See also: IRC log
<scribe> ACTION: all to look at the TAG issue (DONE) [recorded in http://www.w3.org/2006/04/11-i18ncore-minutes.html#action01]
felix: karl said he will not bring the issue to the TAG
... for now
<scribe> ACTION: everybody to look at the mail about charmod+voicexml (DONE) [recorded in http://www.w3.org/2006/04/11-i18ncore-minutes.html#action02]
<scribe> ACTION: Felix ask Max if he can join next week's i18n core call (DONE) [recorded in http://www.w3.org/2006/04/11-i18ncore-minutes.html#action03]
felix: max and kazuyuki will join us later about this
<scribe> ACTION: Felix to write an IRI implementation report (DONE) [recorded in http://www.w3.org/2006/04/11-i18ncore-minutes.html#action04]
<scribe> ACTION: Felix to find out where the tests from Martin are located (ONGOING) [recorded in http://www.w3.org/2006/04/11-i18ncore-minutes.html#action05]
francois: need to follow up on that
<scribe> ACTION: felix to go trough LTLI within the next 2 weeks (DONE) [recorded in http://www.w3.org/2006/04/11-i18ncore-minutes.html#action06]
<scribe> ACTION: Francois to build a current issues list on charmod (PENDING) [recorded in http://www.w3.org/2006/04/11-i18ncore-minutes.html#action07]
<scribe> ACTION: Mary to look at WS Addr (DONE) [recorded in http://www.w3.org/2006/04/11-i18ncore-minutes.html#action08]
mary: looked at it
... they don't need really comments, it seemed to be o.k.
... they use IRIs
francois: good
... so we just forget about it for now
<scribe> ACTION: Mary will make the updates discussed in Mandelieu on ws i18n (ONGOING) [recorded in http://www.w3.org/2006/04/11-i18ncore-minutes.html#action09]
background: mail from Martin, reply to Max' initial mail about why voice xml 2.1 cannot refer to charmod normatively:
http://lists.w3.org/Archives/Member/member-i18n-core/2006Apr/0019.html
http://lists.w3.org/Archives/Member/member-i18n-core/2006Apr/0020.html
francois: it is a problem with IPA
... but "sounds" in charmod is the sounds as human understand it
...
which IPA is designed for
richard: if you deal with phonemes
... the same phonetic symbol may be pronounced differently, in relation
to language
... at the phonemic level, there are not fine distinctions
francois: the full range of a phoneme is language dependent
... but SSML is speech synthesis
... i.e.,
you create the IPA character and create a sound which is not language dependend at all
... we wrote this MUST because of "normal" writing
systems
... where is no 1:1 relation between writing and speech
... but IPA exists exactly for that
... so the MUST NOT needs an
exception clause like
... "except in the case of phonetic alphabets"
... we could integrate this as an errata in charmod
... next
issue: C013
... I agree
Richard: they don't use the term "encoding" in the same way
... they mean a format for representing data,
not "character encoding" as we do
francois: just a misunderstanding, no breaking
... next issue: C043:
... no breakage
... C010:
easy to fix
... C067:
... they should use a different term, it' possible without breaking anything in their specs
... C014:
... I
don't understand the issue
richard: I don't see that as an issue
francois: it is not explicit, that's all
<r12a> http://www.w3.org/blog/I18NBlog.php
richard: it is blog based
... this will make it much easier to keep the home page up to date
... I used
to run XSLT processes and a lot of copying
... now if you have the permisson to write to the blog, you can do it yourself
francois: so it is a content management system?
richard: it is a blog software called D2Evolution
francois: that sounds good
... does it also do excerpting of news items?
richard: yes
francois: a good CMS is wonderful
... so this would be the HP for the activity
... I'm not sure if they
are needed for the working group HPs
... it is a way of handling news
richard: you can put other things in as well
... it is basically the same page, but with other stuff for
handling news
francois: very nice
... it is permisson based?
richard: yes
... any problems with this?
everybody likes the idea
richard: it will be mapped to the normal /International/Overview.html
... an internal mapping, you won't see
it
francois: very good
http://lists.w3.org/Archives/Member/member-i18n-core/2006Apr/0020.html
<r12a> "C001 [S] [I] [C] Specifications, software and content MUST NOT
<r12a> > require or depend on a one-to-one correspondence between characters
<r12a> > and the sounds of a language."
francois: the don't think that this is an error of charmod, but we need to make clear that the scope of this statement are other languages than IPA
maxf: I'm fine with amanding c001 that it is not about phonetic languages, but the other ones
francois: explains C013
... voice xml documents are XML, i.e. in a single character encoding
... DTMF
is another data object, we don't see a conflict with charmod
maxf: I probably misunderstood what was meant by "data object"
francois: if you use XML, you don't violate C013
... C043 : XML breaks that, that is why it is a SHOULD
... your spec does it on the virtue of XML
... you cannot violate XML in your spec
maxf: does it make sense to say that documents should conform to charmod?
francois: yes, as much as it is applicable
maxf: so if it is XML, does it not come automatically?
francois: a lot does, but even then a lot of things need to be discussed
... there are some charmod
requirements about content
... e.g. the 1:1: correspondence of sound and characters, which we will fix
maxf: another question:
... if the first version of our spec did not mandate charmod
... and if we
mandate it now
... do we loose backward compatibility?
francois: I don't know the spec to give an answer
maxf: it is more a question of QA
francois: you are 95% covered because of XML
maxf: yes, but what about the 5% left?
... e.g. the two syntaxes for character escapes
francois: but that does not break XML
... you encourage people not to break a SHOULD, but if you have good
reasons, it is o.k.
... C010:
... if there is a problem about the term "character" in your spec, it should be fixed
... I don't know
where voice xml breaks, but can't you fix it?
maxf: voice xml 2.1 is a diff spec
... it adds features on top of 2.0
... we don't indent to change
2.0
francois: but you could add a definition saying what you mean by character in the 2.1 diff
... if it is
localized in a section, it is possible
maxf: we talk sometimes about DTMS "characters", not XML characters
francois: to me DTMF is tones, not characters
maxf: there is a mapping to characters
<Kazuyuki> :)
richard: as tones are mapped to characters
... that are unicode characters, right?
maxf: in the voice xml spec, there is no mapping defined
... looking for the part of the spec which talks
about it
<maxf> <choice dtmf="1" next="http://www.sports.example.com/vxml/start.vxml"/>
<maxf> <choice dtmf="2" next="http://www.weather.example.com/intro.vxml"/>
<maxf> <choice dtmf="3" next="http://www.stargazer.example.com/astronews.vxml"/>
francois: what is dtmf=1?
maxf: it says "the DTMF recognizer has received a character '1"
francois: so there is a mapping
maxf: but the DTMF document don't say that, they are not explicit about it
richard: characters like "1" and "2" exist in Unicode, so there is a problem in Unicode
maxf: the problem is our wrong use of "character"
francois: so these are XML characters to match these characters
... so the DTMF recognizer does the mapping,
it is outside the voice xml spec
... if there is a different DTMF version, the spec would just use that
maxf: yes
francois: that's no conflict with charmod
maxf: do you recommend that we change our wording?
francois: if you write s.t. like "if we say characters, we mean XML characters"
maxf: o.k.
francois: C014
... no conflict as well
... so all is left is C001