[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: libpkg-guide status ?


On Sat, May 20, 2006 at 11:01:32AM -0500, W. Borgert wrote:
> On Sat, May 20, 2006 at 06:28:11AM +0900, Osamu Aoki wrote:
> > These are in debiandoc-SGML, while yours is in docbook-XML. Basically
> > XML framework is bettwe since it has table etc.
> >
> > No one took effort to convert these to XML yet.

But I and Phillipe took effort to create conversion script with 2
different approaches.

> > So mixing with these are non-trivial task.
> I'm willing to do that conversion. Could you send me the CVS/SVN
> address of the document, please?



There were good efforts on conversion script.
 * debiandoc-sgml package:  debiandoc2xml command.
 * debiandoc2dbxml package: xslt template with perl

I documented this in the appendix of debiandoc-sgml-doc package as much
as I did.


These should give you a good start.  

I tested these tools on the SGML files long time ago and looked all
right.  I did not do real conversion mainly because I was not familiar
with creating XML build script.  That is why documentation stop at
validation.  Conversion itself is almost solved problem. 

Here is what I think is needed. First step is to run doc source under
one of these script.  If you find bad behavior of conversion script, bug
me or phillipe.  Once you get successful source conversion following
what I wrote to escape references to the default.ent contents, you can
provide this sequence into a synchronization script.  Then you may want
to fine tune XML using xslt etc.

Once you script all these process, then start synchronized XML tree with
upstream SGML.  Then ask people to switch :-)

Oh, when you strat XML source, I suggest to have source split in at
least by chapters (preferably by sect1).  So version tracking is easy.
(We need to force version symbol usage more)

Although I document "-1" option use only for debiandoc2xml, I think
debiandoc2xml invoked without -1 should work to make chapter-by-chapter
XML files.  (Less tested)  Any shortcoming such as use of tag, may be
accomodated by bugging us :-)

As you mich imagine, my XML conversion script is nothing-but extension of
HTML conversion tool inherited from ardo.  That is why main file is
index.xml.  It has been a while since I made these.  So my memory is a
bit vague.


PS: Tough part is version tracking.  I guess simplest thing is to save
diff of original english version and translation with tools provided in
the SGML side.

Reply to: