Element domain tokens enable generalization during conref

Another DITA 1.0 foundational idea for @domains tokens: it's supposed to enable some complex processing when resolving content references between two document types that use different levels of specialization. Confused? Yeah, that's because this doesn't happen much.

Here's a specific example of this scenario: I have a document type that does not use the highlighting domain. I reuse a paragraph from your topic; confusingly, your topic includes the highlighting domain. If the reused paragraph includes <b>, what happens when that content is pulled into my topic?

Answer: the specification says I examine @domains, discover that my document doesn't allow <b>, and in the new context I generalize the element to <ph>. I've implemented this, and it works. But …

I've never actually seen this used. In the real world, my document types all use common domains, so all this extra processing is just overhead. Even when I've had help exchange content between organizations with different domains, the occasional use of a specialized element in a new context was 1) unlikely, and 2) not worth worrying about.


If this ever does happen in the real world, I'm no longer sure this is even the right way to handle it. That is - I'd be surprised if I grabbed your phrase and the bold (or monospace or italic) formatting from your document disappeared in my document.

Principle of least surprise says … just leave it alone. If my publishing system supports the specialized element, the special formatting is retained, and it's probably fine. If my publishing system doesn't support the specialized element, I get the general phrase formatting, and I'm none the wiser.