- From: Leif Halvard Silli <lhs@malform.no>
- Date: Fri, 17 Jul 2009 03:47:26 +0200
- To: Mark Birbeck <mark.birbeck@webbackplane.com>
- CC: Steven Pemberton <Steven.Pemberton@cwi.nl>, Julian Reschke <julian.reschke@gmx.de>, Manu Sporny <msporny@digitalbazaar.com>, RDFa Developers <public-rdf-in-xhtml-tf@w3.org>, HTMLWG WG <public-html@w3.org>
Mark Birbeck On 09-07-16 18.34: > In other words, it's quite fundamental that any information required > for processing a graph -- prefix mappings, default language, base > URLs, character encodings, or anything else you can think of -- must > be obtained prior to processing that graph. > On Thu, Jul 16, 2009 at 5:02 PM, Steven Pemberton: >> [...] I have to say I disagree. >> By the time you have the graph, you have done the RDFa processing. The >> statement "This document uses a profile of http://example.com/profile" is >> true, whether or not the profile has been used in the generation of the >> graph. There is no contradiction, Gödelian or other. >> On Tue, 14 Jul 2009 11:06:59 +0200, Mark Birbeck >>> Hi Julian, >>> So the key point is not 'don't use RDFa to describe how to interpret >>> RDFa' -- the key message here is that 'instructions on how to process >>> a graph cannot go into the graph itself'. I appear to conclude like Steven: If you write <html about="[_:Göteborg]" > then you have used a character which may depend on a child element of <head> for its correct interpretation w.r.t. the character encoding. (Ditto for e.g. @title, which HTML5 permits inside <html>. Whereas HTML 4 only permitted ASCII content inside its attributes - @lang and @dir.) In HTML 4, metadata is more or less limited to the <head> element [1]. Thus it was logical with <head profile=...">. But with RDFa, any element (even <html> and <head> itself, I gather) may have RDFa attributes. So, to be completely generic, shouldn't one switch from <head profile=...> to <html profile=... >? Hence, if <head profile="..."> actually is good enough, then it ought to be good enough with <link rel="profile" ...> as well, no? If RDFa is "a CSS for meaning"[2], then - just as the <html> element may be styled via a child element of <head> (which in turn may point to a style found in an external CSS document), then likewise should it not matter if the profile URI is found in a child of <head> instead of inside the <head> attribute. No? [1] http://www.w3.org/MarkUp/2009/rdfa-for-html-authors#Generalise [2] http://www.w3.org/MarkUp/2009/rdfa-for-html-authors#Introducti -- leif halvard silli
Received on Friday, 17 July 2009 01:48:10 UTC