[xliff-tools] PO Representation Guide: The PO Header

[Josep Condal]

Hi Asgeir,

You are right, I guess I had my "blinders" on when I read the text of
the specification under <trans-unit>. But now that I read it at a slower
pace, I think that what XLIFF is trying to convey is that there are only
two ways to say that a segment is not translated yet:

1. There is no target element.
2. The state attribute is "needs-translation".

Only the latter allows for the concept of fuzzy-pretranslation (like in
SDLX or Trados, where a segment of the memory is primed into the
bilingual file as proposal for the given segment, normaly at
pretranslation time).

I understand Rodolfo's point on such a important state being controlled
by a boolean attribute, and that the state attribute is too dangerous
for such a function, but I think that the approved= attribute has
another meaning, and it is not advisable to overload it with this one
(the fact of if a segment is translated or not).

I apologize for the confusion caused by misundertanding of the
specification.

However, I think that maybe a translated="yes|no" attribute would be
practical for the <trans-unit> in a future XLIFF revision.

Regards,

Josep.

-----Mensaje original-----
De: Asgeir Frimannsson [mailto:asgeirf at redhat.com] 
Enviado el: viernes, 18 de febrero de 2005 14:48
Para: xliff-tools at lists.freedesktop.org
CC: Josep Condal; Rodolfo M. Raya
Asunto: Re: [xliff-tools] PO Representation Guide: The PO Header

On Tue, 15 Feb 2005 06:08, Josep Condal wrote:
> Regarding this comment:
>  ---
>  "state" attribute of <target> element is used to qualify a
translation.
> This is intended to aid project managers decide when to add a 
> translation to a TM database. Look at the predefined values in 
> http://www.oasis-open.org/committees/xliff/documents/xliff-specificati
> on .htm#state and you will understand what I mean.
>
> IMO, the "approved" attribute of <trans-unit> is better suited to 
> reflect if an entry needs translation or not. If a segment needs 
> translation, simply mark it as non approved. Once the translator has 
> reviewed the segment, he can mark it as approved.
> ---
>
> How about if instead of the "approved" attribute we use the
"translated"
> attribute?
>
> The following excerpt is from the XLIFF 1.1. specification:
> ---
> The required id
> <http://www.oasis-open.org/committees/xliff/documents/xliff-specificat
> io n.htm#id>  attribute is used to identify the <trans-unit> within 
> the file. The optional approved 
> <http://www.oasis-open.org/committees/xliff/documents/xliff-specificat
> io n.htm#approved>  attribute indicates whether the translation has 
> been approved by a reviewer. The optional translate 
> <http://www.oasis-open.org/committees/xliff/documents/xliff-specificat
> io n.htm#translate>  attribute indicates whether the <trans-unit> is 
> to be translated.
> ---
>
> I I understand it right, the actual transition of the approved 
> attribute from approved="no" to approved="yes" seems to be reserved 
> for a second person, who reviews the translation of the translator who

> makes the transition from translated="no" to translated="yes". 
> Obviously, if the translation project is not to be reviewed by a 
> second person, the translator can directly approve his own
translations as a shortcut.

Do you here mean the 'translate' boolean attribute of <group>,
<trans-unit> and <bin-unit>? I couldn't find any 'translated' attribute
in the specification. 

The 'translate' attribute can be set to 'no' to protect a <trans-unit>
from translation. E.g. in tools such as Catalyst, these entries are
read-only. Can e.g. be used in a customized filter where certain parts
of the text is not to be translated (e.g. a source code example in a
programming tutorial?). 

So there is no relation between the 'translate' and 'approved'
attribute.

cheers,
asgeir