Mark Nottingham wrote: > Likely it's not too > bad, owing to the bad state of @rel in HTML anyway, but it has > effectively created one more thing to sniff in HTML -- "what rel > convention is in use here?" -- with all of the ambiguity and issues that > entails.
My last email explained why this is factually wrong. Did you miss it? RDFa did not add anything more to sniff in HTML: you're assuming that (@profile, @rel) is all you need to determine @rel, and I'm pretty sure that's incorrect. Take GRDDL, which is a REC, and which can be implemented as an HTML4 @profile. (And which was led by Dan Connolly, whom I trust to know what @profile means.) To make sense of @rel in GRDDL, you're instructed to look at <link rel="transformation" href="specificTransformation.xsl">, which itself tells you *how* to interpret @rel. What is the difference between this level of indirection and that of RDFa's? -Ben
