See also: IRC log
Previous: http://www.w3.org/2009/09/24-rdfa-minutes.html
<scribe> ACTION: Ben to update JS xmlns getter code on implementors' guide for xhtml mime type support [recorded in http://www.w3.org/2009/09/17-rdfa-minutes.html#action02] [CONTINUES]
Manu: The core of this issue is that the test cases shove every namespace that is active into the top elements in the literal.
Shane:Ivan seemed to say we should canonicalize...
...Ben seemed to disagree.
Shane: This has to do with whether an
XMLLiteral is supposed to be canonicalized.
... You don't have to canonicalize XMLLiterals in RDF.
Manu:The underlying issue is that processors might generate different xmlliteral. It is possible to tighten up our spec to prevent this. This will affect real world use.
Manu: So multiple processors might raise xmlliterals that are not byte by byte identical and might not compare correctly via sparql
Shane: If we bring HTML into the mix... what happens when there is no defauilt xml namespace
Manu:We could require XHTML namespace to be emitted when generating xml literals from HTML input.
Manu:Should we make it a requirement that a processor assume the default namespace is the xhtml namspace when processing RDFa from an HTML source? Probably. It would not have any negative effects.
Shane: Why does it matter if there are extra xml namespace declarations embedded in an XML Literal ?
Manu: It matters because the rest of the stack assumes the input is canonical XML... even though that is not really a requirement. XMLLiteral is a corner case.... so the rules here won't mess up that much in the wild. We should require that conforming processors always canonicalize an XMLLiteral.
Shane:This would be a conformance change.
Manu:Can we issue an errata that says we are going this way in the future and implementors should do it now.
Shane:Yes, we can.
Shane:What should we do with the test suite? Change the
tests so they require canonical input?
Manu:Yes.
Shane:Add a comment to the tests so they indicate that a failure is okay now, but won't be soon.
<scribe> ACTION: ShaneM to craft errata text about canonical XMLLiterals [recorded in http://www.w3.org/2009/10/08-rdfa-minutes.html#action02]
<msporny> http://rdfa.digitalbazaar.com/test-suite/
<msporny> TCs: 131, 134, 140, 142, 147, 154
Manu: Test Case 131: tests to ensure that characters defined as whitespace by XML are treated as whitespace by a processor
<msporny> +1 for passing 131
+1
<Steven> +1
Test 131 approved.
Manu:Test case 134:
Manu:In XHTML mode, this should generate no triple.
Shane:xmlns spec says http://www.w3.org/TR/REC-xml-names/#ns-decl
Shane:'xmlns' must be in lower case
<Steven> http://www.w3.org/TR/rdfa-syntax/#s_metaAttributes
Shane:XHTML Metainformation Attributes Module redefines rel and rev
Steven:Should we treat reserved words as case-insensitive?
Steven: HTML4 says the author is allowed
to treat rel and rev values as case insensitive... so yes. we should.
... Need to issue an errata about section 9.3 that says they are
treated case insensitively on input and must be transformed to lower
case on ouput in triples
<scribe> ACTION: ShaneM to craft errata about case insensitive reserved words. [recorded in http://www.w3.org/2009/10/08-rdfa-minutes.html#action03]
Test case 134 is NOT YET APPROVED