XLIFF-ITS minutes and comments (Re: [ISSUE 55] log from friday's call to review its-xliff mapping)

Hi Dave, all,

raw minutes are at
http://www.w3.org/2012/10/26-its-xliff-tc-irc.txt
and below as text, for the mail archive. I didn't generate an HTML version,
since the minutes had no topic markers anyway. Some comments below:

- I would advise against re-opening disambiugation, e.g. with allowing
standoff here too. Given that we have four weeks to go to last call and
from the implementors point of view (currently Tadej) this data category is
stable, we should not invest more time on it. There is too much other stuff
on the plate. So in essence, drop
https://www.w3.org/International/multilingualweb/lt/track/actions/260

- It would be great to have this discussion not totally separated from the
MLW-LT WG. How about using some time from the editors call for this? E.g.
5-6 p.m. CET MLW-LT call, 6-7 p.m. XLIFF/ITS/editors call? After all it
seems that most of the time editing works without a dedicated call for it.

- Moving both to ITS 2.0 last call and continuing this discussion is a
challenge without an additional meeting in November. I would propose a
virtual long meeting (two 1/2 meetings) in November. More to be dicusssed
in Lyon.

Best,

Felix

16:07] == daveL [~daveL@public.irc.w3.org] has joined #mlw-lt
[16:08] == Milan [~Milan@public.irc.w3.org] has joined #mlw-lt
[16:08] <daveL> chair: dF
[16:08] <daveL> scribe: daveL
[16:08] <daveL> rollcall: milan, naoto, davidF, shaun, yves
[16:09] <daveL> apologies: felix
[16:09] == dF [~dF@public.irc.w3.org] has joined #mlw-lt
[16:11] <daveL> rollcall: milan, davidF, yves, dave
[16:11] <Yves_>
http://www.w3.org/International/multilingualweb/lt/wiki/XLIFF Mapping
[16:11] <daveL> rollcall: add naoto
[16:13] <daveL> dF: go throug proposal data category by data category
[16:13] <daveL> ... translate has issue in v1.2 as there is not an
inverse of transalte
[16:14] <daveL> ... in tcd-UL prototype we use custom DT and DNT
[16:14] <daveL> Yves_: proposed using protected for no translate and
just something like 'DT' for the inverse
[16:15] <daveL> ... as this is in the spec
[16:16] <daveL> dF: but is protected actually a full match. But this
is not a big deal, so happy to maximise use of existing
[16:17] <daveL> Yves: so are we agreed that we use protected for no
translate and customer for translate=yes, i.e. x-dt for 'do translate'
[16:18] <daveL> df: note are fairly straight forward
[16:18] <daveL> Yves_: but ITS has attribute for alter/description
which doesnt have equivalent in XLIFF
[16:19] <daveL> df; so can we add an attribute to an exisitng node for this?
[16:19] <daveL> yves_ could we use priority, which is value 1-10
[16:20] <daveL> ... so perhaps 1=alert and 10=descption
[16:21] <daveL> df: does anyone use these already
[16:21] <daveL> yves_; not see it use so far
[16:22] == dF [~dF@public.irc.w3.org] has quit [Ping timeout: 20 seconds]
[16:22] <daveL> df: we have no option since private attribute are not allowed
[16:22] <daveL> .. so we must use priority
[16:26] <daveL> yves; in 2.0 there is no extensibility on note
[16:27] <daveL> df: so we need to propose extension for ITS2.0 for its note type
[16:28] <daveL> yves: agreed or ask TC to introduce this into 2.0 core
[16:29] <daveL> action: dF to propose extensible of note to support
ITS type to XLIFF TC
[16:29] * trackbot noticed an ACTION. Trying to create it.
[16:29] <@trackbot> Created ACTION-257 - Propose extensible of note to
support ITS type to XLIFF TC [on David Filip - due 2012-11-02].
[16:30] <daveL> milan: noted that in 1.2 we can have values less than
1, so description needs to be '10'
[16:30] <daveL> dF: for 2.0 mrk there is no extensibility
[16:31] == joerg [~joerg@public.irc.w3.org] has joined #mlw-lt
[16:33] <daveL> Yves: but we have to consider that these aren't parsed
using an ITS processer, since node don't work as ITS nodes
[16:34] <daveL> ... but we want to indicate we are supporting a
mapping to the corresponding ITS attribute
[16:34] <daveL> daveL: perhaps most clear to use same name but with
alternative name space, e.g. itsx or itsxlf
[16:35] <daveL> action: yves_ to ask XLIFF TC what best practice of
mapping ITS into a namespace in XLIFF
[16:35] * trackbot noticed an ACTION. Trying to create it.
[16:35] <@trackbot> Sorry, couldn't find yves_. You can review and
register nicknames at
<http://www.w3.org/International/multilingualweb/lt/track/users>.
[16:36] <daveL> Yves: to ask XLIFF TC what best practice of mapping
ITS into a namespace in XLIFF
[16:36] <daveL> action: Yves to ask XLIFF TC what best practice of
mapping ITS into a namespace in XLIFF
[16:36] * trackbot noticed an ACTION. Trying to create it.
[16:36] <@trackbot> Created ACTION-258 - Ask XLIFF TC what best
practice of mapping ITS into a namespace in XLIFF [on Yves Savourel -
due 2012-11-02].
[16:38] <daveL> yves: ask if we should be consistent in naming as its mapping
[16:38] <daveL> dF: let revist names later, and focus on structure now
[16:39] <daveL> action: dF to review names of new attributed for ITs
to ensure understandability and consistency
[16:39] * trackbot noticed an ACTION. Trying to create it.
[16:39] <@trackbot> Created ACTION-259 - Review names of new
attributed for ITs to ensure understandability and consistency [on
David Filip - due 2012-11-02].
[16:40] <daveL> yves: for local note we still have problem adding
custom attributes in mrk
[16:40] <daveL> dF: XLIFF TC need to revisit introduction of
extensibility in mrk for 2.0
[16:41] <daveL> df: moving onto terminology
[16:42] <daveL> ... suggest treating its disambiguation togeht with
terminology, as they facets of the same lifecycle, which is interested
to terminology folk
[16:42] == joerg [~joerg@public.irc.w3.org] has quit ["Page closed"]
[16:43] <daveL> yves; but this is inconsistent with ITS
[16:44] <daveL> ... perhaps better for wait until disambig is stable
before attenpting mapping to ITS
[16:44] <daveL> df; so focussing just on terminology, but which structure levels
[16:45] <daveL> yves: a term could be at various structure, but could
just use an enclosing mrk if the whole unit correspond to the term
[16:46] <daveL> ... and there are a bunch of attributes in 1.2 for mrk
whihc includes 'term'
[16:47] <daveL> df: can use similar appraoch on mrk in 2.0, but again
this needs extensibility - a recurring issue
[16:48] <daveL> Yves: we have info and info ref
[16:48] <daveL> ... but not both at the same time
[16:48] <daveL> ... so we can have terminfo extension for 1.2
[16:49] <daveL> ... but for 2.0 we have term again and can have value
for value and ref if it is a reference
[16:50] <daveL> topic: directionality
[16:50] <daveL> df: in 1.2 people just using unicode characters
[16:50] <daveL> yves: so do we map it or not
[16:52] <daveL> df: some people use unicode characters
[16:53] <daveL> ... an extension for trans-unit seem good, and works
well at the paragraph level
[16:54] <daveL> ... but would need similar extension on marker
[16:55] <daveL> ... or use unicode characters for inline
[16:55] <daveL> ... but in 2.0 its up to proposal for directionality
[16:55] <daveL> topic: ruby
[16:56] <daveL> yves; do we need mapping? is it implemented
[16:56] <daveL> df: I have an idea on this, but lets skip it for now
to avoid loosing valuable time
[16:56] <daveL> topic; language information
[16:57] <daveL> yves; could map to standard way of mapping to xliff
mean for specifying at structural manner
[16:58] <daveL> ... but it needs to be defined inlien with xml:lang in mrk
[16:59] <daveL> df: but this is a problem since xliff is bi lingual
[17:00] <daveL> yves; but it is still necessary, and we can use inline
markup as mrk for this
[17:01] <daveL> df: should it be its:langinfo for mrk
[17:02] <daveL> daveL: perhaps map straight into XML lang, since
its:langinfo is really playholder for non-standard lang informaiton
[17:03] <daveL> yves: agreed
[17:03] <daveL> df: dow e need extensibility for 2.0 again
[17:03] <daveL> yves; not in this case, but it comes up a lot
[17:04] <daveL> df: yes this exercise strengthen to case for mrk
extensibility in 2.0
[17:05] <daveL> topic; element within text
[17:05] <daveL> yves; doesn't need to be mapped since it is actioned
at extraction time, in its it is an extraction requirement
[17:07] <daveL> df: so lets look at free time in Lyon to do this
[17:07] <daveL> topic; domain
[17:08] <daveL> df; this would be a candidate for another x-its attribute
[17:09] <daveL> yves: a candidate for having a in-file node
[17:10] <daveL> df: let agree to push its local markup for this
[17:11] == dF [~dF@public.irc.w3.org] has joined #mlw-lt
[17:14] <daveL> daveL; we need to make sure to set teleonf time on
wednesday so yves and co can schedule that in
[17:16] <daveL> df; back to disambiguation, use mrk type = phrase and
prose text for providing value
[17:17] <daveL> yves; but are there any text value for disambig value
[17:17] <daveL> milan: it is mostly urls or other repository references
[17:20] <daveL> df; propose using identref
[17:22] <daveL> ... and put text value in a comment
[17:23] <daveL> df: so for xliff 2.0 do we prefer standoff or an mrk extension
[17:26] == dF [~dF@public.irc.w3.org] has quit [Ping timeout: 20 seconds]
[17:26] <daveL> action: dave to raise need to standoff markup for
diambiguation in its to support multiple annotations in ITS
[17:26] * trackbot noticed an ACTION. Trying to create it.
[17:26] <@trackbot> Sorry, couldn't find dave. You can review and
register nicknames at
<http://www.w3.org/International/multilingualweb/lt/track/users>.
[17:28] <daveL> action: daveL to raise need to standoff markup for
diambiguation in its to support multiple annotations in ITS
[17:28] * trackbot noticed an ACTION. Trying to create it.
[17:28] <@trackbot> Created ACTION-260 - Raise need to standoff markup
for diambiguation in its to support multiple annotations in ITS [on
David Lewis - due 2012-11-02].
[17:28] == Milan [~Milan@public.irc.w3.org] has quit ["Page closed"]
[17:29] <daveL> rrsagent, generate minutes
[17:29] <daveL> rrsagent, make logs public
[17:30] <daveL> rrsagent, generate minutes
[17:31] <daveL> chair: daveL
[17:31] <daveL> rrsagent, generate minutes
[17:35] <daveL> rrsagent, generate minutes