Re: DOCBOOK-APPS: Re: [Fwd: [Fwd: First Open Source DocumentationSummit at the O'Reilly Open Source Convention]]

[Sebastian Rahtz <sebastian.rahtz@computing-services.oxford.ac.uk>]

> 
> I don't mean to complain or to sound negative here, but maybe there's
> too much detail being left as an exercise for the reader ...  I'm just

sorry if thats my fault. 

> It seems to me there is a great deal more to rendering a page than to
> string fonts together into horizontal scan lines.  You need only
> compare the output from LaTeX with MsWord to see which one uses more
> typographic knowledge.

"knowledge"? TeX does what its told. of course it has many many more
variables to twiddle than MS Word, and it knows more about fonts.

>  An XML->PDF translator needs to account for
> hyphenation, ligatures, and all those things which are very mature in
> LaTeX.  Agreed, LaTeX is an awkward beast to port, but isn't it easier
> to port LaTeX than to re-invent it from scratch.

maybe. but the fact that its hard does not mean one should be scared of
it. TeX is there as a mine of algorithms, after all, to make it easier. we
cannot go on with Knuth's code for ever as it stands. its not as if its
perfect, after all - any journal typesetter will tell that TeX does not do
very good pagination (well, when I say "TeX", I mean "typical macro
packages")

> Ah ... ok, the shortest _affordable_ path ;) I gather from this that
> the actual answer is "there are no known paths for less than $500"

no, I didnt say that. i do XML to PDF every day (well, most days), using
TeX, at no cost

> I checked this out this afternoon and while it is a pretty nifty kit,
> the only way to use it is to reinvent hand-crafted DocBook stylesheets
> in "XMT"

yes, true. I never said the work had been done on the stylesheets for
Docbook :-} (now, if you used a Real Man's DTD, namely TEI...)

> ... I could not find any reference for transforming XSL to
> XMT
no, that would be hard work. i know of somone who can translate XMT to
XSL, though

> so it appears that xmltex could not, as it sits, parse _any_
> DocBook;
of course it can parse it. it will just not produce very elaborate output

> it has potential, but no work has been done to move in that
> direction.

not so. Michel Goossens at CERN has done a lot of work on Docbook .xmt
files, you could ask him to release them.

> For the curious, xmltex comes with an XMT file for TEI --- the docs
> were too scant to figure out how to install this

blame me for that. but there is no mystery, they are TeX input files. put
them where TeX can find them. on TeXLive (http://www.tug.org/texlive/),
this all just works off the shelf

 pdfxmltex foo.xml

spits out foo.pdf

> implemented".  Like JadeTeX, xmltex appears to be more a proof of
> concept than a general purpose tool.

i would take issue with that. it *is* a general purpose tool which you can
use in anger *IF* you accept the (high) price of a well-configured TeX
setup, and TeX error messages. xmltex, and PassiveTeX (which uses xmltex),
are (of course) largely undocumented. to that extent they are useless.

>     S> good tools, but they are not formatters. FOP is the formatter
>     S> and its not as good as TeX - yet. it may be one day. they
>     S> havent even started on hyphenation yet
> 
> This sounds like FOP is not a formatter either, but intends to be one
> someday.
now you are being picky. many many applications don't 
need hyphenation, and can use FOP as it stands for real work. if your work
needs math, and multi-lingual hyphenation, FOP is not for you. equally, if
you make newspapers, TeX is not for you. come to that, if you typeset
mixed Chinese and Arabic, I doubt if TeX is for you.

> I can live without RTF, but stylesheets do not generate files,
> formatters do.  Stylesheets are just specs to be given to a formatter
> to do the rendering and it appears as though there are no free
> formatters which read these XSL files, only a promise that FOP may
> someday do this. 

yes, PassiveTeX is a formatter which reads XSLFO files and formats
them. so is FOP. so is the RenderX tool, so is Arbortext's Epic. None of
them meet the spec of XSL FO, but then that isnt even finished. Similarly,
Jade does not implement DSSSL. few SGML parsers implement all of
SGML. many XML tools do not implement it properly.

> Since the DocBook XML files are (IIRC) XSL files and there are "no
> seriously reliable XSL FO" tools, then isn't the answer to "Can I
> print DocBook XML?" also "no"? 

you can print Docbook XML using Jade/DSSSL/JadeTeX. you can print it using
FOP or PassiveTeX or Epic. its just that you may not be able to express
every conceivable typographic nuance. sorry, but its not black and
white! if you sent me a Docbook XML file, I could "print" it. I could
probably not meet your arbitrary page design; perhaps nothing could.


> Or are the XSLT tools, which are said
> to not be formatters at all, sufficient to produce commercial quality
> output?  I'm confused.

define "commercial quality", in detail and exhaustively, and I could
answer the question. I have typeset a book for a commercial publisher
using XSL FO and TeX. does that prove the point?


> Last I was asked, admittedly some time ago, the Oxford University
> Press still gave LaTeX stylesheets. 
i dont count them as reputable :-}

> LaTeX or MsWord (some choice!!) --- all the _trade_ publishers I know
> only offer MsWord templates although I know of two (Some Pearson and
> ORA projects) who accept DocBook directly.  

right. so I dont see the urgency of LaTeX output from DB

> I have a problem coming up where I have a chapter written in DocBook
> to submit to a publisher who wants StarOffice ... the only paths I
> know are RTF or HTML, and it is questionable which of those two will
> be less lossy.

if the publisher cant sort it out, find a different one. seriously.

>     S> a high level Docbook to LaTeX transform really is a separate
>     S> project from the generic DB->print. doing it might be useful;
>     S> it would be trivial, once you decided what the mapping was
> 
> Trivial, but tedious: It doesn't sound like a weekend project to me ;)

perhaps not. but one could do 95% of it over a weekend.

> If it is trivial, though, and FOP is barely off the ground, then isn't
> the shortest path to print DocBook still DB/SGML->LaTeX->DVI?

no, the shortest route is Framemaker, or Epic; but you want free, I
expect. in which case a per-project DB to LaTeX would be quick and
efficient. IF LaTeX can do what you want. you can LaTeX can implemeny any
design? I dont.

> Is it true that there is no known way to generate quality print
> material from DocBook without expensive commercial tools, and no way
> at all to generate commercial quality print material from DocBook XML?

you need to define "quality" as well, now, please :-}

why do you regard XML and SGML as differemnt, in this context?

> Is jadetex another netscape navigator that we all hate to use but have
> no other alternative? ;)

god, I hope not!!! if it looks like that, I shall take my name off it, and
leave it in the public domain.

> Maybe I am just getting lost in acronyms again.  The real question I
> want to answer is must simpler: Given a DocBook XML source file, what
> are the shell commands to generate a printable output at least as good
> as DocBook SGML through jadetex?

XML and SGML work in the exactly the same way in this context. you need to
give Jade an XML declaration, of course, but maybe your tools do that
(dont ask me, I dont use them)

sorry if I seem to be eluding the issue, but whether TeX-based solutions
meet your need depends on what you those needs are. 

I typeset my daughters primary school newsletter using TeX, from an XML
master file. I translate it to TeX using XSL.  is that commercial?

Sebastian


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