W3C home > Mailing lists > Public > public-html-data-tf@w3.org > November 2011

Re: Guidance on publishing in multiple formats

From: Jeni Tennison <jeni@jenitennison.com>
Date: Wed, 9 Nov 2011 18:52:50 +0000
Cc: HTML Data Task Force WG <public-html-data-tf@w3.org>
Message-Id: <AB574866-1D2A-49DA-ADB8-00869AC1DC02@jenitennison.com>
To: Gregg Kellogg <gregg@kellogg-assoc.com>
Gregg,

On 9 Nov 2011, at 03:15, Gregg Kellogg wrote:
> Looks really good Jeni. One small comment in the examples: you use both <meta> elements with datesTime values. now that <time> seems to be back, these would more naturally be expressed using the <time> element, which would also preserve the literal dataType:
> 
>  <time itemprop="dtstart" property="startDate" datetime="2016-04-21T20:00:00" content="2016-04-21T20:00:00" datatype="xsd:dateTime/>

Yes. I was lazy and copied the example directly from the schema.org documentation. The way microdata *was* specified, it would have been invalid to use the <meta> element there. I don't know whether, when the change from <time> to <data> is reverted, the restriction on how date/times are specified will be re-introduced or not.

> One of the proposed changes for HTML+RDFa is to process @datetime in a similar manner to microdata, so the @content and @datatype attributes could be removed. It hasn't come up for a vote yet, but it seems non-controversial.

Yep, that will make it much better. That's rdfa-ISSUE-97 isn't it? [1]

> Also, there is a proposal in RDFa to give @property many of the same attributes as @rel, which would also allow the markup to be reduced. Presuming we vote for this on Thursday, elements could be simplified further:
> 
>  <div class="location" itemprop="location" property="location" typeof="http://schema.org/Place>
>    <a property="url" href="wells-fargo-center.html">
>        Wells Fargo Center
>    </a>
>    <div property="address" typeof="http://schema.org/PostalAddress">
>      <span property="http://schema.org/addressLocality">Philadelphia</span>,
>      <span property="http://schema.org/addressRegion">PA</span>
>    </div>
>  </div>
> 
> Note how @typeof now bonds to the object, rather than the subject. This is also true if you continue to use @rel.

I know that you've been proposing changes about this and I was hoping that they would help with this example, including the rather gnarly issue about chaining due to the nested @href. Is there a publicly available test implementation of the changes?

> Also, unless @vocab is used, @property values must be be spelled out using CURIEs or IRIs. In this sense, the Mixing Syntaxes example is currently incorrect.

The top-most div had a @vocab on it:

<div class="vevent"
  itemscope itemtype="http://microformats.org/profile/hcalendar#vevent"
  about="_:event" vocab="http://schema.org/" typeof="Event">
  ...
</div>

but it was confusing because the nested @typeof attributes used full IRIs. I've changed this to specify @vocab wherever there's a @typeof.

I'm tempted to change it to use the schema: prefix throughout since this is more reliable than using @vocab (less likely to get lost through copy/pasting) as per the guidance earlier in that page.

> @itemprop + @itemscope becomre more like @property + @href/@src/@resource and @typeof. Note that without @typeof (even an empty value), @property will not cause chaining in the proposed changes to RDFa 1.1.

Great :)

Jeni

[1] http://www.w3.org/2010/02/rdfa/track/issues/97
-- 
Jeni Tennison
http://www.jenitennison.com
Received on Wednesday, 9 November 2011 18:53:16 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 9 November 2011 18:53:17 GMT