W3C Internationalization Tag Set (ITS) in a non-XML Context

Hi,

Every now and then, questions related to ITS in a non-XML context show up. Felix and myself thus thought it might be a good idea to gather a bit of material around this. Here it comes ...

Best regards,
Christian
==

Central ITS concepts are useful even in a non-XML context. http://www.w3.org/TR/its/#design-decisions mentions for example the abstract notion of data categories. http://www.localisation.ie/xliff/resources/presentations/2010-10-04_xliff-its-secret-marriage.pdf (slide 15) indicates for example that ITS and RDFa could be coupled.

The motivation of thinking about the question "ITS in non-XML context" is for example the following: Content is available in many formats and may be converted several times in a content production chain. Thus, it is not always sure that information related to localization and internationalization is produced and needed only in an XML context. The use of concepts from a well-defined approach like ITS can help to find a suitable approach in a non-XML context for this

When using ITS in a non-XML context, one gets close to the notion of "localization directives" - a concept that is sketched in http://www.opentag.com/locdirectives.htm and implemented in http://okapi.sourceforge.net/Release/Shared/Help/locdirectives.htm .

An example non-XML context would be the following: a "sub-grammar" that includes translatable text in an HTML5 attribute:

       <div x-data-props="title:'Submit', hoverText:'Click to submit ...'">

>From a general point of view, you could take two routes from here:

1. Transform the native format to something more internationalized such as XLIFF

               <group id="g1">
                       <trans-unit datatype="plaintext" id="1" resname="a" restype="button">
                               <source>Submit</source>
                       </trans-unit>
                       <trans-unit datatype="plaintext" id="2" resname="b" restype="tooltip">
                               <source>Click to submit ...</source>
                       </trans-unit>
               </group>

2. Devise your own formalism that implements ITS concepts. This would include at least the following:

a. Ggeneral: define data categories independent of their implementation. Re-use as much as possible. Don't merge existing data categories, e.g. "translate + terminology + ...".
b. For your new (non-XML) context, define a global and a local mechanism to identify pieces of content
c. Keep the precedence mechanisms from ITS: local identification comes first, then global identification, then inheritance (which will be (non-XML) format-specific), then defaults
d. For your format in question (e.g. HTML): use for global identification a selection mechanism that is already well established (e.g. in HTML: CSS selectors). If you need to invent something new, consider something close to XPath

Following these guidelines might help re-using ITS in non-XML contexts. It also helps to keep a door open for using "real" ITS, in case a format conversion will become an option.