RE: [xliff] ref value in translation candidates module element <mtc:match>

["Mr. Rodolfo Raya" <rmraya@maxprograms.com>]

Hello Yoshito,

I don't know what the original goal was, but I'll offer my interpretation.

The specs says that "ref" attribute points to a span of text, that does not mean it must be a <mrk> element, I take it just means that it has to point to an element that contains the text that is being matched.

If my assumption is right, there is an error in example 4.3.1.22 as you correctly pointed out and it should say:

  <mtc:match id="tc01" ref="#seg2">



I certainly hope the original intention was not to force enclosing <source> content in a <mrk> element as was done with <note>.

Forcing the enclosure of <source> content in <mrk> is the only way to point <note> to a segment because <note> lacks "ref". Noted content has to reference the <note> instead of having <note> referencing annotated text. 

I think that <note> and <mtc:match> should both use the same mechanism for pointing at text in a <unit>. I would prefer leaving <source> untouched, but I consider that a consistent mechanism is more important.

The element <gls:glossentry> also uses "ref" but there is no example pointing to a whole <source> content. We may need to add one.

Regards,
Rodolfo
--
Rodolfo M. Raya
Maxprograms http://www.maxprograms.com

-------- Original Message --------
Subject: [xliff] ref value in translation candidates module element
<mtc:match>
From: "Yoshito Umaoka" <yoshito_umaoka@us.ibm.com>
Date: Tue, April 06, 2021 2:20 pm
To: xliff@lists.oasis-open.org

Hello team,

I have an XLIFF specification question about 'ref' attribute used by <mtc:match> element.

http://docs.oasis-open.org/xliff/xliff-core/v2.1/csprd01/xliff-core-v2.1-csprd01.html#candidates_ref

'ref' attribute value is IRI, pointing to a span of text within the same unit, according to this specification. However, I don't understand what is right way to point a span of text with this description.

http://docs.oasis-open.org/xliff/xliff-core/v2.1/csprd01/xliff-core-v2.1-csprd01.html#matchAnnotation

Above specification is explaining use of <mrk> element to annotate a translation candidate. (BTW, this explanation does not look great, because it does not mention <mrk> element directly at all, and just specifying what attributes are used there. It should state <mrk> element is used for the annotation first.)

My question is whether a value 'ref' attribute in <mtc:match> must be pointing a <mrk> element under the same <unit> using the XLIFF fragment identification mechanism always.

I understand the use case example in 5.1.4 Translation Candidate Annotation below is a typical use case.

-----------------
<unit id="1">
  <mtc:matches>
    <mtc:match ref="#m1">
      <source>He is my friend.</source>
      <target>Il est mon ami.</target>
    </mtc:match>
    <mtc:match ref="#m1">
      <source>He is my best friend.</source>
      <target>Il est mon meilleur ami.</target>
    </mtc:match>
  </mtc:matches>
  <segment>
    <source><mrk id="m1" type="mtc:match">He is my friend.</mrk></source>
  </segment>
  <segment>
    <source>Yet, I barely see him.</source>
  </segment>
</unit>
-----------------


What I don't understand is - the example in 4.3.1.22 isolated. I pasted the example below:

-----------------
   <file id="f2" xmlns:abc="urn:abc">
   <unit id="1">
     <mtc:matches>
       <mtc:match id="tc01" ref="seg2">
         <source><sc id="1" isolated="yes"/>Warning:</source>
         <target><sc id="1" isolated="yes"/>Attention :</target>
       </mtc:match>
     </mtc:matches>
     <segment id="seg2">
       <source><pc id="1">Warning: File not found.</pc></source>
     </segment>
   </unit>
 </file>
-----------------

This example is used for explaining <sc> ans <ec>, and includes <mtc:matches>. In this example, you can see <mtc:match id="tc01" ref="seg2">. I think there are two errors in this example.

The first issue is, the value of ref attribute "seg2" is not a valid XLIFF fragment identifier. According to the spec (http://docs.oasis-open.org/xliff/xliff-core/v2.1/csprd01/xliff-core-v2.1-csprd01.html#fragid), it should have prefix '#' always.

The second issue is, I think the intent of this example is to point the text span '<pc id="1">Warning: File not found.</pc>' in <source> element in <segment id="seg2">. Pointing <segment> element does not look right, because it's not a span of text. Or, are we able to use <mtc:match> to point to a <segment> element?


If 'ref' attribute in <mtc:match> can only point to <mrk> element with type="mtc:match", then I think section 5.1.6.5 ref should clearly state the value must point to <mrk> element (with type="mtc:match") within the same unit.

If 'ref' attribute can point to any other elements, although it sounds weird, it should explain what element can be used for specifying a span of text, and should provide an extra example.


Anyway, can someone clarify the original intent, and precise definition of 'ref' attribute value in <mtc:match> element?


Thanks,
Yoshito