OASIS Darwin Information Typing Architecture (DITA) TC






Dear TC 
Members,
 
Pardon my ignorance 
on this particular subject, but I'm looking for clarification on the filtering 
logic for conditional metadata.  To paraphrase the Architecture Spec (and 
please correct me if I'm wrong), the logic is to create a conditional processing 
profile (DITAVAL file) that explicitly identifies the values to be excluded 
(otherwise, the content is included).  I'm not concerned with flagging 
right now.
 

The way I 
understand it now, unless you explicitly declare an exclusion, it will be 
included. I have several clients that want the reverse.  In 
other words, the processing model would:
 
* Include 
all elements where:
    a) there is a matching value for the identified attribute, 
OR
    
b) the identified attribute is not specified on the element
* Exclude elements 
that
    
a) the identified attribute is set, but there isn't a matching 
value
 
Some of 
my clients make heavy use of conditional processing, where it makes more 
sense to specify that the current collection of topics should be processed for 
audience A, platform B and product C rather than explicitly excluding audiences 
D, E, and F, and so forth.
 
What I'm wondering 
is why the specification only identifies the explicit exclusion model.  
Wouldn't it be reasonable to suggest that a DITA processor could implement an 
explicit inclusion model also?  
 
Any insight you can 
provide would be very helpful.  
 
Best 
Regards,
 
Jim 
Earley

Hi Jim,

In DITA 1.1 you can set the default
behavior to exclude or include, and also override for particular attributes
as well as particular values.

For example, you could set default to
exclude for everything, but set the default to include for just platform
attribute values.

Michael Priestley

Lead IBM DITA Architect

mpriestl@ca.ibm.com

http://dita.xml.org/blog/25





"Earley, Jim"
<Jim.Earley@flatironssolutions.com> 
05/06/2008 02:15 PM




To
"dita@lists.oasis-open.org"
<dita@lists.oasis-open.org>


cc



Subject
[dita] Rationale for Filtering Logic
in Architecture Spec






Dear TC Members,
 
Pardon my ignorance on this particular subject,
but I'm looking for clarification on the filtering logic for conditional
metadata.  To paraphrase the Architecture Spec (and please correct
me if I'm wrong), the logic is to create a conditional processing profile
(DITAVAL file) that explicitly identifies the values to be excluded (otherwise,
the content is included).  I'm not concerned with flagging right now.
 
The way I understand it now, unless you
explicitly declare an exclusion, it will be included. I have several clients
that want the reverse.  In other words, the processing model would:
 
* Include all elements where:
    a) there
is a matching value for the identified attribute, OR
    b) the identified attribute
is not specified on the element
* Exclude elements that
    a) the identified attribute
is set, but there isn't a matching value
 
Some of my clients make heavy use of conditional
processing, where it makes more sense to specify that the current collection
of topics should be processed for audience A, platform B and product C
rather than explicitly excluding audiences D, E, and F, and so forth.
 
What I'm wondering is why the specification
only identifies the explicit exclusion model.  Wouldn't it be reasonable
to suggest that a DITA processor could implement an explicit inclusion
model also?  
 
Any insight you can provide would be very
helpful.  
 
Best Regards,
 
Jim Earley