:target pseudo-class vs. XPointer


   Fragment identifiers are subject to the relevant Media Type. For
HTML, [RFC2854] defines them to refer to anchors in the document (i.e.
elements with the corresponding id or name attribute value). For XML
documents (and thus for XHTML documents) the semantics of fragment
identifiers have yet to be resolved (see [RFC3023] and [XHTMLMIME])
but the intention is to delegate the semantics to XPointer ([XPointer])
(at least for XML, see my discussion on www-html and w3c-html-wg).

The CSS Level 3 Selectors Module [Selectors] introduces a new
pseudo-class ':target' to refer to whatever the fragment identifier
identifies (see http://www.w3.org/TR/css3-selectors/#target-pseudo).

The draft currently incorrectly states "URIs with fragment identifiers
link to a certain element within the document, known as the target
element". This is true for HTML documents, but not for most other Media
Types. Depending on the successor of [RFC3023] it won't be true for XML
documents. XPointers may be used to select many elements or just
fragments of text nodes (using the ranges feature).

I think the module has to address the following issues:

  * does the :target pseudo-class apply to XPointers (i.e. non-simple
    fragment identifiers)?

  * If yes, is this feature mandatory for CSS3 conformance (i.e.
    mentioned in the CSS3 profile, requiring support for the fragment
    identifier mechanism of the current Media Type)?

  * If yes, why shouldn't there be a :xpointer() pseudo-class function
    to extend W3C Selectors through XPointer (and by reference XPath)?

  * Are there any additional restrictions for this pseudo-class
    (currently there aren't)?






all the best,
Björn Höhrmann { mailto:bjoern@hoehrmann.de } http://www.bjoernsworld.de
am Badedeich 7 } Telefon: +49(0)4667/981028 { http://bjoern.hoehrmann.de
25899 Dagebüll { PGP Pub. KeyID: 0xA4357E78 } http://www.learn.to/quote/

Received on Thursday, 18 October 2001 15:18:50 UTC