Review of RDFa in XHTML: Syntax

My apologies for only getting this out the day before the f2f...I understand if we can't incorporate them into the discussion.

Review of RDFa in XHTML: Syntax [1] 

In general this is a really exciting document. In addition to a very thorough description of RDFa and how to extract triples from HTML it also has some of the most succinct and clear descriptions of RDF I've ever read.

The 21-Sep-2007 draft I read still had quite a few editorial notes in it, which I didn't address. I also seemed to notice RFC2119 keywords that were not capitalized. My DTD skills are also very rusty, so I was not able to give the appendices the attention they deserved.

My specific comments can be found below:

Section 3.3 

URIs are most commonly used to identify web pages, but RDF makes use of them as a way to to provide unique identifiers for concepts.

Since RDF can use URIs to identify documents as well as concepts shouldn't this be:

URIs are most commonly used to identify web pages, but RDF uses them to also provide unique identifiers for concepts.

Section 3.8

Are CURIEs going to be defined in another document ( or not? If so does the CURIE doc go req track as well? The TR/curie document does not mention 'safe' CURIEs anywhere, although they are described in section 4.2. Would it be possible to include an informative reference for what the well known limitation of QNames is?

Section 4.1

Could an example of a "higher level protocol" be given? For example: HTTP?

Section 5.3

It might be nice if the processing steps have fully qualified numbers, like so that they can be referred to easily.

Section 5.3, item 1, bullet 4

Note that this means that the value of the [current element identifier] will not be the same as the [current resource] since in the absence of an attribute to set its value will be set to a [blank node] ...

If the current resource was set with @about and the current element identifier is set to the value of @about won't they be the same? I realize there can be  situations where they might not be the same, but saying they "will not be the same" seems to strong somehow, but perhaps I'm misunderstanding.

Section 5.3, item 2, bullet 2

The [current object literal] will be set as a [plain literal] if @content is present or the body of the [current element] contains only text (i.e., there are no child elements), or the body of the [current element] does have child elements but @datatype has an empty value.

When I first read this I thought it meant the @datatype attribute wasn't present for the element, but then learned later if it's not present at all then that's a trigger for treating the literal as an XML literal. It might be clearer to say:

or the body of the current element does have child elements and a @datatype attribute but the @datatype is set to an empty value: (e.g. '').

Section 5.3, item 3, bullet 1

Predicates for the [current object literal] can be set by using @property. If present, the attribute must contain one or more [curie]s each of which is converted to an absolute URI using CURIE process rules ... 

CURIE processing rules are mentioned several times, but I wasn't able to find them in this document. Are they available somewhere?

Section 5.3, item 6

I found the discussion of the 'stack' to be confusing, and found myself wondering what the implications were, and then kind of skipped over it. Then I read Diego's thoughts on this and I agree with his analysis. I liked his solution a) better than b) ... but either would be an improvement.

Section 6

I think it would be clearer to rename this section to "RDFa processing in detail".

Section 6.1.1 

When triples are created they will always be in relation to the [current resource].

Except when they are created in relation to the current element identifier as in Section 5.3.4.

Maybe add base to the html so the n triples can easily be written?

When a triple predicate has been expressed using @rel, the @href on the [RDFa statement]'s element is used to indicate the object as a [URI reference].  Its type, just like that of @about is a URI:

might read better as:

When a triple predicate has been expressed using @rel, the @href on the [RDFa statement]'s element is used to identify the object with a [URI reference].

I left off the bit about href because @about is URIorCURIE not just URI. Is href the same?

Section 9.2.5, 2nd paragraph: 




Received on Sunday, 7 October 2007 13:19:41 UTC