[publican-list] Possible alternative to FOP!

[Jeff Fearn]

On 08/22/2011 11:31 PM, Raphael Hertzog wrote:
> Hi,
>
> On Mon, 22 Aug 2011, Jeff Fearn wrote:
>> It's not nearly as annoying as the limitations FOP has, such as no
>> complex text PDFs, and the fact that the current release is
>> basically un-packagable.
>
> Hum, Debian still has 0.95 (and not 1.0). I don't know the reasons but at
> least it gives some plausibility to your statement. :-)
>
>>> The PDF should be a high quality rendering so that the result is perfect
>>> should it be printed as a real book (via a print-on-demand service for
>>> example).
>>
>> This was the original goal of the PDF, it was not attainable, and it
>> is not what PDF consumers use it for. The PDF is now aimed at being
>> a single file distributable, which closely resembles the other
>> outputs.
>
> BTW epub is also a "single-file distributable" (even if it's a just a zip
> file).
>
>> Perhaps sometime in the future we will consider another output for
>> this use case.
>
> I would definitely love this.
>
>> The reason we are switching is because FOP is unmaintainable, if you
>> want to step up and maintain FOP then I'm happy to stick with it.
>> Unless someone steps up and maintains FOP we will be looking to
>> switch, even if we lose some functionality.
>
> I don't think I'm going to step up to maintain the FOP backend but I'm
> definitely interested in a PDF that is of book print-quality and if the
> switch picks a new technology where this is possible I _might_ contribute
> a bit (or pay someone to contribute).
>
>>> Most of the people doing docbook use some sort of LaTeX based backend for
>>> the PDF generation. And it's also what many people expect when it comes to
>>> create a real book.
>>
>> I've never seen any even remotely decent docbook->latex tools, feel
>> free to drop some names and some links if you know of any.
>
> According to my limited experience, dblatex seems to be the most
> popular XSLT stylesheets: http://dblatex.sourceforge.net/
>
> Did you consider it? If yes, what were the main problems?

Yes we looked at it, 5 years ago, at that time it

1: couldn't cope with the complex documents we tested it on, trying to 
build them generated thousands of errors

2: we couldn't get simple documents to work reliably in CJK and Indic 
languages

3: it wasn't easy to base a distributable on it for Windows

4: the style wasn't anywhere near as good as the output you get from the 
XML::FO generated by the default style sheets

Maybe we could give http://docbookpublishing.com/ a whirl :P

Cheers, Jeff.