Re: Paper on use of selectors for adaptive hypermedia

> 2) The special interpretation of the "bc" namespace is dangerous. It
> appears to change the meaning of the following space (or "//" in
> XPath) from "descendant" to "any". But what if the "bc" namespace
> actually occurs in the document to be formatted? Since the CSS
> matching code has to be changed anyway, why not add some syntax as
> well?

For XPath, I came across this possibility:

//info[//bc:webpage[uri="..."]]

This works well with the example given in the paper and does neither
require a change of the XPath language, nor of its processing.

Namespaces: as not the "bc" prefix depicts the namespace, but the URI
associated with it, it should be no problem to use a special namespace
for distinguishing between browsing context and "nude document". It is
the same case as when, for example, I try to look at an XLink linkbase
using an XLink-aware browser: The browser will interpret the "xlink"
namespace as XLink hyperlink and not as "normal" markup as elements from
other namespaces.

We thought that using namespaces is better than introducing new tokens,
for the sake of compatibility with existing, non-browsing context
browsers.

> Ad 2: The Media Queries draft uses an extension of the @media and
> @import rules rather than an extension of the selector syntax.

We thought that it wouldn't fit into media queries, as our approach is
considerably more complex, but it might be worth more discussion.

Michael

Received on Wednesday, 30 January 2002 10:29:14 UTC