Please forgive me for launching these thoughts into the middle of
the discussion without a thorough consideration of the whole thread.
I only just joined the docbook@ list, having been confused about
it vs. docbook-tc@ (where I'm an observer) for quite a long time
now. On discovering my error, Norm suggested that I send some
thoughts I had intended for the TC list but couldn't send there
because of my status.
I was originally attempting to respond to the TC list thread about a
"task" as a sibling to a "section", which relates to the question of
considering a DITA-like "topic" structure. Here's what I had hoped
to send:
====
IMHO, a task should only be allowed as a peer to a section if
there's a document-construction methodology that makes sense of
this, such that the model actually reflects a targeted application.
I wrote a bit about this topic here a while back:
http://www.xmlgrrl.com/blog/archives/2005/11/02/topic-oriented-architecture/
If you were doing an Information Mapping-like model, where they have
a clear notion of seven different categories of "block" that can be
peers and two of them are "task" and "generic info" (actually real
IM would have "procedure" and a bunch of other specialized ones,
like "concept"), then okay. Without that discipline, you'd better
be very sure about agreed processing expectations before allowing
it, because it's not in the common experience of most writers to do
this, or in the experience of most readers to see it.
It's not impossible to come up with such rules, but it seems to me
they would be "made up" rather than codifying existing norms.
====
I hope not to be such a stranger, now that I'm on the list with all
the cool action. :-)
Eve
--
Eve Maler +1 425 947 4522
Technology Director eve.maler @ sun.com
CTO Business Alliances group Sun Microsystems, Inc.
