Talk:Manual Generation 3.0

From Gramps
Revision as of 08:16, 26 January 2008 by Erikdr (Talk | contribs) (What is about previous internal links {xref} ?)

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search

Is it usefull to still work with gramps.pot?

We could have many people translating the wiki manual, and then create different manuals from that. The alternative of translating gramps.pot is in theory better, but means we depend very much on one person doing the translation.

Like current data on wiki, we could use a XSLT script for retrieving data for localized version (de, fr, hu, nb, nl, ru, sk) with our docbooks, but we need to follow the english template and find a new way for tracking changes.

Jerome, you can track changes to english text in wiki with the history. Just select present text, and the last version you translated. You will obtain a list of all changes.
Is that good enough you think? bmcage 04:17, 31 August 2007 (EDT)
Previously (before gettext file), I used diff on XML for having full english model translation. I am not certain that all handbooks could be brought closer as easily as with tools for merge and translations memory (TM), offered by po editors. But I agree that maintenance will be well simplified. Should I paste localized versions ? I made test on i.e ? What could be do for image filename on wiki ? a directory (de, fr, no, nl ...) ? --romjerome 31 August 2007 (EDT)
Correct, po tools are probably much better. However, for manuals, how much can such a tool (TM) do to translate? Is the fact you see a wiki with all text not better to understand the whole view of the text? I would do localized versions like you did in the test. We could wait however until the English version for 3.0 is finished. I think much will change, especially the order of chapters/paragraphs, and that is something difficult to change later. So we could do new english version-> tags2text to gramps.xml, use po translate tools with old language files, then convert this with your script to mediawiki, and load it.
For images we will need to download the images to directory, and change root directory, that is true. Note that this can probably also be somewhat be automized, eg [1] to help/C/figures/Mainwin.png, and for french we would need [2] and download to help/fr/figures/Mainwin.png. What do you think? Perhaps an ftp script to gramps-project.org can be written for that.... or we can just do it manually.
Yes, working on English version and pasting locale data later will be a good start for localized manuals. I made a quick test with one of the current biggest manual page. I need to make minors changes on XSLT script for working with current mediawiki templates, I will wait the last wiki version. I should be able to paste localized manuals (new untranslated strings will be in english, just need a translator).
For images, I never try lftp or gnome-vfs scripts for ftp ... We might try FileZilla3 or gftp.

What is about previous internal links {xref} ?

We have internal links on current docbook/manual using xref. How to transcribe this on wiki for next manual generation ?

I have the impression xref is only for figure reference. Eg: here we need link to the events-fig figure: here.
If we create a new template {{man xref|pos|xref}} then this can be used to link to the image page, while xref can be what we need to use in the extraction script?
To make it complete, inclusion of figures should also happen with a template: so we change [[Image:events.png|right|thumb|150px|Fig. 2.x Events view]] into {{man includeimg|Image:events.png|right|thumb|150px|Fig. 2.x Events view|xref}}
Like that we know how to extract the data of the image into docbook.
What do you think? bmcage 04:33, 31 August 2007 (EDT)
There is reference on text too, see the section called “Entering and Editing Data: Complete Description”. txt2tags allows previous and post process/macros, we can look at example for improving the extract. Also, the localized image need to be renamed or we need a directory ? romjerome, 31 August 2007 (EDT)
Every section should be given an id, so indeed internal text references should be handled in however works best for us.
If the template for images should be introduced as stated above, pls. give erikdr a notification. The Python parser program will otherwise no longer detect the image files --Erikdr 10:16, 26 January 2008 (EST).