OASIS Darwin Information Typing Architecture (DITA) TC

While I think the controlled values proposal will be very useful and may 
be appropriate for also defining the effective values for unconstrained 
enumerated attributes, I think we should also look at the OASIS 
Genericode standard as well, as it may be more appropriate for this 
particular use or may be something that can work with the controlled 
values proposal in a useful way. See http://www.genericode.org/

Tony Coates gave a nice 6.3 minute presentation on Genericode at XML 
2007 (one of the lightning round presentations) and it seemed like just 
the thing. However, it is still under development so that may make it 
unavailable for immediate consideration for 1.2

It seems like a very simple solution to the general problem of simply 
defining sets of "codes" (that is, distinct values within some domain) 
and relating individual code values to some governing definition (e.g., 
a set of topics used in the service of defining a set of controlled 
values).

Genericode has been used with SchemaTron to do provide second-pass 
validation for things like UBL, which uses genericode to define sets of 
codes that must vary by user, country, business domain, and so on.

 From what I've seen, it seems like it would be pretty straightforward 
for editing tools, for example, to integrate support for 
genericode-defined value sets into a DITA-specific authoring environment 
given a little bit of configuration (e.g., for this 
shell/specialization, use this code set definition for this 
type/attribute pair).

Cheers,

Eliot
-- 
Eliot Kimber
Senior Solutions Architect
"Bringing Strategy, Content, and Technology Together"
Main: 610.631.6770
www.reallysi.com
www.rsuitecms.com

Hi, Eliot:


Thanks for mentioning this work.  I'd seen the UBL code lists previously but wasn't aware that it had been spun off as the Genericode initiative.  I can easily see taking advantage of Genericode if good infrastructure comes into being around it.


I wouldn't think, however, that we'd want to drop the controlled values proposal from DITA 1.2 for several reasons:



Genericode is relational -- the XML equivalent of a database table.  It represents the properties for each thing in a flat list.
The DITA controlled values proposal is hierarchical and associative -- it has a primary purpose of representing the subjects covered by the content we create for use in filtering, flagging, and retrieval.  It makes sense to share and exchange the definitions of such subjects with the content. 
Controlled values and controlled vocabulary have a natural representation with topics and their links.  (Indeed, Wikipedia has been processed as one enormous web of controlled values.)  It would be hobbling the potential of DITA not to take advantage of that capability of topics and maps. 
We would benefit from having a single method for specifying enumerations whether selection values, pedagogy types in learning, and so on.


I don't think that's antagonistic to Genericode -- if we can validate an attribute enumeration by flattening a DITA scheme to a Genericode list and then using Genericode-aware Schema validators or a Genericode Schematron generator, we should leverage those tools for those environments -- but we have good reasons for having a DITA method for representing controlled values and vocabularies.



Erik Hennum

ehennum@us.ibm.com



Eliot Kimber <ekimber@reallysi.com> wrote on 12/12/2007 01:59:32 PM:


> While I think the controlled values proposal will be very useful and may

> be appropriate for also defining the effective values for unconstrained

> enumerated attributes, I think we should also look at the OASIS

> Genericode standard as well, as it may be more appropriate for this

> particular use or may be something that can work with the controlled

> values proposal in a useful way. See http://www.genericode.org/

>

>  From what I've seen, it seems like it would be pretty straightforward

> for editing tools, for example, to integrate support for

> genericode-defined value sets into a DITA-specific authoring environment

> given a little bit of configuration (e.g., for this

> shell/specialization, use this code set definition for this

> type/attribute pair).