- From: Leif Halvard Silli <xn--mlform-iua@xn--mlform-iua.no>
- Date: Thu, 27 Jan 2011 14:00:25 +0100
- To: Jirka Kosek <jirka@kosek.cz>
- Cc: Leif Halvard Silli <xn--mlform-iua@xn--mlform-iua.no>, Steve Faulkner <faulkner.steve@gmail.com>, Anne van Kesteren <annevk@opera.com>, HTML WG <public-html@w3.org>
Jirka Kosek, Thu, 27 Jan 2011 11:00:16 +0100: > Leif Halvard Silli wrote: > >> Does DocBook say that <subtitle/> does not belong in the Table of >> Content outline? > > This is presentational behaviour which is not defined i the DocBook > standard. However usually subtitles are not shown in ToC. So as header, but not included in the ToC, usually. >>> it is usually transformed to HTML as: >>> >>> <div class="section"> >>> <div class="titlepage"> >>> <h2 class="title">Foo</h2> >>> <h3 class="subtitle">Bar</h3> >>> </div> >>> ... content of section ... >>> </div> >> >> If there is something wrong with that transformation, then why isn't >> that subtitle header transformed to, for example, <p class="subtitle"> >> rather than <h3 class="subtitle">? > > I'm not sure what's your point here, but to me it seems completely > natural to turn <subtitle> into <hx> as it is still title/heading > although of secondary importance. My point was that while <subtitle> usually isn't kept shown in a ToC, any <hn> usually *is* shown in a ToC. If that's more important than whether the text fragment is some kind of heading, then converting <subtitle> to <hn> is unwise. <hgroup> currently turns all its children into a single heading - according to Ian. Such a thing could be described as information loss. To say that the <hgroup> contains two _headings_ but only one of them shows in the ToC is a different from HTML5's current solution which is that two or more children of a <hgroup> is treated as a multi-paragraph single heading with special rules for which part goes into the ToC. If we want to resemble DocBook, then the most important point seems to me to be that the subtitle element has some kind of "headingness". >> I ask because Anne mentioned W3C TR's, like XML 1.0, which currently it >> creates this outline: > > Well the situation of this document is at least funny as it is being > served as text/html, but containts code like > > <h1><a/>xxx</h1> > <h2>yyy</h2> > > which is interpreted as > > <h1><a>xxx</a><h2>yyy</h2></h1> > > Wonderful isn't it? ;-) Ha ha. Now I understand why there are so many strange :hover/:focus effects on that page! > Time to put polyglot serializer into pipeline probably. Indeed. >> Extensible Markup Language (XML) 1.0 >> W3C Recommendation ... >> Abstract >> Status >> TOC >> 1. Appendices >> 1.1 Intro >> 1.2 Terminology >> 2. Documents >> >> I cannot say that it would have been a clear benefit if the "W3C >> Recommendation ..." was concatenated with the top level heading/hidden >> from outline. > > Personally I think that it doesn't make sense to have "W3C > Recommendation", May be that depends on how "informative" one wants the outline to be. Just now I discovered that search results pages on www.google.com creates this outline: Google Search results <!--this one is visually hidden< --> 1st result 1 2nd result 2 3rd result 3 Joining the two first into a single one, would have taken away information, IMO. > but as the document doesn't use structural elements > like article/section I think it is correct to show "W3C Recommendation" > as child of "Extensible Markup Language (XML) 1.0". -- leif halvard silli
Received on Thursday, 27 January 2011 13:01:00 UTC