W3C home > Mailing lists > Public > public-vocabs@w3.org > October 2011

Re: RDFa 1.1 Lite

From: Manu Sporny <msporny@digitalbazaar.com>
Date: Fri, 21 Oct 2011 22:16:05 -0400
Message-ID: <4EA22765.608@digitalbazaar.com>
To: W3C Vocabularies <public-vocabs@w3.org>
On 10/21/2011 01:08 PM, Jason Douglas wrote:
> There's a lot to like here.

Glad to hear that. Jason, apologies for the late-joining on my part - 
but are you actively involved with the Rich Snippets stuff at Google, or 
just interested in it from a design perspective?

> However, the two biggest issues raised in
> the discussion at the workshop seem unaddressed / unacknowledged to me
> by this document:

Just to clarify - the document was meant to be a quick and dirty 
introduction - not a complete specification. So, there were many things 
that were left unsaid/unexplained.

> 1/ Why both property and rel?  There was concern expressed that
> implementers wouldn't understand the difference (honestly, I'm not sure
> I do) and it wasn't made clear why having both is necessary. If the
> distinction is literal value vs. url, why not use value precedence (like
> itemprop in microdata)?

You could say that the distinction is "literal value vs. URL", but 
that's a bit of an over-generalization. @rel exists for a number of reasons:

1. It has always been the mechanism used in HTML to specify relevance
    (aka: relationship information) to links on the Web.
2. We found that often people want to create a link relationship and
    express a literal value at the same time, on the same element.
    Having @rel and @property allows them to do that without having
    to change their document structure.
3. The use was re-inforced in the Microformats community after
    extensive study of usage patterns and it seems like people
    were getting its usage correct.
4. Creative Commons supports the usage of @rel and many of their
    implementers get it.

In the 3 years since RDFa 1.0 has been in the field, we see very little 
abuse of @rel. In fact, we see far more examples of correct @rel usage 
in RDFa 1.0 than we do incorrect usage. So, while I do understand the 
concern, is there any publicly available study or data to back up the 
concern?

The reason we didn't go with value precedence is for the reasons listed 
above. Value precedence also doesn't work when you really want to 
express a literal, but there just happens to be an @href on the element. 
We had considered this during the design of RDFa 1.0, but found it to 
not be adequate for the use cases we were considering for RDFa.

Is the existence of @rel a deal-break for anyone? If so, why?

> 2/ Layer-ability on existing markup.  That's probably too abstract a
> label... basically, implementers have found the itemref feature of
> microdata to be useful in layering vocabulary on top of existing pages.

I have a hard time believing this claim. I've found @itemref to be one 
of the most difficult parts of Microdata to use, and have not really 
found it to be that helpful in most documents. Do you have any numbers 
on the number of documents that use @itemref, or the frequency in which 
it is used on the Web?

>   For example, if the name of a product is in one part of the DOM tree,
> but it's properties are in an entirely different one, how do you combine
> them?

Use @about.

> If both the about URI and the markup
> with the about attribute have RDFa in them, how are those properties and
> types combined?

Like this:

<div about="#jason" typeof="schema:Person">
    <span property="schema:name">Jason Douglas</span>
    ... other stuff about Jason ...
</div>
... Lots and lots of HTML ...
<span about=#jason" property="schema:jobTitle">Product Manager</span>

Any time that you want to talk about the same subject again - you just 
use the same @about value. Make sense?

-- manu

-- 
Manu Sporny (skype: msporny, twitter: manusporny)
Founder/CEO - Digital Bazaar, Inc.
blog: Standardizing Payment Links - Why Online Tipping has Failed
http://manu.sporny.org/2011/payment-links/
Received on Saturday, 22 October 2011 02:16:30 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 22 May 2012 06:48:56 GMT