[prev in list] [next in list] [prev in thread] [next in thread] 

List:       kde-i18n-doc
Subject:    Re: Generating docs for kdeextragear
From:       Lauri Watts <lauri () kde ! org>
Date:       2005-06-21 7:09:29
Message-ID: 200506210909.33772.lauri () kde ! org
[Download RAW message or body]


On Tuesday 21 June 2005 08.26, Krzysztof Lichota wrote:
> Nicolas Goutte wrote:
> > On Tuesday 21 June 2005 00:17, Krzysztof Lichota wrote:

> Step 2 (compiling) - translated docbook is compiled using meinproc into
> index.cache.bz2 and docbooks/images/index.cache is installed in system.
> This step is done by user, who compiles and installs application in the
> system.
> Step 3 (displaying) - khelpcenter transforms docbooks using XSLT into
> HTML and displays them.

Not quite.

Step 2, during compile, the docbook is converted *by meinproc* with xslt to 
html, which is what is in the index.cache.bz2.  This is what khelpcenter 
displays.   KHelpcenter *can* do the conversion, but it's slow and it's a 
last resort, only if the index.cache.bz2 is not present or not usable, or 
there have been changes requiring an update.

Step 3 is a fallback, if you are seeing khelpcenter regenerate docs every 
single time, then there is something going wrong.
>
> Entities are needed in step 2 and 3. I would like to move expanding them
> into step 1, so that they are not necessary later.

It's really not that easy.  

> > - write again author names, emails etc in full instead of using an entity
> > for such critical docs. (However this is also again going back to the old
> > KDE behaviour.)
>
> This is rather radical. I would prefer to have entites expanded
> automatically during generation of docbooks, as I described above.

Except, there is no generation of the docbook, as I explained.  The only point 
you could expand the entities is during po -> xml conversion.  And you only 
need write out the *new* entities, the ones that have been added since 
whatever version of kdelibs the hypothetical author is supporting.

> > - the translation should have the possibility to
> > manipulate the !DOCTYPE, so that translators can add entities there. (As
> > far as I know, this is not possible currently.)

This is a much better solution.

> But it means we would have to add all possible entites to headers of all
> translated docbooks? It would be a lot of work to keep it consistent
> everywhere. I think this was the reason why common entities were
> introduced.

No, it would mean you *could* include entities that are local to the document 
(and those that are for the purpose of supporting older kdelibs)

More to the point, a useful text editor can expand the entities (heck, a 
regexp and awk probably can.)  It's a very simple problem, the issue here is 
as much social as technical - we still have to know that it needs fixing, 
before it can be.  Expansion of entities not available in older kdelibs is a 
special case, the normal case does not require it at all.

Regards,
-- 
Lauri Watts
KDE Documentation: http://docs.kde.org
KDE on FreeBSD: http://freebsd.kde.org

[Attachment #3 (application/pgp-signature)]

[prev in list] [next in list] [prev in thread] [next in thread]