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

RE: experimental release of linuxdoc-tools (based on sgml-tools 1 .0.9)

ok, I've snipped a few things from this cross-post, I'm tired of getting 8
copies of everything.  :-)  Everybody must be on one of these lists, since
they replied at some point.

> -----Original Message-----
> From: cg@cdegroot.com [mailto:cg@cdegroot.com]
> Sent: Wednesday, May 17, 2000 12:10 PM
> To: Gregory Leblanc
> Cc: 'sgml-tools@via.ecp.fr'; Greg Ferguson; Taketoshi Sano;
> ldp-discuss@lists.debian.org; ldp-discuss@lists.linuxdoc.org;
> dave@lafn.org; guylhemlistes@free.fr; gferg@metalab.unc.edu;
> debian-sgml@lists.debian.org; JF@linux.or.jp
> Subject: Re: experimental release of linuxdoc-tools (based on 
> sgml-tools
> 1 .0.9)
> GLeblanc@cu-portland.edu said:
> > It's probably true that creating a DSSSL stylesheet for Linuxdoc
> > wouldn't be all that complex, but I'm still a little troubled by the
> > lack of a good text backend.  I haven't seen any 
> mass-migration of the
> > LDP authors to DocBook yet, so I haven't really pushed to get
> > something better than Lynx.  
> Does the term "chicken-and-egg problem" ring a bell? 

Yeah, but I'm not sure how it applies here.  We know that we need better
processing into text from DocBook, and at least some of the LDP authors are
using it.

> There's a better backend-processing candidate for the text 
> stuff (html2text, 
> IIRC), and it's even on the SourceForge SGMLtools-Lite 
> project page as a bug 
> report that this must be implemented. I think that grabbing 
> this bug would be 
> a more useful passtime than writing new stuff for Linuxdoc.

I took a look at that, and I'm pretty sure that my skills aren't up to
figuring out how to do the configure stuff (that is one UGLY file), which is
the only place that real work is required for integrating this into the
sgml-tools project.  I don't think I proposed writing anything new for
Linuxdoc, since the support for text there is decent (except for a bug that
I think has to do with deeply nested sections, I haven't been able to track
it down).  I haven't looked at how well html2text does, and since tables
aren't the only thing that lynx does poorly, it's worth looking at.  URLs
are the other thing that comes to mind, they really ought to be printed out
full text. Later,

To UNSUBSCRIBE, email to ldp-discuss-request@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org