Skip to main content
B
18-Opal
bfriesen18-Opal
18-Opal
November 15, 2011
Question

Translation: The ins and outs.

  • November 15, 2011
  • 2 replies
  • 3781 views

We are very new to arbortext and very new to the thought of translation. We are creating our manuals in english currently and we are wanting to go to Russian, Porchagese, and Spanish. I am looking for someone with experience, how the translation process works? Is there a document or something that I can review that will give me some options of how this is supposed to work. Do we need to have multiple dita maps?, What happens when the original language is modiifed, do you need to keep track so the same sections in the translated book? Does it do it automatically? Lots of questions, few answers.

Thanks for your time,

Bryon Friesen

    2 replies

    J
    JeffStevenson1-Visitor
    1-Visitor
    November 22, 2011

    What you would probably need is a globalization or translation management system. A GMS is able to track changes to the source language and store translations. It can also create translation projects for new and modified content. Arbortext and Windchill can integrate with these systems although some more closely than others. Typically, a GMS will be able to use XML as a form of transporting content which makes it able to integrate with Arbortext.

    If you are translating a lot of content, it is probably best to do an ongoing translation process where translation is occurring in parallel to content updates. Otherwise, you will need to translate large batches of content periodically.

    Hope that helps!

    R
    RayStachowiak1-Visitor
    1-Visitor
    December 13, 2011

    Interesting question. Arbortext is an editor and creates content, in your case English content. If you are doing DITA what you end up with is English modules (DITA Topics) and a DITA Map. Now you want to go to Russian, Portuguese and Spanish. You need to run this English data now through either a machine translation logic engine or a translation memory system or send it out for translation. You would send out the DITA Map and the modules and you would get back a translated set of modules and a map in the desired language. You would store this in a directory on your system. In the end you would have 4 streams of modules. When an English module changes you would send that module and the associated module from each language to your translator and have it updated. How you manage versions could be an issue unless you don't and just store the latest of each module, but a version number will insure that you translate all the modules that changed during an authoring period. If you have 10 modules, easy to keep track, but if you have a thousand modules, making sure that all the versions have been translated in all the languages starts to become an effort which is why you should look at a CMS to help manage the DITA modules and translations. The CMS can maintain versions, changes, insure all language versions are changed when an English version changes. Without a CMS this can get unwieldy very fast. CMS's are available as SaaS (Cloud) applications which you can subscribe to so that you can get into the process at a reasonable cost and then move into an installed system at a later date. Hope this helps.

    B
    18-Opal
    bfriesen18-OpalAuthor
    18-Opal
    January 20, 2012

    Thanks for the responses so far. This is still pretty fuzzy.

    To see if I am understaning

    We are authoring in English creating dita files, we currently do not have a content management system in place (IT is supposedly working on this). In our situtation, we would need to have a russian version of the ditamap and all modules (dita topics). We then send this whole folder out to the translators (ditamap and all modules) or do we need to save the file off as something else to send to the translator? This only needs to be done the first time for each language? When I get the files back, I would need to overwrite the original folder to give me an actual Russian manual. When I go to publish will all my generated text be translated also, Figure, table etc...? Do they also translate all of our DWCS headings? I thought these are buried in the DTD.

    So I would now have a ditamap and modules that has been translated into Russian, when I go to publish this ditamap am I going to need a my stylesheets modified because of the length of words potentially.

    When revisions are requried I will manually have to keep track of what is needing to modified, do the dita topics have a version tracking ability that can help me?

    Thanks

    Bryon

    B
    18-Opal
    bfriesen18-OpalAuthor
    18-Opal
    January 22, 2013

    I am back at this translation topic again. The process I have figured out so far is

    Open styler, go to tools, language properties
    Under Generated text , set your target languages, by adding the languages you want to translate to.
    Under Document Language mapping, Click Add, set document language and language to use, to same language. Add the same ones that you added to the generated text.

    Export Generated Text
    To export generated text for translation:

    1. In Arbortext Styler, select the Tools ▶ Export Generated Text menu option to open the Export Generated Text for Translation dialog box.
    The menu option is not available if you have not specified any target languages for generated text.

    2. In the dialog box, provide the information required to set up the XLIFF files.
    If you are re-exporting a stylesheet’s generated text for translation, you may wish to export only those translation units that have changed since the last export. Select the Include only generated text that does not have a current translation option.
    Click OK to confirm the settings and begin the export.

    3. Your XLIFF file(s) will be exported to the requested destination. If your stylesheet contains no translations, you will see a notification message and the file will not be created.

    4. Open the XLIFF file for the desired language output search for <target state=”new”>. Change the text in between these tags. Then change the “new” to “Translated”.

    5. Locate the XLF file and use xmsgfmt to convert it to a .amo file. (found under styler help "Localizing Generated Text") (not sure if you actually need to do this or not).

    6. Place both the .amo and .xlf files in a locale directory associated with the language. (not sure if this should be in my com.arbortexst.sma folder or if it should be under program files, PTC... locale\ folder).

    7. Using styler, import the .xlf file(s). (Do I need to do this to both stylesheets if we have a editor and a print version?)

    At this point I think that I am well on my way. In my ditamap I set the language value at the top level to ru. It then shows up for all the topics in my ditamap. I open the first topic and all my generated text is still english. I have found that if I reopen styler and preview in editor button that sometimes rhe hazard statements change and sometimes the title for the images changes.

    Thanks

    Bryon

    Terms & ConditionsAccessibility statement