DITA Translation Subcommittee Meeting Minutes -- 12 March 2007

["JoAnn Hackos" <joann.hackos@comtech-serv.com>]

 DITA Translation Subcommittee Meeting Minutes: 12 March 2007

 

(Recorded by JoAnn Hackos)

 

The DITA Translation Subcommittee met on Monday, 12 March 2007 at 08:00am PT
for 60 minutes.

 

1)  Roll Call

 

    Present:
        JoAnn Hackos
        Robert Anderson
        Rodolfo Raya
        Don Day
        Andrzej Zydron
        Nick Rosenthal

 

    Regrets:

 

2)  Accept Minutes from 26 February 2007
    Accepted. [JoAnn moved, Rdolfo seconded, no objections]

 

3)  Review open action items

 

    3.1 ACTION: Gershon to investigate whether he can use a client's samples.

 

        CONTINUED. (client is awaiting legal OK)

 

    3.2 ACTION: Gershon and Don will present the approved best practices on
        indexing and Translation Memory as committee drafts for approval by the DITA
        Technical Committee.

 

        ON HOLD (until DITA TC closes 1.1 spec)

 

    3.3 ACTION: Everyone to review the document JoAnn posted on 13 November
        (subject: Document from Christian Lieske).

 

        COMPLETED

 

    3.4 ACTION: Don to raise glossary work item at the TC after the 1.1 draft to
        see if it needs its own subcommittee.

 

        ON HOLD (until DITA TC closes 1.1 spec)

 

    3.5 ACTION: JoAnn will revisit the charter and draft a revision.

 

    3.6 ACTION: Andrzej will draft a preliminary statement of the acronyms issues for the
 DITA 1.2 issues list.

 

        CONTINUED

 

    3.7 Rodolfo will continue to work on the Best Practice for XLIFF

 

    CONTINUED

 

    3.8 Gershon will prepare the multilanguage and the other Best Practices in the OASIS

    template for the TC approval.

 

4) Returning business:

 

    4.1 Approved the new draft of the Multi-language Best Practice (Nancy Harrison)

 

      Action completed to approve this Best Practice. Rodolfo moved, Nick seconded.

      Gershon to add examples and prepare the complete set for this TC approval. Include the Chinese

      examples which are not copyrighted. These are in the public domain.

 

    4.2 Reviewed Andrzej's acronym proposal. Andrzej will revise

 

       Issue of acronym's for inflected language. Possible problem in typing the acronym. Consider too 

        much typing for authors. Consider problem of translated text in the attribute.

 

    Notes from Kara Warburton

Using the acronym itself as the id attribute is not a good idea since it may not be unique. Many acronyms have multiple different expanded forms, even within the same domain. For instance, in computing, ACL can stand for both access control list, and application connectivity link. When you are dealing with multiple domains, there are likely to be more than two expanded forms.

Also, how does this relate to the <term> element that already exists in DITA? An acronym is a type of term, so I would like to see these two items handled together somehow, such as through a "type" attribute on the <term> element, that attribute using one of the values in TBX-lite.

Notes from Richard Ischida

I just skimmed this, and thought I should point out a couple more things to consider. Note that acronyms also change in translation, eg. NATO becomes OTAN in Spanish. Another good reason to avoid using the acronym as the id.

Also, it would be useful to search the public w3c html mailing lists (going back a few years) for discussions about acronyms and abbreviations. Things you'll find there include discussions about whether it makes sense to separate out acronyms and abbreviations - ie. can people always decide which is which. And Web Accessibility requirements that lean towards a need for a pronunciation guide that is separate from the expanded form, eg. for clarifying that CSAIL (Computer Science and Artificial Intelligence Laboratory) is pronounced see-sail, not C.S.A.I.L.

    4.3 Rodolfo proposed adding to the Indexing best practice. 

 

    Need to account for languages with multiple character sets. Japanese should be sorted by the Hiragana

    alphabet, based on pronunciation, not the Unicode ID.

 

    Rodolfo noted that translators must type in the sort order in Hiragana. However, the translator might

    not always know the correct Hirgana pronunciation for the Kanji or Katekana characters.

 

     Human entry is important. You cannot achieve good results with automatic processes alone.

 

    Andrzej -- noted other languages with similar problems for indexing. Serbian, Croatian, Adzerbajani,

    Albanian, Mongolian, Georgian, Caucasian, perhaps others.

 

    The text must be sorted in multiple orders.

 

    JoAnn will try writing the addition to the indexing Best Practice for review.

    4.3 Review Christian Lieske's article on XML, DITA, and Translation requirements.
        
    4.4 Discussion of new issues to explore for the committee going forward.

 

--Meeting adjourned at 9:00--

 

JoAnn T. Hackos, PhD
President
Comtech Services, Inc.
710 Kipling Street, Suite 400
Denver CO 80215
303-232-7586
joann.hackos@comtech-serv.com