Opentag / Localization

From: Marc van Woerkom <marc(at)netcologne.de>
Date: Sun, 15 Nov 1998 17:36:25 +0100 (CET)

(Message forwarded by Wolfram)

> Has anyone else (particularly any of the translation teams) had a look at
> OpenTag, <URL:http://www.opentag.org/>?

I just had a look but returned not much enlightened.
The example I saw, created two files from input, one with formating information,
one with the contents.
To be honest, I have no clue how this is supposed to help us.

Right now I take the original en-1.sgml, copy it to de-1.sgml and walk
through it, line by line, replacing English terms with German ones.

(Which itself is a good thing, for I need to know the approximate logical
 structure of the doc - thus an approach that simply extracts all messages and
 replaces them with msg numbers, leaving me with a just a big message pile
 to translate is bad)

The problems will begin, when someone changes the original English sgml file
into en-2.sgml:

    1. I would need to know that this update happened

    2. diff en-1.sgml en-2.sgml,
       resulting in a list of (pos, new bit of text) pairs of the English text

    3. find the corresponding changed en-1.sgml locations in de-1.sgml

    4. create a de-2.sgml, where I put in the new (English) text bits at
       those locations

    5. translate the new bits into German

Obviously, step 1 could be assisted by a Makefile with appropriate dependencies
which could also exec the diff for step 2 and further ops.

Step 3 is more tricky. How can one relate a position between the two
language versions?

When saying position, it is not about a (line, column) pair - this formatting
will change during translation, but this should be a logical position,
like the 105th sgml tag, 3rd sentence.

Guess I wish a process that acts like this:

    (cvs update)
    cd /usr/doc/doc/de
    make update
    (cogwheels turn)

and then find a file de-2.sgml that looks like this:

    <!- de-2.sgml translated from en-2.sgml ->

    Dies ist ein bereits in de-1.sgml enthaltenes Stück Text.
    
    <!-
    UPDATE 981115
    ------------------------------
    This is some new english text (part of the diff to en-1.sgml)
    with lots of interesting new stuff
    or subtle changes that should be put in exactly here.
    ------------------------------
    -->

    Und hier gehts ganz normal weiter wie vorher ..

    <!- end of de-2.smgl ->
    

A little bit of perl should the trick.. what do you think
about it?

I'll upload a demo implementation on the German translation ftp site soon.

Regards,
Marc
Received on Sun 15 Nov 1998 - 17:35:20 CET

search this site