Re: Transformation needs [was Re: returning a nodelist]

[Brandon Ibach <bibach@xxxxxxxxxxxxxx>]

Norman Walsh said:
> 
> I'd like that, too.  I'd love to transform DocBook input that uses blocks
> inside paras into DocBook that doesn't:
> 
>   <para>xxx<orderedlist>...</orderedlist>yyy</para>
> 
> would become
> 
>   <para>xxx</para><orderedlist>...</orderedlist><para>yyy</para>
> 
> This would make the HTML that results from the DocBook to HTML
> transformation "legal" HTML in more cases.
> 
   This is just the sort of thing I'd like, too.  In my case, I have
data which is represented item by item in the SGML, but needs to be
built one column at a time for the HTML output.  It would be so much
easier to two-step it so that a transformation pass lines the data up
for output, and a styling pass actually generates the HTML.

> SMOP, I suppose, to fix up the input with a transformation using current
> Jade techniques or the HTML afterward.
> 
   That would be a way to go, but I think the actual transformation
language (if you can figure it out) could provide more compact,
powerful transformation capabilities.

> The other thing that I need, and I'm beginning to feel pretty
> desperate about it actually, is easier access to PIs.  James, do
> you have any comment on what the implementation effort would be
> to support PIs in Jade?  By support, I mean have a construction
> rule (or any function, honestly) fire automatically when one is
> encountered.
> 
   Couldn't you just put in a (default) rule, and check for the class
of the node to set if it's a PI?  Ok, never mind (just checked the
standard... default only processes nodes of class element).  Okay, how
about an alternate version of (default) which catches all classes of
nodes?  :)

-Brandon :)


 DSSSList info and archive:  http://www.mulberrytech.com/dsssl/dssslist