Auto-translation and the concurrent localization method

Auto-translation is always enabled in the concurrent localization method, but its behavior is very flexible.

Note: This section applies to the new concurrent localization method only. It does not apply to the node-based localization method.

Configurable behavior

By default, auto-translation does a character-by-character comparison, so changes in the element name, an attribute, or an attribute value, as well as any punctuation or white space are considered a difference that prevents the auto-translation.

However, there may be differences that should be ignored during the auto-translation process. For example, a new draft-comment added by the writer or a punctuation change may not need to be translated. You want to be able to reuse the last valid localization for changes that you consider minor.

You can configure the following items to make the auto-translation process more flexible:
  • Ignore white spaces: If the contents of the current source and the previously translated source differ only in spaces, carriage returns, new lines, or tabs, you can configure the CCMS to consider both elements as identical. In this case, the CCMS will use the text of the previously localized revision during auto-translation.
  • Ignore certain characters: If the contents of the current source and the previously translated source differ only in certain characters (such as single quotes, double quotes, commas, etc.), you can configure the CCMS to consider both elements as identical. In this case, the CCMS will use the text of the previously localized revision during auto-translation.
  • Ignore certain elements: You can configure a list of elements that will be ignored during auto-translation. If the contents of the current source and the previously translated source differ only in certain elements (for example, <draft-comment> or <data>), you can configure the CCMS to consider both elements as identical. In this case, the CCMS will use the text of the previously localized revision during auto-translation.
    For example, consider the following segments:
    Authoring Localization
    English Rev 1:
    <p locid="1">The following diagram shows the localization process.</p>
    		 French Rev 1:
    <p locid="1">Le diagramme suivant illustre le processus de localisation.</p>
    English Rev 2:
    <p locid="1">The following diagram <draft-comment>Note to writer: Please review in next release</draft-comment>shows the localization process.</p>
    		 French Rev 2:
    <p locid="1"  translate="no">Le diagramme suivant illustre le processus de localisation.</p>
    The <p locid="1"> element is considered identical in both revisions even if revision 2 contains an additional element (<draft-comment>), since this element was configured as ignorable. In this case, the CCMS uses the localization from Rev 1, removes the <draft-comment>, and adds the translate="no" attribute to the paragraph.

Non-configurable behavior

In the concurrent localization method, the auto-translation process will consider as identical two elements that differ in the element name only. For example, consider the following segments:

Authoring Localization
English Rev 1:
<p locid="1">This is the <i locid="2">title</i> of the book.</p>
 French Rev 1:
<p locid="1">Ceci est le <i locid="2">titre</i> du livre.</p>
English Rev 2:
<p locid="1">This is the <cite locid="2">title</cite> of the book.</p>
 French Rev 2:
<p locid="1" translate="no">Ceci est le <cite locid="2" translate="no">titre</cite> du livre.</p>

The only difference in revision 2 is a change in element name. In this case, the CCMS will consider both segments as identical and will use the translated copy of the previously localized revision during auto-translation, changing only the element name and adding the translate="no" attribute.