- From: Brian Sexton <discussion-w3c@ididnotoptin.com>
- Date: Fri, 17 Dec 2004 17:54:29 -0800
- To: <www-html-editor@w3.org>
- Message-ID: <004d01c4e4a4$83489cc0$651aa143@desktop>
Hello! I see from the XHTML 2.0 Working Draft dated 22 July 2004 that "target" attributes are to be included for "a" elements (see http://www.w3.org/TR/2004/WD-xhtml2-20040722/mod-hyperAttributes.html#adef_hyperAttributes_target). What I do not see (not because it is definitely not there; possibly because it is lost in a sea of other text) is whether they will be part of the strict document type definition. Please then forgive me if the issue about which I am writing has already been resolved. This issue has been discussed by many people at many venues, but my contention is simple: if a markup language such as XHTML provides a way to embed another document into the current document via an element such as XHTML's "object" element, that markup language should provide a way for those embedded documents to specify whether their targets open within their containing elements or at the parent document's level. I believe this should be done with a single DTD--NOT with a strict DTD and a loose DTD--to allow strict documents to be embedded within other strict documents without having their links trapped within their containing elements. It is true that using a loose DTD for the embedded document presently allows the use of the "target" attribute for breaking out of that element, but forcing developers to use loose DTDs is not much better than forcing them to use scripting (an available, but unreliable and possibly DTD-violating method of link-targeting). It has been argued that "target" attributes are presentational rather than structural in nature, but I disagree. A road is structural. A train track is structural. A ladder leading down into a pit is structural. To say that such a ladder is structural when it leads down into the pit but presentational when it leads out just doesn't make any sense. And even if by some clever permutation of logic one could show some sense in that assertion, it just doesn't make any sense to give Web developers ladders for end-users to climb down into "object" elements, but require a separate DTD, scripting, or a future version of CSS--the latter two of which the end-user may not have access to--to provide a way for the end-user to climb back out. Kind regards, Brian Sexton
Received on Saturday, 18 December 2004 01:54:30 UTC