Re: [xml-dev] (In)Validate My Assumptions on Linking.

[Ari Krupnik <ari@cogsci.ed.ac.uk>]

Ben Trafford <ben@prodigal.ca> writes:

> 	1) Stylesheet languages need some sort of way to display links
> from generic XML. This is so we can interact with them in user
> agents. By "stylesheet languages," I am specifically referring to
> XSL-FO and CSS.

I think FO is less of a target for this effort because you cannot
"apply" FO to arbitrary XML - you have to build an XSL-FO document
based on your source XML, so the marginal cost of having to put XLinks
or fo:basic-links into the transformed document is trivial.

> 	3) XLink is -conceptually- on the right side of the
> 80/20. Forget the syntax, and focus on the actual ideas -- do they
> cover what needs to be covered? Especially if it were possible to
> easily extend them in the future.

I don't know how common of a use case this is, but I've often found
myself writing

<a href="{@link}">
  <xsl:apply-templates/>
  <xsl:if test="not(node())">
    <xsl:value-of select="@link"/>
  </xsl:if>
</a>

<example>See <a href="http://example.com/~john";>John's blog</a> and
<a href="http://example.org/weird";><!-- I can't be bothered to think
of a good title for it now, so just show the URL to the reader-->
</a>.</example>

CSS can almost do it, but has no way to differentiate between elements
that have content and empty elements:

a[href]:before {
  content: attr(href);
}

Ari.

-- 
Elections only count as free and trials as fair if you can lose money
betting on the outcome.