Would the following help:
Use a <learningObjectMap> to create standalone aggregations of <learningContent> together with its supporting plan, overview, summary, and assessments, according to the needs identified by the learning goals
and objectives.
While it is possible to include multiple <learningObject> in a <learningObjectMap>, how processors deal with those additional instances of <learningObject> will be processor dependent.
Any <topicref> elements that occur before the <learningObject> are intended to be resource-only references such as key definitions.
The last sentence can be struck from the spec since it’s no longer possible to add topicref elements with the changes in the errata. Since we can’t add any new language that is normative, then what I’m trying
to say is that you can have multiple learningObject, but how they are processed is in the hand of the processors.
Éric Sirois
DITA Toolsmith ?
IXIASOFT
825 Querbes, Suite 200, Montréal, Québec, Canada, H2V 3X1
tel + 1 514 279-4942 / toll free + 1 877 279-4942
mobile + 1 647 462-3620
eric.sirois@ixiasoft.com /
www.ixiasoft.com From:
dita@lists.oasis-open.org [mailto:
dita@lists.oasis-open.org]
On Behalf Of Kristen James Eberlein
Sent: August 17, 2016 7:48 PM
To:
dita@lists.oasis-open.org Subject: Re: [dita] Issue with leaningObjectMap content model
We cannot make the change to the written spec that Eliot suggests; it would introduce normative wording.
The grammar files are normative, so what we should do is change the spec language to reflect the reality of the grammar files.
With DITA 1.2 (and perhaps earlier), we stated that if there was a discrepancy between the written specification and the grammar files, the written spec took precedence. However, as we prepared DITA 1.3, OASIS made it clear to us that their processes required
that the grammar files be normative.
Best,
Kris
Kristen James Eberlein
Chair, OASIS DITA Technical Committee
Principal consultant, Eberlein Consulting
www.eberleinconsulting.com +1 919 682-2290; kriseberlein (skype)
On 8/17/2016 6:35 PM, Éric Sirois wrote:
Hello,
France Baril mentioned in a post on dita-users that the content model for learningObjectMap does not match what is allowed in the DTD,RNG, and XSD.
<snip>
The documentation says: "The <learningObjectMap> references one and only one <learningObject> ".
However, the schema allows for an unlimited number of learning Objects in a learningObjectMap.
</snip>
With a reply from Eliot in an email thread:
<snip>
The RNG also allows multiple instances of <learningObject>--we never attempted to implement a constraint that would limit it to exactly one while also allowing the other topicref types that should be allowed
before or after <learningObject> (mapref, keydef, ditavalref, topicgroup).
That level of constraint would require completely redefining the content model for the learningObjectMap element, which we could do but which we chose not to do (or failed to consider the need for).
The only constraint we implemented was the learningAggregationsTopicrefConstraint, which disallows <topicref> but allows the other items from the mapgroup domain.
The simpler change might be to adjust the the reference entry to change "references" to "SHOULD reference".
</snip>
We should discuss this issue at the following TC meeting.
Éric Sirois
DITA Toolsmith ?
IXIASOFT
825 Querbes, Suite 200, Montréal, Québec, Canada, H2V 3X1
tel + 1 514 279-4942 / toll free + 1 877 279-4942
mobile + 1 647 462-3620
eric.sirois@ixiasoft.com /
www.ixiasoft.com