RE: [docbook-tc] proposal: add relatedlink element to topic

["Rowland, Larry" <larry.rowland@hp.com>]

>"Rowland, Larry" <larry.rowland@hp.com> writes:
>> I guess it is a usability issue.  I find that indexterms are
>> much more robust than keywords in our content because they are
>> embedded in the content they represent rather than remotely
>> located from it.  Adjacency is a strong principal in interface
>> design and general human factors.
>
>I suppose. I don't think there's a compelling argument to forbid them
>From occurring inline, it's a question of model design.
>
>I'm still waiting for someone to provide some concrete examples of
>where/how a relatedlink element would be used. I don't have a good
>sense of the processing expectations. Are they sorted, collated, are
>duplicates coalesced?

My understanding is that duplicates are coalesced and the resulting
list is collated and sorted.

>I'm not entirely comfortable with a model that says these elements
>might or might not resolve and that that might or might not be a
>problem.

This is an unfortunate problem of any reference in content that is
intended for assembly into arbitrary collections.  Unless the person
defining the assembly knows all the references made among the bits
and pieces being assembled, there will most likely be references
that do not resolve unless references that cross topic boundaries
are forbidden.  That would, of course, make content much less useful
to the consumer of the content.

>How are relatedlinks in a document related to (no pun intended) the
>relatedlinks section in an assembly file?

My understanding is that the relatedlinks in a topic would
be merged with the relatedlinks specified in an assembly and
duplicates would be coalesced.  Once the list is generated, they
would be sorted and collated and presented at the end of the
topic.  I felt it would be desirable to have the ability to 
indicate where the related links belong, since a section with
multiple topics in it would probably not be best with the 
relatedlinks for each topic to follow the topic, but people more 
familiar with DITA felt that a rule based model would be adequate.

Regards,
Larry

=======================================================================
[ Larry Rowland             | If you want to build a ship, don't drum ]  
[ ESS/ATG                   | up the men to gather the wood, divide   ]
[ Hewlett-Packard           | the work, and give orders. Instead,     ]
[ 3404 East Harmony Road    | teach them to yearn for the vast and    ]
[ Fort Collins, CO  80528   | endless sea. - Antoine de Saint Exupery ]
[ Phone: 970/898-2280       +-----------------------------------------]
[ E-Mail:larry.rowland@.hp.com                                        ]
=======================================================================