DITA translation background

[Don Day <dond@us.ibm.com>]

I was asked to provide some background on DITA's current design decisions
regarding translation and localization.

DITA took some initial inspiration for translation support specifically
from Richard Ishida's early set of XML best practices, "Localisation
Considerations in DTD Design"
(http://xml.coverpages.org/IshidaDTD-Paper.html).
   –  translate=“yes/no”
    • Indicates whether the content of the element should be translated or
not.
  –   xml:lang=“value”
    • Specifies the language of the element content when the language is
different from its context.

When no xml:lang value is supplied, the default value of English is assumed
(a processing default in the absence of a specific value at the root
level). This application of xml:lang for any element is expressed also in
the W3C's core XML Recommendation
(http://www.w3.org/TR/REC-xml/#sec-lang-tag).

Another best practice for translation reuse is to provide unique attributes
for all potentially reusable strings.  The DITA Architecture for conref and
topic referencing guarantees unique ids for reusable content, as well as
the usual validation of correct reuse of referenced elements in context (an
improvement on usual entity set declarations).

The DITA DTDs were originally designed to support the use of specialized
DTDs to provide translated title text via attributes.  Practice has proven
the general advice not to include translateable text as attributes, hence
there is now a work item already approved by the TC for DITA 1.1: #11
"Create elements for text attributes with translatable text."

The original DTD DTDs represent a balance of best practice for translation
with authoring ease and practicality.  For example, the possible country
values for the xml:lang attribute are left for authoring tools and
processing tools to validate, since the list values are potentially
changeable.

DITA anticipates the usual case, especially in scientific literature, of
quotes, names, terms and other examples of languages that are foreign to
the root language of the topic. Use of the xml:lang attribute when known
about the origin of these quips is a hint for proper treatment of such
content in it translated context.  Multilingual topics are not the usual
case for DITA content, other than perhaps safety notices consolidated in a
single, multilingual topic.

Regards,
--
Don Day
Chair, OASIS DITA Technical Committee
IBM Lead DITA Architect
Email: dond@us.ibm.com
11501 Burnet Rd. MS9033E015, Austin TX 78758
Phone: +1 512-838-8550
T/L: 678-8550

"Where is the wisdom we have lost in knowledge?
 Where is the knowledge we have lost in information?"
   --T.S. Eliot