- From: Leonardo Rojas <lrojas@gmail.com>
- Date: Fri, 15 Jul 2005 04:00:44 +0000
- To: www-style@w3.org
Hello, I'm new to this list and found this thread very interesting and would like to share some thoughts that some might find useful. Please feel free to correct me if I misunderstood or misrepresented anything. -- When laying-out a document the author is essentially restructuring it based on some criteria such as aesthetics, a specific communication intent (e.g.: marketing) or some special needs of the intended audience. While CSS is a great mechanism for styling, as of now, IMHO, it is still constrained to the structure of the document. This often leads to the author weakening the semantics in order to have more flexibility in the layout. Another negative side effect is the loss of generality, or the coupling of the two. During this thread some examples where given to show that the current CSS mechanics allow for the creation of any layout that the new language is capable of. While this is true, the samples were artificial, in the sense that the structure of the documents were created to suit the layout (this is desired but not always possible with the current CSS constructs). In other words it works fine when the blocks are the first siblings of the first visible node of the document. But what happens when the blocks are nested in different hierarchies? What the "new language" offers is a way to extract an arbitrary set of nodes from the document hierarchy and rearrange it into a different hierarchical structure according to the needs of the presentation. (while client side XSLT could do this, I understand that maintaining the simplicity of CSS is desired). The styling rules of the "new language" are not assigned to a specific node, they create a new separate structure, the layout, that will contain arbitrary sets of nodes from the semantically structure document. This is usually the intention when authors create a set of empty boxes and places them on the page. This also analogous to how the most common layout software for print works. It's also worth noting that the usual reason to reject proposals for CSS rules that flow overflown content into other nodes is that empty nodes would be required, which would not make a good semantic structure. My understanding to this point is that CSS as a language is fit to fulfill the styling needs that authors expect, a new language or syntax is not really needed. What is needed is the ability to establish an independent layout structure that is not tied to the semantic structure in a way that requires a compromise between the two. Thak you very much.
Received on Thursday, 28 July 2005 15:20:54 UTC