best practices"

Mary: in some cases encoding was part of a locale

Francois: in posix, but that was not good
... this could be a subject we could address

discussion on various locale models

mark: java and cldr are almost purely on language

felix: besides "locale versus language", how about matching? should LTLI talk about it?

mark: could be
... as Richard pointed out, many people put in some language information not exactly
... we did a search what people put in accept-lang , and it is a mess
... e.g. many people put in the underbar instead of a dash
... we would canonicalize those
... the use of xml:lang or lang as attributes gets even worse

felix: a kind of BP for matching in e.g. http and xml?

mark: bp for receivers / senders

francois notes IE7 and language negotiation

mark: it is extremly fuzzy what a "locale" means
... does "en-US" means my citizenship? the place I am on vacation?

francois: agree, but the language tag received from a browser is not from where you are travelling

mark: it is the language of my browser setting
... there is nothing we can do to influence people to set up their browser

francois: people will not read the w3c spec and set up their browser

mark: we do language detection on documents
... we don't rely on the tags
... that includes both xml:lang and the http header
... there is a lot of material mistagged
... going back to locale:
... I would recommend BP for a user should do , a browser / UA, a server, a receipient

felix: sounds good

francois: on web applications
... which resemble more desctop applications than web pages
... for whose, we need to describe how an application could gather the information

mark: at least make the relationship between language and locale clearer would be good
... we don't want to say too much on how locales are used

francois: I agree
... let's just say "it varies", and the application needs to find out how to deal with it
... we will end up with examples saying e..g "the country is not everything!"
... if a user contacts a server, there is an initial request
... you have to start from that point and use the right language
... the language again is in the center

mark: language is the core of any notion of locale

francois: is there any other stuff than language which is always available?

idna issue

mark: I looked it over
... the klensin document was seriously flawed

felix: we could reply as a group, or as w3c liaison

mark: I would suggest i18n core looks over the proposal, and w3c sees if you reply as liaison to ietf

francois: why are they excluding so much?

mark: the ietf draft is very suspicious on combining characters. you can't write e.g. indic language without these
... the approach is very euro centric

francois: is the appraoch in principle o.k., i.e. to say what is included, and not what is excluded?

mark: at least the unicode identifiers should be available
... they are removing lots of stuff in response to not clearly defined problems

francois: the underlying problem is security, right?

mark: excluding symbols like a "heart" does not help security