XSL/XML and XSLT and/or Docbook

[Jeff Garland]

Craig White wrote:
> Been struggling for a while with this and am in a quandry.
> 
> What I have is a manual of several hundred Microsoft Word documents
> linked together only by subdirectories or a 3 ring binder. These
> documents must be converted no matter what because their formatting
> falls apart, even if moved to Macintosh OSX and Microsoft Word.
> 
> What I have identified so far is that we would like to be able to have
> these documents available in various forms...
> - html (xhtml is probably more likely)
> - pdf as a complete set that we could print with a single command.
> - pdf subsections since some of the sections are used 'stand alone'
> 
> Thus the little knowledge that I posses told me that I could convert
> these documents into Docbook format and then use either dsssl or xslt to
> process the xml into the various formats that I might want to use.
> 
> My experimentation with this leaves something to be desired though
> because the maintainable data structures in XML don't seem to be very
> friendly for administrative staff to edit/maintain.
 >
> I also have hear from Joeseph that docbook format is clunky and
> uncooperative and thus far, my experiences with using docbook2html /
> docbook2pdf vs. xsltproc haven't been too dissimilar to convince me that
> either is a better way but I was hoping to hang on to docbook just in
> the hopes that I could have administrative staff use something like OOo
> as their editor so as not to completely scare them off.


Having some experience maintaining technical documentation in "BoostBook" 
which is basically DocBook with a few little add-ons specifically for C++ 
docs, I'd have to agree the editing and maintaining DocBook is painful.  Of 
course I use emacs to do it, but I've tried other 'GUI editors' and none of 
the do much of use.  I haven't tried OO v2, so maybe it can do something more. 
  We've also had lots of trouble with the pdf rendering not working real well.

> I am almost thinking that I might have to simply eschew the common word
> processor as the data tool and end up coding a complete application in
> something like ruby on rails to output the xml bits to ultimately be
> assembled into final form by xsltproc or other and am wondering...my
> question for the list...
> 
> Am I missing something or are my only choices to get something like this
> done is through a web based application that takes the user supplied
> data, writes out xml and then post processes via xsltproc or simply 
> use 'master pages' and assemble books from linking all the various OOo
> documents together? - which of course is gonna leave me in a world of
> hurt for html.

Not that I'm aware.  Docbook sounds good in concept, but it seems mostly like 
a geek thing at this point.

I was thinking you might be able to convert to Open Office 2 as it has both 
xhtml and pdf export.  Just trying it before I said it in public, though, I 
notice that my version seg faults on xhtml output...ymmmv.  But in any case 
the storage for open office is xml so in theory you could write your own 
backend if need be.

Jeff