RE: [xliff-comment] Re: [xliff] Version Control Commit by DavidFilip

["Yves" <yves@opentag.com>]

> My point is: If a segment has an isolated <sm/> or <em/>, we can just complement 
> it with the corresponding start in the CTR <item>. My more general point being 
> that <item> is a representation of the object model, not of the unparsed original 
> XLIFF markup.

Actually, storing:

 <target>...data<em startRef='m1'/> text</target>

as:

 <item property='content'><mrk id='m1' translate='no'>...data</mrk>

would be an issue when trying to compare annotations, as we technically break one annotation into two identical. I suppose that could be detected because they would have the same ID value, but seeing that while working on a comparison for one of the segments would be difficult to do.

So, a possible solution is to make sure the added start or end tag is recorded too. And that can be done with a CTR attribute:

 <item property='content' ctr:added='start'><mrk id='m1' translate='no'>...data</mrk>

A more complete example:

<unit id='u1'>
 <ctr:changeTrack>
  <ctr:revision appliesTo='target' ref='s1'>
   <ctr:revision  datetime="2016-09-05T10:00:00+6:00">
    <ctr:item property='content'>Text. <mrk id='m1' translate='no' ctr:added='end'>Data. </mrk></ctr:item>
   </ctr:revision>
  </ctr:revisions>
  <ctr:revision appliesTo='target' ref='s2'>
   <ctr:revision  datetime="2016-09-05T11:00:00+6:00">
    <ctr:item property='content'><mrk id='m1' translate='no' ctr:added='start'>... data</mrk> text</ctr:item>
   </ctr:revision>
  </ctr:revisions>
 </ctr:changeTrack>
 <segment id='s1'>
  <source>Text. <sm id='m1' translate='no'/>Data. </source>
  <target>Tekst. <sm id='m1' translate='no'/>Data. </target>
 </segment>
 <segment id='s2'>
  <source>... data<em startRef='m1'/> text</source>
  <target>... data<em startRef='m1'/> tekst</target>
 </segment>
</unit>

Cheers,
-yves