Error terminology [was: Are indexterm ranges backwards incompatible?]

["Grosso, Paul" <pgrosso@ptc.com>]

	From: Esrig, Bruce (Bruce) [mailto:esrig@lucent.com] 
	Sent: Wednesday, 2006 August 16 05:45
	To: Paul Prescod; Dana Spradley; Michael Priestley
	Cc: Chris Wong; dita@lists.oasis-open.org; JoAnn Hackos; Grosso,
Paul
	Subject: RE: [dita] Are indexterm ranges backwards incompatible?
	
	The wiki page
	
http://wiki.oasis-open.org/dita/Indexing_issue_summary?action=show 
	is again more up to date.

________________________________

Thanks, Bruce.

I'm happy with leaving word-smithing to when we
have actual draft text to discuss, so I'm not
suggesting we open up our decisions at this point,
but I do need to note that saying something like
"<index-see> is allowed at the deepest levels only"
is not sufficient for our spec because we cannot
*prevent* such from happening.  Therefore, we need
to say what should occur when <index-see> does occur
at other than the deepest level.

We might consider doing what the XML specs do which
is to define "error" [1] to be "a violation of the 
rules of this specification; results are undefined....
Conforming software MAY detect and report an error 
and MAY recover from it."  That allows for giving
warning or error messages, but also allows simply
ignoring it.  (A good implementation would probably
give the user a choice as to whether to receive
warnings or not.)

Then instead of having to repeat ourselves throughout
the spec, we can just say something like "it is an
error for an <index-see> to occur at other than the 
deepest level".

paul


[1] http://www.w3.org/TR/REC-xml/#dt-error