An idea of a "semantic" stylesheet
in [CSS]
|
From: Jukka K. Korpela on 8 Aug 2010 12:42 Out of curiosity more than anything else, I started playing with the idea of a stylesheet that overrides any other style sheets and presentational HTML features, as far as that is possible, and imposes a "semantic" way of rendering. By this I mean that the defined semantics of HTML markup is used as the basis, rendering elements in a manner that more or less reflects that semantics. For elements like <p>, old typographic tradition of first-line indent would be used. Links might be preceded by an arrow, as cross references in encyclopedias. Forms and tables would be rendered in a uniform manner, clearly distinguished as forms and tables. For some elements, interesting renderings might be used, like preceding any <kbd> element by a keyboard icon and following it with that same icon as grayed out. Fonts would be used in a disciplined way: one copy text font, another font for headings and tables, and a monospace font for <pre> and <tt> for which that's part of the definition. Demo: http://www.cs.tut.fi/~jkorpela/styles/semantic-test.html I'm not sure what that might be useful for. But it was an interesting idea, and using such a style sheet as a user style sheet (e.g., using Web Developer Extensions in Firefox, letting you add a style sheet on the fly) would reveal the "true" (= marked-up) structure, if any, of a page and might work as a sanity saver when you wish to read a page but get rid of its fancy formatting. -- Yucca, http://www.cs.tut.fi/~jkorpela/
From: William Gill on 8 Aug 2010 15:21 On 8/8/2010 12:42 PM, Jukka K. Korpela wrote: > Out of curiosity more than anything else, I started playing with the > idea of a stylesheet that overrides any other style sheets and > presentational HTML features, as far as that is possible, and imposes a > "semantic" way of rendering. Interesting, and useful. > ... For elements like <p>, old typographic tradition of > first-line indent would be used. ... I used to try to replicate this, and the convention of double spacing between sentences, but have abandoned the practice under the assumption that a new media created a new convention. > I'm not sure what that might be useful for. But it was an interesting > idea, and using such a style sheet as a user style sheet (e.g., using > Web Developer Extensions in Firefox, letting you add a style sheet on > the fly) would reveal the "true" (= marked-up) structure, if any, of a > page and might work as a sanity saver when you wish to read a page but > get rid of its fancy formatting. At the very least it serves as a good basis on which to build one's own consistent "typographic" style, if not a good foundation for an interesting philosophical discussion. Bill
From: Swifty on 9 Aug 2010 02:45 On Sun, 8 Aug 2010 19:42:00 +0300, "Jukka K. Korpela" <jkorpela(a)cs.tut.fi> wrote: >I'm not sure what that might be useful for. Well, it has taught me a number of HTML tags that I hadn't come across before. I already knew basic HTML tags when HTML first appeared as they are the same as the SGML markup languages that I'd been using. After that, I learned new tags "on demand". I can use Opera's "User mode" to see the effect of the tags if there were no CSS in effect. Your page is like the laser. Everyone could see that it is useful, but no one could think of a use, at first. :-) -- Steve Swift http://www.swiftys.org.uk/swifty.html http://www.ringers.org.uk
From: dorayme on 9 Aug 2010 04:48 In article <C7B7o.8557$136.648(a)uutiset.elisa.fi>, "Jukka K. Korpela" <jkorpela(a)cs.tut.fi> wrote: > Out of curiosity more than anything else, I started playing with the idea of > a stylesheet that overrides any other style sheets and presentational HTML > features, as far as that is possible, and imposes a "semantic" way of > rendering. > > By this I mean that the defined semantics of HTML markup is used as the > basis, rendering elements in a manner that more or less reflects that > semantics. For elements like <p>, old typographic tradition of first-line > indent would be used. Links might be preceded by an arrow, as cross > references in encyclopedias. Forms and tables would be rendered in a uniform > manner, clearly distinguished as forms and tables. For some elements, > interesting renderings might be used, like preceding any <kbd> element by a > keyboard icon and following it with that same icon as grayed out. Fonts > would be used in a disciplined way: one copy text font, another font for > headings and tables, and a monospace font for <pre> and <tt> for which > that's part of the definition. > Demo: > http://www.cs.tut.fi/~jkorpela/styles/semantic-test.html > > I'm not sure what that might be useful for. But it was an interesting idea, > and using such a style sheet as a user style sheet (e.g., using Web > Developer Extensions in Firefox, letting you add a style sheet on the fly) > would reveal the "true" (= marked-up) structure, if any, of a page and might > work as a sanity saver when you wish to read a page but get rid of its fancy > formatting. I take it that your idea is to do better than to simply disable all author styles and rely on the basic no frills styles provided by most browsers. By no frills I mean that if no author styles were included in or for an HTML doc, the browser would supply a basic style and this basic style reflects to some degree the semantics in the sense that if the author is not too clueless and more or less uses the right elements to house his meanings, the basic styles already bring the meanings out. I am thinking that your aim is to give particularly distinct styles to elements, ones that are more prominent or more useful than the no frills one that browsers provide. The no frills ones could also come under the description of showing the semantics up to a point. The paragraphs, for example, in the no frills ones more or less get to look like paragraphs, links are underlined and blue and are instantly recognised as links, lists go down the page and have bullets and so are also easily recognised. I confess not examining your treatment of other elements yet, this is where you might well be on to an improvement over what already exists. Since there is already a sanity saver to get rid of fancy author styles, your proposal is that yours does this too but in addition emphasises the meaning better. I imagine that some people might find this useful and especially if they take their cue to fashion their version of your sheet to their own liking. -- dorayme
From: Jeff Thies on 9 Aug 2010 08:39
On 8/9/2010 2:45 AM, Swifty wrote: > On Sun, 8 Aug 2010 19:42:00 +0300, "Jukka K. Korpela" > <jkorpela(a)cs.tut.fi> wrote: > >> I'm not sure what that might be useful for. > > Well, it has taught me a number of HTML tags that I hadn't come across > before. I already knew basic HTML tags when HTML first appeared as > they are the same as the SGML markup languages that I'd been using. > After that, I learned new tags "on demand". > > I can use Opera's "User mode" to see the effect of the tags if there > were no CSS in effect. > > Your page is like the laser. Everyone could see that it is useful, but > no one could think of a use, at first. :-) > The problem, as I see it, is not in the usefulness but in creating the markup. For someone with expertise in html and it could be very useful in switching in a stylesheet like this to see how the document was structured. Or at least it could be with some style changes. I can see uses for that in abstracts and studies where precision in meaning is important. It could also help in indexing the content. For the run of the mill web content, much less so. Just my 2 Markk Wuggas. Jeff |