I suppose you're right about programmers not seeing the need for high
quality typesetting. This is just a disappointment for me, because I
do see it as important. The last update of FOP was 1 1/2 years ago,
and it doesn't seem like there is any rush to move it forward. I
understand that work on FOP may be very difficult; but it lacks many
things I need to make nice documents, such as widow and orphan
control.

LaTeX is an excellent system for typesetting, but I think the
*concept* behind it is inferior to FO. LaTeX uses what I call a text
type system, in which new lines and backslashes mean something in
contrast to reading a tree. I think that going from a tree (TEI for
example) to a tree (FO) makes a lot more sense than going from a tree
to text, in which you have to escape certain characters and watch out
for newlines. Plus, the syntax for FO will seem easier to understand
if you are working with XML--attributes and elements will make more
sense than {\i }.

My feeling --and I don't have hard proof--is that the open source
community is a bit infatuated with text type processing such as you
can achieve with perl. I don't think the open source community has
fully grasped the power of trees.

I was learning quite a bit of LaTeX before I discovered XML and XSLT,
and I abandon it for FOP, thinking it was just a matter of time before
FOP replaced LaTeX in open source publishing. But now it seems that
work is going so slow with FOP that it may never reach the quality of
LaTeX.

Given the status of FOP, I am willing to have another look at LaTeX.
How do you convert from XML to LaTeX? I assume you use PassiveTeX?
Any details would be greatly appreciated.

Paul


On Tue, Feb 15, 2005 at 11:02:30AM +0000, Peter Flynn wrote:
> Date:         Tue, 15 Feb 2005 11:02:30 +0000
> From: Peter Flynn <[log in to unmask]>
> Subject:      Re: tei tables -> fo -> pdf using fop
> To: [log in to unmask]
>
> On Mon, 2005-02-14 at 21:19, Paul Tremblay wrote:
> > On Mon, Feb 14, 2005 at 04:46:31PM +0000, Sebastian Rahtz wrote:
> > > >
> > > writing typesetting formatting engines is hard work,
> > > as the FOP folk have found. Its a good example
> > > of an area when the open source system paradigm
> > > has cut in, but to your disadvantage: there does
> > > not seem to exist a community which cares enough
> > > to make FOP as good as XEP. Some of the people
> > > who care are in the mad mad world of TeX instead :-}
> >
> > This obsession with TeX has really disappointed me.
>
> I'm not clear what the problem is here. TeX is just a tool which
> happens to solve some problems which other systems don't handle as
> conveniently. It has advantages and disadvantages, but at the
> moment the former outweigh the latter.
>
> > The open source community doesn't think PDF is important, I believe,
> > thinking that everything will go electronic and any print medium is
> > akin to cave-age thinking. (Or is there another reason for it?)
>
> I don't think they believe it to be unimportant per se, but IMHE most
> people in the Open Source programming field -- perfectly naturally --
> don't have any background in publishing, so they don't see high
> quality formatting as important (whether PDF, PostScript, HTML, XML,
> or even plain ASCII text).
>
> > There is a great document system called LyX, but it uses TeX as its
> > back end. If only we in the XML community had something has nice!
>
> Surely that should be "There is a great document system called LyX,
> *and* it uses TeX as its back end" :-)
>
> But seriously, I think you risk conflating the synchronous typographic
> interface with the underlying file markup format.
>
> LyX is very impressive, but it still has a large number of interface
> problems, and it doesn't save as standard LaTeX, so there are some
> portability problems.  The XML community has several commercial systems
> which are capable of producing publication-quality output, and have
> synchronous typographic interfaces -- that is, you make changes
> directly to the typographic appearance without the markup being visible
> (as with Word, for example) in the hope that these will be stored
> sanely, rationally, and sensibly in the underlying markup system.
>
> Unfortunately there is as yet no system capable of interpreting all
> the things people want visually and storing them in the "right" places
> in (for example) the underlying TEI file, because the visual paradigms
> currently in use have no mechanism for explaining *why* you want this
> word italicised.
>
> What You See Is What You Want requires some intellectual input, either
> direct from the user, or by inference and experience by the software
> (which will never be as good), or by a combination of the two (this
> is part of my current research).
>
> For the moment, I side with Sebastian. XSL:FO -- whatever engine --
> still requires the reinvention of too many wheels that LaTeX already
> possesses. Some of it (XEP is a good example) is getting there, but
> the more you try to cover up the specification of what gets formatted
> how with layer upon layer of visual interfaces, the more cumbersome
> the process ultimately becomes, as you try to substitute dexterity with
> the mouse for intellectual experience. It makes for good marketing,
> but it's aimed at the commercial customer who knows that mouse-pushers
> are cheap but markup-cum-formatting experts are expensive.
>
> You're therefore stuck at this stage with FOP on the one side, and
> XSLT-->pdfLaTeX on the other. As I explained earlier, I prefer the
> latter largely because I'm used to it. It's not hard: for someone who
> knows XML and XSLT the learning curve for LaTeX is nowhere near as
> steep as for a complete beginner with markup systems, and it's quite
> short -- days, not weeks. It's not as mad as Sebastian paints it :-)
>
> ///Peter

--

************************
*Paul Tremblay         *
*[log in to unmask]*
************************