Re: [docbook-apps] Do different xslt processors and parsers really make a difference for pdf output?

[Norm Tovey-Walsh <ndw@nwalsh.com>]

Kevin Dunn <kdunn@hsc.edu> writes:
> Thanks, Norm. Yes, I evaluated fop, xep, and ahf, and bought the
> Antenna House Formatter. My question now is between xsltproc, Saxon
> 6.5, and Saxon 10.

Sorry I misunderstood your question.

> All three are working for me, and I don't have a reason to prefer one
> to the others. Is there one?

Saxon 6.5 is written in Java. If you need or want extension functions
(like the DocBook extension functions that work out image sizes), they
are probably easier to write and use in Java then they are in C. But
xsltproc might be faster in some cases because C is sometimes faster
than Java. Saxon 10 is an XSLT 3.0 processor. It will be processing 1.0
stylesheets in backwards compatible mode. Thatâs probably ok, I canât
think of any areas where itâs likely to be a problem off the top of my
head, but it does mean you can accidentally write XSLT 3.0 instructions
into your customization layers and theyâll slip past the processor. That
might be really useful or really confusing depending on circumstances.
Saxon 10 is being actively developed. If you find a bug in Saxon 6.5,
itâs exceptionally unlikely (IMHO) that a bug fix would be released.
That said, Saxon 6.5 has been in use for going on a couple of decades,
so itâs not likely youâll find any new bugs.

Iâd probably use Saxon 6.5 or xsltproc for XSLT 1.0 stylesheets, but Iâm
not sure. Saxon 10 is probably a better choice if you have an eye
towards someday using maps or arrays or packages or any of the modern
features of the XPath/XQuery/XSLT ecosystem.

I found XSLT 3.0 compelling enough to switch.

                                        Be seeing you,
                                          norm

--
Norman Tovey-Walsh <ndw@nwalsh.com>
https://nwalsh.com/

> We are constantly invited to be who we are.--Henry David Thoreau

Attachment: signature.asc
Description: PGP signature