Hi Mirek et al,
I think you make some very good points
and agree that we should probably remove the 'phase-name' attribute from
<trans-unit> and <bin-unit> as it can be retrieved, as you
state, from the child <target>.
I also think though that we should add
this attribute to <alt-trans>, in doing this we will allow specification
of when (i.e. which phase) a particular <target> was introduced to
the XLIFF document AND when it was changed from a child <target>
of <trans-unit> to a child of <alt-trans>.
My take on 'phase-name' is that it should
be used to indicate when an item (e.g. a <target>) was introduced
into the XLIFF document. Following from this and extending on Yves' observations,
I think that we should provide another attribute in <alt-trans> to
indication the reason why a given <alt-trans> is an alternate translation.
This attribute could be 'reason' and by making it required we could enforce
more stringent use of <alt-trans>. Values for such an attribute that
spring to mind are 'TM Suggestion', 'MT Suggestion', 'Rejected-Inaccurate',
'Rejected-Spelling', 'Rejected-Grammar' and 'Rejected-Length'. Since this
attribute would be providing mostly informational content, it could be
just free-format plain text, though I think a supplied list of values would
allow for better tooling features, i.e. nobody would want to update a translation
memory with a translation that was rejected for a misspelling, but an inaccurate
translation for a given case may be useful.
Anyone?
Regards,
Mark
Mark Levins
IBM Software Group | Dublin Software Laboratory | Airways Industrial Estate
| Cloghran | Dublin 17 | Ireland
Phone: +353 1 7046676 | IBM Tie Line 166676
Mirek Driml <>
04/12/2002 15:46
To
cc
Subject
[xliff] Phase name attribute
use
Hi all,
I've been looking at the phase-name attribute issue
and also have discussed
it with my coleague Milan Karasek who participated in XLIFF before. The
result of my observation is:
phase-name attribute is used for in <target>
and <trans-unit> elements (and
also in phase, bin-unit and bin-target elements)
We use <source> and <target> elements
inside the <trans-unit> element to
keep the current translation. There are two cases:
1. there is no <target> element - i.e. the unit hasn't been translated
yet
hence we don't need to store any information about phase
2. there is <target> element containing phase-name attribute which
says in
which phase the translation unit occurs
Since the current phase information can be retrieved
from the <target>
attribute I think we don't need phase-name attribute in the <trans-unit>
element and my proposal is to remove it from there.
Concerning alt-trans element and use of phase-name,
following Yves'
observation, I've found we can use it
1. to identify a different suggestion from a TM
2. to capture the evolution of translation during the translation process
3. to identify rejected translations
ad 1) As Yves observered the different suggestion
from the TM should always
have quality-match attribute but no phase-name attribute until it wasn't
used.
ad 2) The evolution of translation during the translation process is fully
covered by the phase-name attribute used in the <target> element
that
resides inside the <alt-trans> element. Only the <target> elements
with
phase-name attribute keep the text used during the process and the
phase-name attribute defines which phase it was used in.
ad 3) I'm not sure what exactly is meant by "rejected translations"
a) if rejected translation is the translation that was used in the process
and later it was corrected e.g. by proofreader then it is included in the
case 2) and we can simply find what and when was changed
b) if rejected translation is the suggestion from any other source than
TM
that has never been used then it should be stored in the <alt-trans>
element
without any phase-name and quality-match attributes
Taking into account the observation above - my proposal
is to remove
phase-name atttribute from <trans-unit> element and to describe the
use of
the <alt-trans> element in the specification in more details to avoid
any
confusion. Similarly it would be removed from <bin-unit> element.
I'm looking forward to your response.
Best regards,
Mirek
PS:
Milan is considering re-joining the XLIFF TC.
Miroslav Driml
SW Group Manager
=============================
Moravia IT, a.s.
Hilleho 4, 602 00 Brno, Czech Republic
Phone: + 420-545-552-173
Fax: + 420-545-552-233
E-mail: mailto:
WWW: http://www.moravia-it.com
=============================
News and latest developments at:
http://www.moravia-it.com/news/
----------------------------------------------------------------
To subscribe or unsubscribe from this elist use the subscription
manager: <http://lists.oasis-open.org/ob/adm.pl>
