I see no harm in the spec giving
indications to vendors as to what they should be developing.
This is an important function
that's been in my requirements doc for six years or more, with no vendor
supporting it. Process as follows: Writer marks an item in a topic with
<term>. Processing harvests every <term> in all the topics in a map
(or other aggregation). If a given <term>-tagged item also appears in an
associated glossary document (associated in the publishing environment), it is
included in a glossary for the publication that is assembled and published as
part of that document, and appropriately linked in online renditions. If a given
<term>-tagged item does not appear in the glossary, a flag is raised for
someone to create a glossary entry for it, and until then that item is not
rendered as a glossary item (link, etc.) or in any special
At present, glossaries are labor
intensive, error prone, and often omitted although desired.
This is an example of how use
cases may not be being communicated to vendors, and the spec can help close that
loop. Relevance to adoption is obvious.
I pretty much like working with xquery and the idea
that this problem should be left to impementers or actually be solved by the
stylesheet. But I am not sure if that is reliable with all languages and
Regardless if that task can be solved by implementers
in all cases, we should not generate expectations to future DITA
implementations in the langref over a period of 3 DITA
Either we should provide
associative linking to matching glossary entries as
it is mentioned in the langref,
or we mention that the implementer can link the
<term> with the corresponding <glossterm> via stylesheet i.e.
I do not propose "deleting" but "giving
unambiguous descriptions who do not create expectations we may not going
of the DITA Technical Committee
of the DITA Machine Industry
I recall that there was a semantic linking feature on the table for
DITA 1.1, which may have been associated with the <term> element. I
believe it died more to lack of inertia than anything else.
concur with Paul Grosso's assessment that this would be application behavior,
best left to implementers.
Alan Houser, President
Group Wellesley, Inc.
Christian Kravogel wrote:
I have just
checked the langref of the element <term> in version 1.0, 1.1 and 1.2.
All langref descriptions are identically with the following
element identifies words that represent extended definitions or
explanations. In future development of DITA, for example, terms might
provide associative linking to matching glossary
development of DITA" is mentioned within 3 DITA releases we may have either
to change this text or to provide a solution. Or do we already have solved
this, then we may have to update the text.
Member of the DITA Technical
Chairman of the DITA Machine