See also: IRC log
checking attendance
http://lists.w3.org/Archives/Public/public-i18n-its-ig/2014Jul/0017.html
serge: if there is a public
service which will enable content to be published
automatically, everbody will be affected
... we need to provision some hocking points for communities
that will be engaged
... to indicate that they will participate
... want to demonstrate that this is relevant to them as
well
... if you speak to data annotaiton to the public, they will
say "what?"
... definitely LT community will be affected
<dF> will join audio shortly
serge: language service community
will also be affected
... e.g. preparing data, testing, other stuff
... so we need to mention them
... so some simple language that makes that clear
... then about providing feedback: nobody will do that
blindly
... in the future that will be avail. for automatically
translated data too
... it has to be in the data itself - the feedback
... in industry sector there is much talking about European
language cloud
... but there is no detailed info about this
... but if we use the token, people will find the
document
... so it is an important keyword to place
... this is a very important document, I like it
... it is laying out good foundation for public services and
automated translation
... but it is important to lay place for future engagement of
other stakeholders
... now to B: quality measurement
... I don't see it practical to blindly publish material
... to some extend it has be manual, to some extend
automatic
... things have to be in the data, that is the concept
... that is why I believe that ITS LQI is important here
felix mentioning linked data represention of ITS and MQM (ongoing) as a nice way to move quality aspect in data world forward
serge: mentioning ITS and RDF
would be appropriate here I think
... about point 1: previous draft had a question "what is
standards?"
... one should mention ITS and RDF, and the standards bodies
that are concerned with those
... no two: reference model graphics: it has QA component
... but I don't think it is an industry term
... the term encompasses severa items - strictly speaking it is
a standards components box
... before MQM and DQF and the QTLP project, there was an
incorrect notion
... if you improve the process the quality will improve by
itself
... that is a problem of many quality related standards
... that is why something like MQM is rquired: you need to be
able to build the metrics
... many companies provide LQ assurance as a separate process
step
... clearly this is a building block
... also required for full automatic translation
... and that is why qtlp started and mqm came up
... LQA is an industry term, one should refer to that
now number 3:
scribe: we need to change data
mgmt requirement 5 to m (mandatory)
... without a feedback channel it will not be possible to
improve machine translation
... it should not be optional, but mandatory
... like google, which allows you to edit mt text
... point 4:
... proposal to improve requirement 5
... point 5:
... one should mention here MQM
... with all these changes this document will be more
complete
... about point C:
... we have a certain lack of stakeholders for the doc
... need to do certain outreach that will engage relevant
stakeholders
... right now list of contributors is quite narrow
... one of the recent calls there was a question: why was there
no call for comment to industry?
... with this changes the document is in a good shape to give
feedback to wider industry
... one should initiate wider feedback
discussion on how to move the docuemnt forward - felix saying that there is time for that, no need to hurry, and now we can reach out to outside telling: give your input
serge: I would reach out to
others and say: this is our current state
... one should contact projects and say: please provide
input
felix: happy to send out in MLI for feedback
<scribe> ACTION: felix to draft outreach mail for doc for others to re-use [recorded in http://www.w3.org/2014/07/16-i18nits-minutes.html#action01]
<trackbot> Created ACTION-49 - Draft outreach mail for doc for others to re-use [on Felix Sasaki - due 2014-07-23].
serge: we can solicit feedback from gala, we have a huge list of contacts
<scribe> ACTION: serge to draft outreach mail for doc for others to re-use [recorded in http://www.w3.org/2014/07/16-i18nits-minutes.html#action02]
<trackbot> Created ACTION-50 - Draft outreach mail for doc for others to re-use [on Serge Gladkoff - due 2014-07-23].
close action-49
<trackbot> Closed action-49.
david: nothing I disagree with,
looking into it now
... many aspects are preparatory, but that is important
... standard emphasis, interchange capabilities
david: XLIFF 2.0 majority vote
passed yesterday
... that means: within a week OASIS membership vote will
start
... that is like: after WG makes something final it goes to AC
vote
... so all OASIS members will vote
... there are currently 283 members
... only contributors or sponsors can vote
... we need to get at least 43 votes to get OASIS standard
label
... it is not so easy
... out of 283 I know only 20 are on the TC
... we really need to get some help to reach out to primary
representatives to get positive votes
... I have a list of OASIS members who are also in W3C, 32
orgs. So AC rep in W3C may be the same like OASIS
... I think it would help to make AC reps aware of this
development
felix: focus on technical aspect of relation between XLIFF2 and ITS2, not so much relation between standards bodies
david: like tech approach
here
... explain that bitext is needed in multilingual
transformation, here is the relationship
... it solves the problem, preserve the metadata
discussion on OASIS support and W3C support rules
jirka: for docbook it was easy
david: we do it the way you did:
finding contact owners for companies
... need to freeze wiki mapping as a snapshot
... stop in the ITS IG and make this an XLIFF 2.1 feature
serge: what stage is xliff 2.1 in?
david: TC decided to work on
it
... there is informal consensus on the features that should be
in:
... the oasis admin is aware that we are working on it
... the timeline is: at feistiltt people agreed on releasing
minor version yearly
... for 2.1 there should be deadline for reference
implementations in November
... january - may should be spent on admin steps
... tech stuff should be achieved betwee the summer and
November
... feature set is narrow
... but big features. two major features:
... offical ITS support
... a lot of ITS features are possible in xliff as
extentions
... but a module is a better status for the feature
... the ITS support should become and official part of the
spec
... ITS may be several modules to avoid interdepencies
... the other feature is advanced validation support
... that is also why felix joined the TC
(discussion on validation topic in oasis)
serge: happy to create news item for the industry about this
david: great
... it is not yet recorded as an offical feature yet because of
summer vacations etc.
serge: will create a blurb from the minutes, for you to approve
david: would be great to make
some fuzz about XLIFF 2.0 becoming a standard
... CNGL will make a press release about that
serge: I'll do that too
david: you can announce xliff as
a standard, saying that the final vote is happening these
dates
... I will forward the notification to this group so that you
know that it is official
david: will have a feisgiltt
event in vancouver
... it overlaps with w3c tpac unfortunately but it is important
to raise standards awareness in loc community
... focus will be on TBX/RDF, ITS mapping etc. - an outreach
effort, bring the news to north america
... if you have any input to call for paper, let me know
... we are using same PC that we had for Dublin feisgiltt
<philr> Apologies, I have to leave for another meeting.
serge: added ITS/RDF to https://www.w3.org/International/its/wiki/Open_Data_Management_for_Public_Automated_Translation_Services#Terminology , need some links
felix: I'll add the links
adjourned
This is scribe.perl Revision: 1.138 of Date: 2013-04-25 13:59:11 Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/ Guessing input format: RRSAgent_Text_Format (score 1.00) Succeeded: s/features/feature/ Succeeded: s/why/also why/ No ScribeNick specified. Guessing ScribeNick: fsasaki Inferring Scribes: fsasaki Present: fsasaki philr serge dF jirka Regrets: dave tatiana olaf-michael cLieske Agenda: http://lists.w3.org/Archives/Public/public-i18n-its-ig/2014Jul/0013.html Got date from IRC log name: 16 Jul 2014 Guessing minutes URL: http://www.w3.org/2014/07/16-i18nits-minutes.html People with action items: felix serge[End of scribe.perl diagnostic output]