- From: Jason White <jason@jasonjgw.net>
- Date: Tue, 31 Jul 2007 18:17:20 +1000
- To: public-html@w3.org
On Tue, Jul 31, 2007 at 09:33:49AM +0200, Sander Tekelenburg wrote: > Yeah. I ignored that on purpose ;) because "new" is not the only situation to > which this applies. Consider the situation where one person edits existing > content that was authored by another person. The edit may well consist of a > language change, perhaps a partial one. So with every edit, not just on > "new", the authoring tool should allow the user to explicitly define @lang. A > new document is just a special case, where the tool must require either user > input, or insert no lang attribute. Correct. even if the authoring tool can infer the language from the grammar of the text, it should still allow @lang to be configured, in case the text classification algorithm yields an incorrect result. Based on those parts of the draft which I have read, there are few authoring tool requirements specified, and comparatively many user agent requirements. This makes good sense, in that the most important requirement of an authoring tool is that it produce conformant documents. On the user agent side, however, predictability and interoperability are desirable in the parsing, rendering and interactional aspects of HTML processing. In order to maximize flexibility in the design of authoring tools, it may be decided to impose only minimal requirements on such implementations. If this is indeed the position of the working group, then suggestions related to @lang handling by authoring tools may be out of scope.
Received on Tuesday, 31 July 2007 08:17:48 UTC