A few. I do think, when proposing these things, it is important to remember
that XSL-FO is not intended to implement all possible typsetting operations,
that it still needs to remain "easily" implementable.

I guess one question I have is how different should XSL-FO 2.0 be from 1.1?
Should it just be some minor changes/fixes, or are they considering some
pretty substantial changes? Because if it is the latter, I've got some
suggestions with that regard:

* Eliminate entirely the notion of unbounded page lengths/viewports. That
is, browser-like viewing with scrollbars and such. This, among other things,
has the effect of making the viewport stuff unnecessary and substantially
clarifies the specification and any descriptions thereof. The primary
purpose of XSL-FO is for paged media; we already have tools for unpaged

* Seriously reconsider having block-level elements and inline elements be
interleaved. It probably makes FO processor implementation a bit more
difficult. Plus, it's just needless; you can easily wrap that inline stuff
in a block.

* Page specification (simple-page-master) could be made so much more
intuitive. It is far too easy to put a header in the middle of your content
by accident, and while I support that functionality, it should not be the

If it's just minor changes:

* Allow for lists that periodically reverse the left/right (IPD,
technically) positioning of the blocks. Generally, to allow for lists that
alternate one after another which side the content and which side the list
item is on. It would, also, be useful to alter other attributes when doing
this. There is a FO-processor that has an extension to do it, but I forget
its name. This alternation would likewise be reset at every page/region

* In that same vein, similar alternation patters for table row properties
that reset at page/region breaks. This would allow for alternate color
backgrounds, but that always start (wherever they are visible) on a
particular boundary.

* FO 1.1 added the ability to have multiple body regions on a page, and a
flow map to dictate how data flows from one to another. What was not added
were appropriate keeps/breaks to deal with the possibility of switching to a
new region without leaving the page itself. This distinction should be made.

* The ability to specify 2-up/4-up/etc style page generation. This cannot be
done even with FO 1.1's multiple body regions, because one lacks the ability
to have multiple static content regions on a page, as well as where those
regions get placed. Something like a "multiple-page-master" that physically
shrinks several simple-page-masters onto a single page.

* Footnote numbering/indexing per page or region, rather than leaving it up
to the FO document. The "numbering", of course, needs to be flexible and
user-definable (perhaps as a sequence of character possibilities that is

* Meta-data needs to be handled in some way. A section, perhaps just after
the page master section, would be ideal.

* Please give us a RELAX NG+Schematron schema for XSL-FO. It would be so
incredibly useful to have such a thing.
View this message in context: 
Sent from the FOP - Dev mailing list archive at Nabble.com.

Reply via email to