keyref and keydef: change to the DITA 1.2 spec., DTDs, and XSDs needed?

["Ogden, Jeff" <jogden@ptc.com>]

Included below is an e-mail exchange between Robert and myself.  The long and short of it is that we think that the "special" defaults for @toc, @print, and @linking on <keydef> in the DTDs and XSDs need to be removed and the DITA 1.2 specification needs to be updated to have a somewhat longer list of attributes that should not be merged as part of key reference resolution processing.  This note is to see if the rest of the DITA TC agrees.

   -Jeff


> Sent: 03/16/2010 11:13 PM
> From: "Ogden, Jeff" <jogden@ptc.com>
> To: Robert D Anderson/Rochester/IBM@IBMUS
> cc: "Grosso, Paul", "Helfinstine, Dave" <dhelfinstine@ptc.com>, "Nitchie, Chris" <cnitchie@ptc.com>
> Subject: keyref and keydefs?
>
> Robert,
>
> I know I retired, but I’m trying to finish up a few loose ends …
>
> In the following example, what is the effective value for @toc, @linking,
> and @print on the topicref after the key reference has been resolved?
>
> <map>
>   . . .
>   <keydef  keys=”abc”  href=”topic.dita” />
>   <topicref  keyref=”abc”  href=”othertopic.dita” />
>   . . .
> </map>
>
> I ask because @linking, @toc, and @print have explicit DTD defaults
> on keydef and when those values are merged using the conref merging
> rules, we end up with what I think are unexpected values for @linking (no),
> @toc (no), and @print (no) on the resolved topicref.
>
> Am I looking at this correctly?
>
> If I am, I suspect that the way out is to remove the explicit DTD
> defaults for @linking, @toc, and @print on keydef and let
> @processing-role=”resource-only” take care of things.
>
> What is the list of attributes that should not be merged from a key
> definition?  Is it?
>
>       @id
>       @conref, @conkeyref, and the other conref attributes
>       @processing-role
>       @xml:lang, @dir, @translate (special rules apply)
>       @keys, @keyref
>
> These probably don’t matter, but for good measure:
>
>       @class
>       @domains
>
> In the draft DITA 1.2 spec. only @id and @keys are mentioned as not
> merging. The special rules for @xml:lang, @dir, and @translate are
> mentioned.
>
>   -Jeff


>> From: Robert D Anderson [mailto:robander@us.ibm.com]
>> Sent: Wednesday, March 17, 2010 8:56 AM
>> To: Ogden, Jeff
>> Cc: Nitchie, Chris; Helfinstine, David; Grosso, Paul
>> Subject: Re: keyref and keydefs?
>>
>> Hi Jeff,
>>
>> I think that's right (both the list of how keydef is working - in a
>> confusing way - and the list of what should not be merged). The
>> toc/linking/print values were given the default of "no" before we had
>> the
>> processing-role attribute, and that was never undone.
>>
>> Robert D Anderson
>> IBM Authoring Tools Development
>> Chief Architect, DITA Open Toolkit


>>> Sent: 03/17/2010 04:54 PM
>>> From: "Ogden, Jeff" <jogden@ptc.com>
>>> To: Robert D Anderson/Rochester/IBM@IBMUS
>>> 
>>>
>>>     "Grosso, Paul" <pgrosso@ptc.com>, "Eric Sirois" <esirois@ca.ibm.com>
>>> Subject: RE: keyref and keydefs?
>>
>>> Thanks.  I'll make a local change to the DTDs and XSDs here and let Chris
>>> and Dave get the updated versions from you and Eric when they become
>>> available.  I guess an update to the DITA 1.2 spec. is needed too.
>>> 
>>> Do we need to send something out to the DITA TC list to make this
>>> official? If so, do you want me to do it or would you like to do it yourself?
>>> 
>>>    -Jeff


>>>> From: Robert D Anderson [mailto:robander@us.ibm.com]
>>>> Sent: Wednesday, March 17, 2010 5:30 PM
>>>> To: Ogden, Jeff
>>>> Cc: Nitchie, Chris; Helfinstine, David; Eric Sirois; Grosso, Paul
>>>> Subject: RE: keyref and keydefs?
>>>> 
>>>> Hi Jeff - yes, I think it needs to go through the TC. I'd be happy if
>>>> you want to send it along. :-)
>>>> 
>>>> Thanks -
>>>> 
>>>> Robert D Anderson
>>>> IBM Authoring Tools Development
>>>> Chief Architect, DITA Open Toolkit