Re: [lojban] Re: The CLL project, technical directions

[Robin Lee Powell <rlpowell@digitalkingdom.org>]

On Sat, Aug 09, 2014 at 05:44:17AM -0700, TR NS wrote:
> 
> On Thursday, August 7, 2014 9:19:44 PM UTC-4, Robin Powell wrote:
> >
> >
> > So let me start off by saying that pandoc is certainly not out of 
> > the question and you're welcome to try to do something like that.  I 
> > played with it myself for a while. 
> >
>
> Would it be possible for me to get a copy of the source material,
> even just a chapter or two portion, to work with? 

Oh, yes, absolutely.

The current state of affairs is on github, and works as-is.  The
problem is that the current state of affairs uses XSLT *heavily*,
really just because docbook uses that and so I followed along.

It turns out that unlike Ruby or Python or Haskell or to a certain
extent Perl, where people in the Lojban community might
theoretically help out, *NO-ONE* will touch XSLT.

Having done a bunch of XSLT work, frankly, *neither will I*.  At
least, I won't ever touch it again.

Hence all this back and forth.

By the way, our docbook is heavily modified, things like the cmavo
tables take like a billion tags in docbook, so I basically made
macros.  That's what the XSLT is for.  I still wouldn't use XSLT if
I had it to do over.

https://github.com/lojban/cll

Note that the toolchain is a bit harsh; find me in #jbopre on
freenode if you want an account on the server that already has it
all.

The Ruby branch is origin/docbook-ruby , but it's just way not
ready.  Soon.  Keep bugging me, please, I really mean it.

> > Teaching a system that has no concept of an index whatsoever how to 
> > render one means thoroughly learning a new system that, at the time, 
> > I had no indication that anyone else would be familiar with or be 
> > able to help with.  That sounds tedious and upleasent to me, just 
> > because of the time that would be required to learn pandoc. 
> >
> > In addition to my primary goal being accessibility for other coders 
> > and the specific issue of the index, though, it's not just the 
> > index. 
> >
> > http://vrici.lojban.org/~rlpowell/media/public/pandoc_vs_docbook/ 
> >
> > Look at the HTML, specifically the "I sell this-thing/these-things 
> > to that-buyer/those-buyers." example.  I don't know why the tables 
> > align differently; I'd have to go figure that out.  The examples 
> > have no numbers.  etc, etc, etc; there's a ton of work there 
> >
> 
> The HTML shouldn't be too hard to fix, at least mostly. I am more concerned 
> about the PDF output though, that a bit harder to fix.and pass
> them through dblatex with special tags, which pandoc doesn't know
> about.

At least some of that is the fact that the PDF table management was
*so bad* with dblatex that I actually write the tables in LaTeX and
pass them through dblatex with special tags, which pandoc doesn't
know about, so raw LaTeX is arriving in the output.

> > I have no idea how much work it will take to teach pandoc
> > everything I need to teach it.  I have a pretty good idea how
> > long the Ruby solution will take, though (I'd say on the order
> > of a hundred hours). 
> >
> 
> Ok. Well, maybe that is best choice at this time then. I am a
> (very good) Ruby programmer so I may be able to help a bit there
> too.
> 
> But I would like to see how far I can get the Pandoc solution to
> work.

I'm totally OK with that.

> Thing is, I have my own book to publish eventually. So I have an
> ulterior motive of perfecting a publishing platform for myself as
> well, and I am absolutely adamant about KISS. Using the LLC as
> base material for working that out should be perfect b/c it
> contain many of the same kinds of constructs I will need
> (examples, tables, index, etc). 
> 
> If I can't get the Pandoc to work, or find something similar, than
> I may just join your Ruby-platform alliance ;-)

Nice to have someone motivated.  :)  I must say I'm *quite* startled
by how hard this (multi-format book output using only FOSS tools)
is.  If you want to stick to free tools and you don't want to
descend in the mire that is docbook-to-pdf, it appears that Sphinx
and Pandoc are really the only two choices, and neither of them do
serious books out of the box (i.e., multiple indexes).

The reason I didn't do Sphinx, by the way, is that I would have had
to write a converter from our modified version of docbook too
Sphinx's markdown(?).  I figured if I was going to have to write a
converter for *every* tag *anyway*...

-- 
http://intelligence.org/ :  Our last, best hope for a fantastic future.
.i ko na cpedu lo nu stidi vau loi jbopre .i dafsku lu na go'i li'u .e
lu go'i li'u .i ji'a go'i lu na'e go'i li'u .e lu go'i na'i li'u .e
lu no'e go'i li'u .e lu to'e go'i li'u .e lu lo mamta be do cu sofybakni li'u

-- 
You received this message because you are subscribed to the Google Groups "lojban" group.
To unsubscribe from this group and stop receiving emails from it, send an email to lojban+unsubscribe@googlegroups.com.
To post to this group, send email to lojban@googlegroups.com.
Visit this group at http://groups.google.com/group/lojban.
For more options, visit https://groups.google.com/d/optout.