- From: Jukka K. Korpela <jukka.k.korpela@kolumbus.fi>
- Date: Fri, 16 Aug 2013 15:39:36 +0300
- To: "public-html@w3.org >> HTMLWG WG" <public-html@w3.org>
2013-08-16 14:56, Karl Dubost wrote: > Jukka K. Korpela [2013-08-16T07:35]: >> Since <cite> is in practice just one of the ways to italicize text (along with <i>, <em>, and <var>), there's no reason to assume that authors haven't used it that way inside <blockquote> elements, too. So assigning a semantic role to it when appearing in <blockquote> would be arbitrary and lead to wrong conclusions about existing documents. > Assuming the world as an hypothesis for valid usage is a possible option, but it basically leads to invalidates almost all HTML elements, because all of them are used for something which was not the intended purpose. Not at all. I am not proposing that <cite> element be declared invalid. (I might support an idea of deprecating it in favor or <i>, but that's a different issue.) What I'm saying is that specifications should not assign meanings to existing, old elements that have been widely deployed in manners that are not consistent with such meanings. The <cite> element has always been vague; when I first learned HTML, I took <cite> as meaning inline quotation, and later I have observed that half of the world has misunderstood it the same way. The element never had a rigorous, or even reasonably objective definition. It's too late to restrict <cite> to any particular meaning, still less assign new meanings (like relating to enclosing <blockquote>). > I prefer the option of "here, this is how it is supposed to be used" and let communities, even if small, make good usage of it (communities of agreements). Some of these communities will do a good usage. > > If anything could be done for softwares is to give more hooks. Maybe. There's a lot that could be done in such areas. Usually proposals to add such markup meet resistance due to lack of proven usage. It can be rather frustrating, but redefining existing elements is really a worse option. Communities can use classes, or custom tags (which work fairly well in practice, no matter what we think of them in practice), or conventions (e.g., the last element inside a <blockquote>, or the first element after a <blockquote>, specifies the source of the quotation). If you think that there should be special, standardized markup for specifying the source and author of quoted material, it is best to propose a new element, without any implied default rendering except perhaps display: block. Then software that is interested in it can easily get it, with no fear of getting wrong information like you would have if you assumed anything about <cite>. -- Yucca, http://www.cs.tut.fi/~jkorpela/
Received on Friday, 16 August 2013 12:40:00 UTC