Re: [xsl] XSL:FO approach for facing-page translation

[Martin Holmes <mholmes@xxxxxxx>]

On 14-02-07 12:57 PM, Tony Graham wrote:
> On Fri, February 7, 2014 6:43 pm, Martin Holmes wrote:
> ...
> > I need to do something similar in PDF, using XSL:FO. The ideal would be
> > to have the original text on the verso page, with the equivalent
> > translation on the facing recto. But I can't figure out a practical
> > approach to this. For one thing, I don't see how I can render individual
> > pages from one text and then the other text in alternating fashion, and
>
> Flow maps from XSL 1.1: use different flows on recto and verso pages.

Unfortunately I'm using XEP, which doesn't appear to support flow-maps.

> > for another, I don't see any way to keep the two texts in sync. I can
> > disallow page-breaks within paragraphs, but I'd have to be
> > extraordinarily lucky for the two texts to end up breaking at the same
> > points every time; it would just take one slightly-shorter paragraph on
> > one side to put them out of sync by a paragraph. Such a strategy would
> > also be rather a paper-waster, because many pages would have a lot of
> > blank space at the bottom.
> >
> > Has anyone every done anything like this, and if so, do you have any
> > advice? I can imagine that it might be done in a horribly manual fashion
> > by trial and error, working page-by-page, but I really don't want to get
> > into that. I'm happy to pre-process the text multiple times before it
> > goes to the XSL:FO stage., Perhaps there are ways to measure (for
> > instance) how much space a paragraph will take, and then adjust
> > page-margins or spacing by small increments to preserve alignment
> > between the two texts, but I haven't seen examples of such an approach.
>
> See my Balisage 2013 talk, 'Decision making in XSL-FO formatting' [1].

I've just been looking at it -- very intriguing.

> The Print and Page Layout Community Group @ W3C [3] has an extension
> function for use with Saxon or Xalan and either FOP or Antenna House that
> runs the XSL-FO formatter within the XSLT transform so you can make
> decisions based on formatted sizes as you go.  That would give you the
> flexibility to format each paragraph pair in turn to work out their sizes,
> which could make things simpler or harder, depending on how you approach
> it.

This does seem very promising, but the rest of the web application I'm 
working with is dependent on XEP, and I don't think FOP yet supports 
some of the features I'm depending on XEP for (although it's a good 
while since I've done a detailed test of this). I have no funding to buy 
into the AH Formatter (and the price is frankly jaw-dropping for an 
academic user) so it would have to be FOP.

I'm beginning to wonder about a far cruder approach: render both 
languages and see which is longer, then implement hard page-breaks in 
both versions at the point the longer one would break. Then run XSLT to 
interleave the two versions to create a continuous document. Very 
fragile, of course, and likely to require a lot of tinkering, but it 
would work.

Cheers,
Martin

> Regards,
>
>
> Tony Graham                                         tgraham@xxxxxxxxxx
> Consultant                                       http://www.mentea.net
> Chair, Print and Page Layout Community Group @ W3C    XML Guild member
>    --  --  --  --  --  --  --  --  --  --  --  --  --  --  --  --  --
> Mentea       XML, XSL-FO and XSLT consulting, training and programming
>
> [1]
> http://www.balisage.net/Proceedings/vol10/html/Graham01/BalisageVol10-Graham01.html
> [2] http://www.w3.org/community/ppl/wiki/FOPRunXSLTExt
> [3] http://www.w3.org/community/ppl/