Xliff
TC Teleconference Meeting - 22nd
April 2003
Present - Doug, Enda, Gerard, Milan, Mirek,
Tony, David. John, Bryan, Christian
2. Last Weeks
Minutes
John moved to accept,
Mirek seconded. The minutes were accepted
3. Review
state-qualifier ballot
12 of 17
votes cast
=>Motion
caries @ 58%, state-qualifier will be added
General comment made by some members of TC -
online ballot system is great!
Shigemichi pointed out
that both needs-localisation and needs-l10n were listed as possible values for
state.
This has been noted and will be clarified,
needs-localisation will be removed so that only needs-l10n will
remain.
4. Design
Loophole
Agenda item carried over from last week.
Problem is that there are a number of different ways to
specify the language of text items in an xliff file.
<xliff> has an xml:lang
attribute
<file> has a source-language
and target-language attributes
<source> has an xml:lang
attribute
<target> has an xml:lang
attribute
While xliff was originally intended to be bi-lingual,
the loophole technically allows for multi-lingual
files.
Different <source>s can have different
xml:lang values and the same goes for <target>s
Also, it is technically possible for the
xml:lang values in <source> and <target> to be different from those
specified in <file>
The TC expressed the desire to leave the specification
deliberately vague so that people could use the xliff format for different
purposes.
It is also technically difficult to close the loophole
at this stage.
The TC decided not to close the loophole at this
stage.
The Chair pointed out that this leaves 2
options
It was decided to make a strong recommendation in
the spec that the <file> element was intended to be uniformly
bi-lingual.
So, all <source>s in a <file> should
have the same language, and all <target>s should have the same
language.
The exception to this is the <alt-trans> element
which could, for example, display a different language's translation for
reference purposes.
John agreed to work on a paragraph for this to be
included in the spec.
5. New suggestions for
restype
Mirek
& Milan feel these 3 are missing
binary?
trans-unit
with type of binary, is that not the same as bin-unit?
It was decided that binary is actually the same
as rcdata (in
windows) which is already
present.
For this reason, binary will not be
included.
usercontrol?
usercontrol
could be useful for custom controls
versioninfo
Version information resources
A phone vote to accept usercontrol (userdefined)
and versioninfo was held
Unanimous
acceptance!
A concern is expressed by Christian and agreed with by
the Chair that restype was supposed to be closed off already and that we need to
stick to our schedule.
6. - size-unit glyph not accurate, suggest grapheme-cluster
Chair -
we've closed this before
Gerard
moved not to discuss this, Doug seconded
Chair
called for any objections?
No
objections!
=>Item
left as is, i.e. glyph
7.
Plans and
timetable!
Today's
changes will be added by Tony and sent around as Draft13 along with
John's suggestion for the recommendation in
item 4
Any review
comments from Doug and John, or anybody
are welcome at this time.
Tony will
try to get final RC for next week
Chair
would like to send out final spec for next week,
receive some feedback, and resubmit the amended final RC for approval at the
next meeting (2 weeks from now)
Reconvene 6th May with a view to accepting the
spec
8. - AOB
IPortal
meeting tomorrow.
Phone
number sent out on Fri
Meeting
adjourned
Best regards,
Enda