Re: [dita] The collection-type attribute

[Michael Priestley <mpriestl@ca.ibm.com>]

Right now the toolkit doesn't provide
any difference in output between unordered and choice - the difference,
semantically, is that "unordered" implies that the children are
not in an ordered relationship, and have no specific relationship to each
other; whereas "choice" implies that each of the children is
an alternative or option out of a set. 

Here's a rough outline of the differences:

(default) or unordered: no semantic
implication. Child topicrefs are collected under the parent for a reason,
but the reason doesn't imply stucture or behavior (ie one or all of the
subtopics may get read).
Example:
>Things I own
   >A car
   >A house
   >A computer
Current result: parent topic links to
children in bulleted list, children link back to parent


choice: child topicrefs are collected
under the parent to indicate that they are alternative paths or implementations
of the parent type, ie only one of the subtopics is likely to get read
Example:
>Getting to the meeting
   >By car
   >By bike
   >By public transit
   >On foot
Current result: parent topic links to
children in bulleted list, children link back to parent

sequence: child topicrefs are collected
under the parent to indicate that they are typically read in sequence
Example:
>Getting there by bike
   >Packing for the trip
   >Leg 1: from A to B
   >Leg 2: from B to C
   >Where to lock up your
bike
Current result: parent topic links to
children in numbered list, children link back to parent, children link
to next and previous siblings


family: child topicrefs are collected
under the parent to indicate that they are typically all read together
- while the reader could start at any place, the topics are all closely
related and typically the reader will need to read all of them to answer
their question
Example:
>Walking versus biking
    >Walking
    >Biking
Current result: parent topic links to
children in bulleted list, children link back to parent, children link
to all siblings

I hope this helps,


Michael Priestley
mpriestl@ca.ibm.com

Christopher Wong <cwong@idiominc.com>

07/06/2005 03:04 PM

To	Don Day <dond@us.ibm.com>, dita@lists.oasis-open.org
cc
Subject	Re: [dita] The collection-type attribute

Thanks, but these descriptions aren't very complete.
What's the 
difference, for example, between collection-type="unordered"
and 
collection-type="choice"? Right now, we're trying to figure out
how to 
render these things by running samples through the Open Toolkit, but 
that is sort of putting an implementation before the standard. It seems
that we are still missing the original intent or meaning behind each of
these possible values (unordered, choice, sequence, family).

Chris

Don Day wrote:

>It's there, in a link to the %topicref-atts; group of attributes:
>http://docs.oasis-open.org/dita/v1.0/langspec/topicref-atts.html
>
>Granted, this description is terse.  Michael provided more detail
that you
>can apply to implementation behavior in the map attribute descriptions
>here:
>http://docs.oasis-open.org/dita/v1.0/archspec/dita_spec_23_common.html
.
>The DITA Open Toolkit should demonstrate these processing-based behaviors.
>
>  
>