When XHTML2 arrives: Will people replace:
Sat 14 May<br> 2005
… with:
<l>Sat 14 May</l> <l>2005</l>
… or actually use the most semantically correct way I’m currently aware of for this kind of situations:
Sat 14 May 2005
… of course, making that line break appear on the screen is done through CSS. The semantics however aren’t affected as this particular construct happens to be inside an element that treats its content as non white space significant.
Phrased differently: Will people just look at solutions or at correct solutions? I think the former.
If the only cool part of XHTML2 is that you can embed stuff on every element and make every element a link — many people feel about XHTML2 this way — why not implement XLink support in browsers for XHTML documents; write some RelaxNG or XML Schema grammar around it and start using it?
Unless XHTML2 is easier to implement of course. But that is not the case. And besides, XLink and XHTML are here today; XHTML2 is not.
(Perhaps I’m looking too much at this from a practical point of view. Theoretically XHTML2 is a lot of fun.)
How many people see XHTML2 as a replacement for XHTML1? Is it? If I recall discussions correctly the W3C once stated it isn’t, but developers will likely need to use it anyway as people seem to require the ‘latest and greatest’. (One reason I really like the upcoming HTML5; which will solve that problem for HTML.)
How does XHTML2 interact with the DOM? An example: What is DOM events processing model for nested links?
How many semantic elements will be duplicated by the arrival of XHTML2? Is that a bad thing? What’s the difference between xhtml:em and xhtml2:em where xhtml and xhtml2 are bound to http://www.w3.org/1999/xhtml and yet unknown respectively.
I have heard rumors that XForms 1.1 will be embedded in XHTML2 sharing its namespace. While XForms 1.1 alone suffers from a lot of semantic duplication by creating a new namespace for just a minor update letting it inherit the namespace of its host language goes against all ideas of semantic purity to me. This rumor seems to be confirmed by the Namespaces for XForms 1.1 section in the XForms 1.1 specification (nowhere near recommendation so object now) which defines a so-called Chameleon schema to be used when XForms 1.1 inherits its namespace.
I don't think most people will write semantically in the near future, probably never (as least for those popular "non-geek" websites). Not even XHTML 1.1. :-'(
Correct me if I'm wrong, but I don't think, that line is a semantic information. The information part is day and year. To display these in different lines is presentation.
For me XHTML2 is just a very good xml format for webpages. To use it as the rendering format is probably not a good idea. CSS is just not good enough to transform semantic content into the preferred rendering.
I really love how I can use XSL to transform XHTML2 (+XInclude+XMLEvents+XForms) to HTML4+CSS+JavaScript. I might even do a transformation to SVG+CSS+Ajax some day. The important part is that I have my content in a format that describes the semantics just right (for now at least).
The XHTML 2 Namespace is known. It's defined in the conformance definition of the current draft.
It is also not to be used as the specification hasn’t reached a certain maturity and therefore I didn’t use it in my post. The final namespace is yet unknown and will most likely look like http://www.w3.org/2006/xhtml2. Unless they sneak it in earlier, that is.
In my mind, people tend to gravitate toward something that makes sense, which is a rather subjective item.
And until there is an understanding that XHTML is a data exchange format for delivering information to the UA, and not the markup language for Web pages, it will make sense for developers to write semantically invalid code.
Will people just look at solutions or at correct solutions? I think the former.
Like they always (seem to?) do.
How does XHTML2 interact with the DOM? An example: What is DOM events processing model for nested links?
Just like the way it happens right now I think. I don't see some special information on the new way to handle it, so I assume it will just stay with the system of event bubbling like it is now. Or do I misinterpret your question about this phenomenon?
I'm more concerned about software that generates code. Seems to me, that most who write their code manually, are hip to what they should be doing. So the real issue is, how do we encourage frontpage, nvu, and others, generate something proper? This ensures proper coding techniques by anyone who doesn't have time to learn how to do it right or has no interest in learning to code.
Devon > the best way to ensure NVu generates proper code is to contribute to it or help Daniel Glazman by bugfeeding him, or whatever you want to do.
Frontpage, on the other hand...
Your date example gets superfluous when people hopefully start to use ISO 8601. :-)
The question on XHTML 2 which I have on my mind for the last months is: Do we need a new MIME media type for XHTML 2? Perhaps with versioning information (which was brought into discussion again for the actual JavaScript/ECMAScript media type proposal). I am sure that there will be lots of browsers that will choke on XHTML 2 code fed to them while requesting application/xhtml+xml, but only knowing XHTML 1.x.
Lars Kasper, they won’t. They will treat it as any namespace they don’t know. And since browsers should support every */*+xml MIME type in theory and Opera does, that isn’t an effective solution. For XML browsers rely on namespaces anyway. Think of SVG or XHTML send as application/xml in Mozilla.
What we really need is some namespace negotiation header in HTTP.