> Are there any vendors who implement @conref who *don't* do what the
Toolkit
> does? If so, then we have an interoperation problem. If not, then we can
> probably safely codify the Toolkit behavior by adding a specific
statement
> about @href.
The implementations I worked on early in DITA's life blindly copied the
href as-is, and when people started getting more adventurous with conref
across directory boundaries, it came up very quickly as a bug. To me, it
seemed obvious what the behavior *should* be - href is adjusted to remain
valid - but I can see Eliot's point that a strict reading of the spec would
copy href as-is with other attributes.
I'm happy clarifying the spec as Eliot wants to do, because I do think this
is the expected and reasonable behavior. In the years we've been adjusting
href during conref resolution, I don't think I've ever heard from any user
who was surprised by it. But, I'm willing to re-evaluate that opinion if
there are vendors or users today who have a reason to support the blind
copy.
Robert D Anderson
IBM Authoring Tools Development
Chief Architect, DITA Open Toolkit
