Under Community Review

Thanks  for logging this - could you perhaps see if you can provide this case in English also? This would make it easier to handle for us and also to vote on for other users potentially. Thanks!

ターゲットファイルに余計なDITAタグを追加しない様にして頂けないでしょうか。　Please don't add extra DITA tags in a target file.

Trados Studioで以下のDITA形式の和文を翻訳して英文版を作成し、xmlファイルを出力すると、元の和文ソースには無かった<menucascade></menucascade>のタグが追加されます。

●和文

<cmd><menucascade>
<uicontrol>システムとセキュリティ</uicontrol>
<uicontrol>Windowsファイアウォール</uicontrol></menucascade>を選択してください。
</cmd>

●英文

<cmd>Select
<menucascade>
<uicontrol>System and Security</uicontrol>
<uicontrol>Windows Firewall</uicontrol></menucascade>.
</cmd>

●出力された英文xmlファイル（余計なmenucascadeタグが追加される）

<cmd><menucascade></menucascade>Select
<menucascade>
<uicontrol>System and Security</uicontrol>
<uicontrol>Windows Firewall</uicontrol></menucascade>.
</cmd>

テクニカルサポートからのご回答を頂きましたが、こちらは仕様とのご説明でした。

以下の理由でこの仕様はおかしいのではないかと感じ、ご意見させていただいています。

1. 中身が空の<menucascade></menucascade>というタグの使い方はDITAの文法違反であり、Tradosを使うと文法違反のソースが出力されるという仕様は疑問に感じる。

2. オリジナルのタグの数と合わなくなる。（DITAで翻訳する際は、ソースとターゲットのタグの数を一致させるのが常識と思います。Tradosを使うとターゲットのタグ数が増減するという動作はおかしいです。）

3. 和文が動詞節を後置しても、英文は動詞を先頭に配置する訳が自然であり、この対訳が正しく出力できるよう考慮してしかるべきと感じる。

是非、この仕様の改善をお願いします。

Naomi Hatano

  • Naomi Hatano-san, thank you for logging this problem with RWS.
    We are also having the same problem and ask for RWS to check the automated command when creating a translation kit.
    This seems to be a system error or might be a bug.

    The error occurred from the very beginning when generating a translation kit.

    The translation kit code:
    </source></trans-unit><trans-unit id="74cb3d82-cb66-4a19-9831-53e6cf34b3cb">
    <source><g id="89"></g><g id="89"><x id="90"/>検索</g>(8)ボタンをクリックします。</source>
    <    seg-source><g id="89"></g><mrk mtype="seg" mid="15"><g id="89"><x id="90"/>検索</g>(8)ボタンをクリックします。</mrk></seg-source>
    <target>    <g id="89"></g><mrk mtype="seg" mid="15"><g id="89"><x id="90"/></g></mrk></target>

    The return kit code:
    </source></trans-unit><trans-unit id="74cb3d82-cb66-4a19-9831-53e6cf34b3cb">
    <source><g id="89"></g><g id="89"><x id="90"/>検索</g>(8)ボタンをクリックします。</source>
    <seg-source>[This is a code that do not show up on Trados Studio but shows on the exported XML data and published PDF]<g id="89"></g>
    [This is what is shown in Trados studio]<mrk mtype="seg" mid="15"><g id="89"><x id="90"/>検索</g>(8)ボタンをクリックします。</mrk></seg-source>
    <target>[This is a code that do not show up on Trados Studio]    <g id="89"></g>
    [This is what is shown in Trados studio]    <mrk mtype="seg" mid="15">Click the <g id="89"><x id="90"/>Search</g> button (8).</mrk></target>

    DITA format:
    Source data:
    <cmd>
    <uicontrol>
    <image href="XXX" placement="inline" width="3.3mm" />検索</uicontrol>(8)ボタンをクリックします。</cmd>

    Exported Target data:
    <cmd>
    <uicontrol>
    </uicontrol>Click the <uicontrol><image href="XXX" placement="inline" width="3.3mm" />Search</uicontrol> button (8).</cmd>

    I believe this problem could occur to any sentence segment where the start tag and end tag are in different line.
    I am not sure if this occurred because the beginning of the segment start with a tag that uses indention or with a continues tag+indention.
    This problem could easily occur to source languages like Japanese where the main subject is written at the beginning of the sentence and need to be quoted.

    We hope that you could check the system and take actions to solve this problem.

    Kind regards to all.

  • Although this is a bit beyond my ken, I endorse the principle enunciated by Naomi that the number of tags in the target segments should match the number of tags in the source. That is, unless there is some special feature of DITA formatting that compels otherwise.

    With kind regards to you all,

    Adam Warren.

  • Thanks for your advice to my comment. I put the translated comment in the following.

    ---

    When we translate the following DITA format Japanese text in Trados Studio to create an English version and output the xml file, the <menucascade></menucascade> tags that were not in the original Japanese text source will be added.

    ● IN Japanese

    <cmd><menucascade>
    <uicontrol>システムとセキュリティ</uicontrol>
    <uicontrol>Windowsファイアウォール</uicontrol></menucascade>を選択してください。
    </cmd>

    ● IN English

    <cmd> Select
    <menucascade>
    <uicontrol> System and Security </uicontrol>
    <uicontrol> Windows Firewall </uicontrol></menucascade>.
    </cmd>

    ● Output English xml file (extra menucascade tags are added)

    <cmd><menucascade></menucascade>Select
    <menucascade>
    <uicontrol> System and Security </uicontrol>
    <uicontrol> Windows Firewall </uicontrol></menucascade>.
    </cmd>

    We received a reply from technical support that this is a specification.

    I feel that this specification is strange for the following reasons, and I would like to make an opinion.

    1. The usage of the empty <menucascade> </menucascade> tag-set is a DITA grammar violation, and the specification that Trados outputs the source of the grammar violation is not appropriate, I think.

    2. Trados does not keep the number of original tags. (When translating DITA document, it is common sense to match the number of tags of the source and the target. It is strange that the number of tags of the target increases or decreases when using Trados.)

    3. Even if the verb clause is placed at the last part of the Japanese sentence, it is natural that the verb is placed at the beginning part of the translated English sentence, and I feel that it should be taken into consideration so that this translation can be output correctly.

    We hope this specification be improved.