Re: Review RDFa LC Primer & Lite documents

Guus,

this is my answer as editor of the Primer document.

Because this document is not Rec Track, all the updates of the document will not be done right now; we will concentrate on that once the Rec Track documents are published as CR. Nevertheless, I have already folded in some of your comments in the editor's draft:

http://www.w3.org/2010/02/rdfa/sources/rdfa-primer/Overview-src.html

Some more comments below.

[skip]

> 
> 
> ====================================
> RDFa 1.1 Primer:
> http://www.w3.org/TR/2011/WD-rdfa-primer-20111208/
> ====================================
> 
> This is also an easy-to-read document. My main problems are with the
> choice of examples, in particular in the "Going Deeper" section.

Yes, they are a little bit a hodge-podge, collected from other documents. Manu and I will have to go through them, and maybe review/change them


> 
> 1. The social network example in Sec 3.2 contains three blank nodes,
> despite the fact that you stated a few lines before that this is not
> good practice.  I suggest to include an "about" attribute for Bob/Eve/Manu.
> 
> Question: would it not be good practice to leave out the typeof
> attribute for Bob/Eve/Manu?  I would include the knows property earlier
> in the examples, and then we can let the RDFa processor work this out.
> 

It so happens that if we follow that route, the resulting RDFa structure becomes, in fact, more complicated. Not at the point you refer to, but some examples below, where the @property="knows" is used. To achieve the same output, but keeping the same level of simplicity, the right attribute to use at that point would be @resource and not @about (to ensure that the person object is the object of @property="knows"). However, at this point, @resource is not part of RDFa Lite... 

Going down this route is, therefore, related to an open issue in RDFa Lite:

http://www.w3.org/2010/02/rdfa/sources/rdfa-lite/#about

Ie, I leave this issue pending for the primer for the time being. Also, as Manu referred to in his reply: the current structure, though uses FOAF, is pretty much following the pattern used by schema.org. Which shows that a blank node (well, an anonymous node...) is syntactically useful, so maybe the good practice issue should be reformulated instead...


> 2. My guess is that the explanation in Sec. 3.1. of why "rel" is needed
> and how it works is too brief for newcomers. A nontrivial notion plus
> rationale is described here in one short paragraph. If I try to think as
> a newcomer and consider the example below this paragraph, e.g.
> 
>  <li typeof="Person">
>        <a rel="homepage" href="http://example.com/bob/"
> property="name">Bob</a>
> 
> then I would have real problems understanding what is happening (e.g.
> how the three property values are related). I suggest to explain this in
> significant more detail.

I added/changed that parg.


> 
> 3. Figure 7 has again blank nodes; suggest to try to get rid of these;
> see 1st comment.
> 


See my answer above.

> 4. Sec. 3.3, 2nd example
> 
>  [...]
>  <h3 rel="creator" resource="#me">Alice</h3>
>  [...]
> 
> You explain why the example uses |resource" instead of "href", but not
> why you use "rel" instead of "property" (like the CC license link in the
> last example of 2.4. Clarify this.

The clarification is that this is a leftover and @property could be used just as well:-) I changed the example.


> 
> 5. Genera: " avoid terms like "of course" and "easily"

I tried, by removing most of them...

Thanks again Guus

Ivan


> 
> 
> 
> 
> 
> 
> 
> 
> 
> 


----
Ivan Herman, W3C Semantic Web Activity Lead
Home: http://www.w3.org/People/Ivan/
mobile: +31-641044153
FOAF: http://www.ivan-herman.net/foaf.rdf