(K. Ari Krupnikov) wrote:
| "Thomas B. Passin" <> writes:
|> Arjun seems to want us to disambiguate everything
|> with some combination of control attributes plus context.
I'm not sure how I introduced the notion of context.
| Exactly what I said. Instead of (attribute name fixed, value affects
| interpretation)
|
| xmlns:xsl="http://www.w3.org/1999/XSL/Transform",
|
| Arjun proposes
|
| xslLiteral="true"
|
| (attribute value fixed, name affects interpretation).
I didn't propose this. My proposal is based on
(attribute name declared, value affects interpretation - including
non-assertion of the attribute altogether)
The difference between "fixed" and "declared" is that "fixed" applies
across all instances, "declared" for only this instance.
| The same scoping problems remain.
There are no scoping problems unless one chooses to implement an inference
scheme. But infering the value of a missing attribute specification (as
"scoping problems" typically become) is not the point. It's to treat
non-assertion as significant.
