- From: Alexander Christiaan Jacob <ego@acjs.net>
- Date: Tue, 24 Apr 2007 15:44:03 +0200
- CC: 3w-html <www-html@w3.org>
Yahia Chlyeh wrote: > > I would like to express my disagreement where the XHTML2 spec says: > The xml:lang attribute is required on [the html] element. > <http://www.w3.org/TR/xhtml2/mod-document.html#edef_document_html> > > The problem is that xml:lang's value cannot contain more than one > language-code, nor it can contain none (actually validating againt > current XHTML documents, sends errors when encountering empty > xml:lang= and lang= attributes) > > You might say it isn't a problem, because a document must have a > primary language. But let's consider this case: > My XHTML2 homepage has links to my various websites in various > languages, and this homepage itself do not contain prominent > language-specific text. It may contain only the titles of the websites > the links refer to. (It also may contain only images as links to those > websites.) > > So, what in that case should the value be used in 'xml:lang='? > I see three solutions: allow xml:lang= > 1. with no value, > 2. with a comma-separated list of value (or both 1. and 2.), > 3. or unrequire xml:lang at all in the root element of XHTML2 documents. > > PS: allowing xml:lang= to be empty is good when you want to cancel the > inheritance of the document's language, and don't know in which > language a part of your document is (it can happen, or more commonly > do not know the code for the language, as it can be very unfrequent). > > --Yahia Chlyeh > <http://yahia.ma/antiblog/> > > You could consider using one of the following language codes on the root element: * mis (miscellaneous languages); * mul (multiple languages); * und (undetermined); * zxx (no linguistic content). (Depending on the exact use case, of course.) HTH, ACJ
Received on Tuesday, 24 April 2007 13:44:16 UTC