From kde-i18n-doc  Wed Feb 28 18:00:44 2001
From: "Erik K. Pedersen" <erik () binghamton ! edu>
Date: Wed, 28 Feb 2001 18:00:44 +0000
To: kde-i18n-doc
Subject: Re: poxml ready for prime time
X-MARC-Message: https://marc.info/?l=kde-i18n-doc&m=98338335721423

Did you move it to kdesdk? kdenonbeta is so unwieldy.

Erik

Onsdag 28 Februar 2001 11:16 skrev Stephan Kulow:
> Hi!
>
> I think I implemented poxml now so that it can go out to the
> masses. So I need now feedback on the solution I found. The
> porting to po files is a bit touch if the translation doesn't
> fit 100% (and I guess none does), but it's managable. But what
> I need feedback for is the kind of splitting. Lauri once posted
> a test case which I took as reference:
>  <para>
>   <variablelist>
>     <varlistentry>
>       <term>
>          Text 1
>       </term>
>       <listitem>
>          <para id="blah">
>             Text 2
>            <note>
>              <para>
>                Text 3 "everything"
>                <footnote>
>                   <para>
>                      Text 4 \"even more\"
>                   </para>
>                </footnote>
>              </para>
>            </note>
>          </para>
>       </listitem>
>     </varlistentry>
>   </variablelist>
>  </para>
>
> I have a list of single tags and a list of cutting tags.
> cutting tags are things like para, note, caution, etc.
> White space is simply converted to a simple space -
> allowing for consequenceless (does this word exist? :)
> reformatting in the english docs.
>
> These build blocks (without the tags), then the single
> tags are removed from start and end. Then tags that only
> wrap the block are removed (above the variablelist and
> variableentry). Then it starts the whole process from
> the beginning on the resulting string (meaning building
> blocks, removing tags...).
> This all leads to the following:
>
> #: index.docbook:14
> msgid "Text 1"
> msgstr ""
>
> #: index.docbook:14
> msgid "Text 2"
> msgstr ""
>
> #: index.docbook:14
> msgid "Text 3 \"everything\""
> msgstr ""
>
> #: index.docbook:14
> msgid "Text 4 \\\"even more\\\""
> msgstr ""
>
> This is exactly what I think is the translatable part. I went for
> compactness over context as I think the context too often simply
> repeats and doesn't add any informations - correct me if I'm wrong.
>
> I checked the kwrite docu and the faq and both are very nicely
> translatable (and split is able to extract the po informations from
> the german translation after only some tweaking).
> Sometimes the paragraphs get too long to be overviewed easily, but
> then it's real text and it's still way easier than it was before
> and somewhen kbabel will be able to show differences in longer texts.
> kbabel also needs later some changes to not mark two entities as
> accelerator clash :)
>
> Frederik already changed KDE customization to allow easier migration
> from SGML to XML, so it's all on it's way. Please try to use it and
> give me feedback. Don't start relying on it though as we might change
> it as feedback arrives.
>
> Greetings, Stephan
> P.S. We have too many mailing lists :)