- From: Lenny Domnitser <ldrhcp87@yahoo.com>
- Date: Wed, 24 Mar 2004 15:11:00 -0800 (PST)
- To: www-html@w3.org
In the latest strict versions of HTML, the `ol` element has no `start` attribute. The removal of this attribute breaks core functionality of ordered lists. The CSS alternative is a complicated counter-based solution that is poorly supported by browsers. However, even if browsers properly supported CSS counters, a structural part of the list is being deprecated. Imagine using a non-CSS browser and instead of I. First Section a. sub-section b. yet another an interrupting paragraph c. some more II. foo a. bar you saw 1. First Section 1. sub-section 2. yet another an interrupting paragraph 1. some more 1. foo 1. bar Now imagine being told to refer to point II.a. This is not an issue of styling: it is critical to the meaning of the list. Also note that, although slightly less important than `start`, `type` is also semantic. Suggesting that list indices should be generated by CSS is like creating the entire body of a page by CSS's `content` property: a poorly supported non-semantic solution that is pushed by the same zealots that removed `target` from `a`. At the very least, `start` should be returned, however, there are still many properties of ordered lists in HTML that make it very inflexible. For example, suppose David Letterman wanted to create a website for his famous "top ten" lists. For those of you unfamiliar with Letterman, "top ten" is an inane countdown that is presented with the tenth item first and the first item tenth. Already established is that properties of ordered lists are semantic and should be handled by HTML. Top ten is clearly ordered, however, because it does not conform to the standard "1, 2, 3" form of list, it cannot be properly marked up. Letterman, in fact, does have a "top ten" site [topten], on which the items are wrapped in paragraphs. What is a possible solution to inflexible lists? Perhaps another attribute such as `step`. So Letterman's web lackeys can write <ol start="10" step="-1"> <li>Once when he was 30, he smiled.</li> <li>Has experimented with drugs, alcohol and Botox.</li> ... </ol> This, however, is still too rigid. I propose a format more like that of definition lists, with an arbitrary index that is explicitly defined by the author, in, say, an `lindex` element, perhaps abbreviated to `lx`. So Letterman can now use whatever type of ordering he wants, instead of limiting himself to the predefined list styles [zvonlst]. An example would be <ol> ... <lx>8.</lx> <li>Had he decided not to go into politics, it would have been the Captian, Tennille and Kerry.</li> <lx>7.</lx> <li>Belonged to an exclusive secret society called the Columbia Record and Tape Club.</li> ... </ol> XHTML2 is already moving toward explicit definition, specifically with the `quote` element [xhtml2q], replacing the `q` element. Aside from the name change, the main difference is that while `q` depended on the browser to automatically surround its contents with quotation marks, `quote` elements include quotation marks. For example, <q>Nay, I say!</q> becomes <quote>"Nay, I say!"</quote> Although CSS is becoming more widely used, we must remember that it is not a replacement for proper markup, but merely a complement. One standard should not be dependent upon full support for another one, especially one that is broken in the most widely-used browser. --- [topten]: http://www.cbs.com/latenight/lateshow/top_ten/ [zvonlst]: http://zvon.org/xxl/CSS2Reference/Output/prop_list-style-type.html [xhtml2q]: http://www.w3.org/TR/xhtml2/mod-inline-text.html#edef_inline-text_quote __________________________________ Do you Yahoo!? Yahoo! Finance Tax Center - File online. File on time. http://taxes.yahoo.com/filing.html
Received on Thursday, 25 March 2004 02:08:31 UTC