[xliff-tools] Translator Comments

[Tim Foster]

Hi there,

On Tue, 2005-02-22 at 01:37, Asgeir Frimannsson wrote:
> > For what it's worth, we're taking the comments
> > from .po files and storing them as bits of <context> information,
> > attached to each trans-unit.
> 
> Ok, what you're doing in your filter is:


> # This is a comment about this file
> msgid "this is text"
> msgstr ""

Yep :

<trans-unit id="a1">
       <source>this is text</source>
       <count-group name="word count">
          <count count-type="word count" unit="word">3</count>
       </count-group>
      <context-group name="message id">
          <context context-type="record">this is text</context>
      </context-group>
     <context-group name="context info">
          <context context-type="record">
# This is a comment about this file
</context>
      </context-group>
    </trans-unit>

> What are the benefits of using this approach? Why is a general '# ' style 
> comment context-information in your implementation? 

Well, drawing the line between what's a note and what's context is subtle :
we usually use <note> elements for actual translator comments, rather than
stuff that came from the source file. We don't know that the comments in
the source file are directed at the translator, so all we can do is store
them in file, and let translators look at them if they want to (in our
case, using the Edit->View Source Context radio button : pops up a floating
window that shows context for the highlighted segment, and updates the context
should you switch to another context)

On the other hand, <note> elements get more visibility in the editor - as
they're (perhaps) written for translators specifically : either by
translators who've done a 1st draft of the translation, or from file formats
that support then (can't think of any off hand, but that was our thoughts on
it)

> For what it's worth, I think PO comments and auto-comments are important for 
> translators to be able to view. Many XLIFF editors hide most 
> context-information from developers, but most editors have support for 
> display of <note> elements.

Right - we show both, just at different levels : context information
we tend to show in a separate floating-window if the translator asks
for it, notes get an icon beside each trans-unit, so that the translator
always sees when a note is present.

(eg. "Fred, can you look over this translation, I'm not sure it's correct")

I don't think it's a big deal whether you go for <note> or <context> when
displaying comments ultimately, just that it's nice to have the distinction
(somehow) between stuff that came directly from the source file, and stuff that 
was written by a reviewer or previous translator.

> I don't drink coffee, and haven't seen self-awareness in PO files. Make your 
> own conclusions :)

Right - good point, I'll try to cut down ;-)

	cheers,
			tim

-- 
Tim Foster - Tools Engineer, Software Globalisation
http://sunweb.ireland/~timf http://blogs.sun.com/timf
http://www.netsoc.ucd.ie/~timf