W3C home > Mailing lists > Public > public-rdfa-wg@w3.org > December 2010

Re: Obsolete tests in RDFa 1.1 Test Suite

From: Gregg Kellogg <gregg@kellogg-assoc.com>
Date: Fri, 17 Dec 2010 13:23:21 -0500
To: Toby Inkster <tai@g5n.co.uk>
CC: "public-rdfa-wg@w3.org WG" <public-rdfa-wg@w3.org>
Message-ID: <174E22D6-25DA-411A-8BF0-A0F9946E7601@kellogg-assoc.com>
On Dec 17, 2010, at 6:50 AM, Toby Inkster wrote:

> On Tue, 14 Dec 2010 20:38:27 -0500
> Gregg Kellogg <gregg@kellogg-assoc.com> wrote:
> 
>> test 0086: NO triple for a non-reserved @rel value, (negative test) 
>> Now generates a triple with default vocabulary 
> 
> In XHTML+RDFa the default vocabulary is initially undefined, so this
> should still generate no triple.

Thanks, I had missed that! It was defined in earlier drafts but I missed when it was changed to be undefined.

>> test 0105: inner @rel neither CURIE nor LinkType  Now generates a
>> triple with default vocabulary
> 
> Ditto.
> 
>> test 0116: Reserved word used in @property should not generate triple
>>  Now generates a triple with default vocabulary
> 
> Should not generate a triple using default vocabulary, as XHTML+RDFa
> has no initial default vocabulary. It should generate a triple though
> owing to the term being defined in the default profile.
> 
>> test 0125: datatype XMLLiteral with other embedded RDFa, (negative
>> test)  XMLLiterals are now deeply parsed
> 
> Note to self: need to implement this change myself!
> 
>> test 0142: xmlns prefix 'xml' with correct URI  <xml:test> is a
>> legitimate URI
>> test 0160: Undeclared prefix 'xml' <xml:test> is a legitimate URI
>> test 0162: Undeclared prefix 'xmlns' <xmlns:test> is a legitimate URI
> 
> Not sure of the best solution for these. According to XML Namespaces,
> the QName prefixes 'xml' and 'xmlns' are predefined. This needn't have
> effect on how CURIEs work, but we may want to consider building them in.

Perhaps this should be added as an ISSUE? Personally, I think that allowing elements that are intended to be CURIEs, but for which a prefix mapping is missing is a problem. I'd rather see some further constraints on valid URI schemes to avoid aggressive generation of URIs where CURIEs were intended.

I'll go ahead and remove 0125, but leave the rest until there's some further resolution. 0142, 0160 & 0162 I'll change to "buggy".

> -- 
> Toby A Inkster
> <mailto:mail@tobyinkster.co.uk>
> <http://tobyinkster.co.uk>
> 
Received on Friday, 17 December 2010 18:24:49 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:19:50 UTC