Possible Additional Requirements for V2 - from EPUB

I've just finished working on a neat little EPUB pipeline and I've found
some bits that we should consider for our V2 work.  I'm not really sure
what we've put specifically on our plate.

1. For processes that build in stages, such as building bits of the EPUB
and then packaging it, dependencies are really important.  I was easily
able to use Calabash's cx:depends attribute [1] to implement this.  We
should consider that approach in trying to solve this kind of problem.  It
seem to work really well as you just wrap certain sections of your pipeline
in a p:group.

2. I think we have an issue with XSLT 2.0.  According to the XSLT 2.0
specification, it is possible for the primary result document (initial
result tree?) to be an empty sequence if the stylesheet only uses
xsl:result-document (see [2] and I'm not sure I'm using the right
terminology).  The problem is that it should be an XProc error if the
output on the result port is empty.  In this case, all the output ends up
on the secondary port and the result port is an empty sequence.  We should
probably make that only apply to XSLT 1.0.

3. Calabash's cx:message is very useful but its invocation is messy.

4. I would love attribute value templates for option values.

5. The zip step really needs to allow specification of directories and
recursive packing of their contents.  It is impractical to think that
someone will specify all the files in the manifest.  I had to resort to
p:exec to package the EPUB.

I've attached my XProc for converting my XHTML slides (slidy flavor) into
an EPUB.  You can see lots of room for improvement but the XProc is fairly
clean.  Attribute value templates and cleaner messaging (cx:message) would
make this XProc so much easier to read.

[1] http://xmlcalabash.com/docs/reference/extensions.html#extattr
[2] http://www.w3.org/TR/xslt20/#executing-a-transformation

-- 
--Alex Milowski
"The excellence of grammar as a guide is proportional to the paucity of the
inflexions, i.e. to the degree of analysis effected by the language
considered."

Bertrand Russell in a footnote of Principles of Mathematics