Re: DOCBOOK-APPS: XInclude support in Saxon and Xalan

[ed nixon <ed.nixon@lynnparkplace.org>]

See below...

Elliotte Rusty Harold wrote:
> At 8:58 AM -0400 10/10/02, Jeff Beal wrote:
> 
>> I didn't know about this.  Does it try to do any sort of validation?
>>
> 
> No, validation is separate from XInclude. You can validate if you want 
> to or not. You can validate before or after inclusion or both or 
> neither. It helps to think of your system as a chain of black boxes 
> applied in sequence where the individual boxes for validation, 
> inclusion, etc. can be swapped or reordered or duplicated rather than as 
> a single black box which does everything at once.

We've seen the discussion about validation, XInclude, xmllint, etc. last 
week and before. If I remember it properly, the questions arose as to 
whether or not and when to validate content components included into a 
"larger" DocBook structure. I remember the discussion as being 
inconclusive about best practices or even approaches. There were issues 
around DTD design and generalized inclusion.

I guess it's obvious that a particular processing chain or application 
will be designed such that validation issues are dealt with properly. 
However, in a more generalized situation, e.g. where included content is 
being exchanged, shared, etc., it's difficult if not impossible to make 
assumptions about validity.

Are we at the point with DocBook that it's possible to "black box" 
validation as Rusty suggests? Do it before or after or neither and, 
ultimately, swap and reorder the black boxes at will?

            ...edN