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

Re: Syntax for itemprop breadcrumb

From: Aaron Bradley <aaranged@yahoo.com>
Date: Mon, 19 Dec 2011 09:00:33 -0800 (PST)
Message-ID: <1324314033.59562.YahooMailNeo@web161003.mail.bf1.yahoo.com>
To: Jason Douglas <jasondouglas@google.com>, Gregg Kellogg <gregg@kellogg-assoc.com>
Cc: John Panzer <jpanzer@google.com>, Jeni Tennison <jeni@jenitennison.com>, public-vocabs <public-vocabs@w3.org>, Tim van Oostrom <tim@depulz.nl>
Thanks everyone for weighing in on this (and sorry for the double post - a blank message seems to have been delivered rather than this one).

I was surprised to see John Panzer's parser results supporting Jeni
Tennison's option C - "the examples are wrong and the itemprop should
be on individual breadcrumb items" - not because it doesn't make sense,
but that the examples are consistently wrong across the board in regard to
schema.org microdata.

I checked on the main Google Breadcrumb article...
http://support.google.com/webmasters/bin/answer.py?hl=en&answer=185417
... and discovered there that, indeed, breadcrumbs are marked up individually,
both in RDFa and for (data-vocabulary.org) microdata.

A snippet of their microdata example:
<div itemscope
itemtype="http://data-vocabulary.org/Breadcrumb";
  <a href="http://www.example.com/dresses";
itemprop="url"
    <span itemprop="title">Dresses</span
  </a›
</div
<div itemscope
itemtype="http://data-vocabulary.org/Breadcrumb";
  <a href="http://www.example.com/dresses/real";
itemprop="url"
    <span itemprop="title">Real Dresses</span
  </a›
</div

A friend weighed in on my G+ thread with an observation very similar to yours,
Jeni:
"Well.. if you read the specification the Expected Type for
"breadcrumb" is text, so it doesn't really matter if its a link or
not. Semantically, a breadcrumb is really only a category marker and it
shouldn't matter if its a link or not. In modern uses, it is hyperlinked for
usability, but technically, it doesn't need to be linked for a human or bot
to parse out the categorization structure."

So my takeaways are two-fold:

1.  The schema.org examples are incorrect, and each individual breadcrumb should
be declared with itemprop.
2.  While schema.org describes the breadcrumb property as "[a] set of
links" the expected property is text rather than URL, which at least
technically affirms that breadcrumbs need not be links.

Thanks!

>________________________________
>From: Jason Douglas <jasondouglas@google.com
>To: Gregg Kellogg <gregg@kellogg-assoc.com
>Cc: John Panzer <jpanzer@google.com>; Jeni Tennison
<jeni@jenitennison.com>; public-vocabs <public-vocabs@w3.org>; Tim
van Oostrom <tim@depulz.nl>; Aaron Bradley <aaranged@yahoo.com
>Sent: Saturday, December 17, 2011 10:18:36 AM
>Subject: Re: Syntax for itemprop breadcrumb
>
>
>On Sat, Dec 17, 2011 at 9:00 AM, Gregg Kellogg
<gregg@kellogg-assoc.comwrote:
>
>Interesting that this is one place where RDFa @rel does what you want.
>
>
>I was under the impression @rel was still allowed in RDFa Lite, just not
required.  Named links seems like a use case where it does make things simpler.
>
> 
>It also requires @inlist to maintain the breadcrumb order.
>>
>>
>>RDFa Lite will have the same restrictions as microdata, requiring a
@property on each <aand depend on the vocabulary-specific handling of
schema:breadcrumb to maintain order in RDF.
>>
>>Gregg Kellogg
>>Sent from my iPhone
>>
>>On Dec 17, 2011, at 7:43 AM, "John Panzer"
<jpanzer@google.comwrote:
>>
>>
>>For my parser, (c) appears to be the answer today.
>>>On Dec 17, 2011 12:30 AM, "Jeni Tennison"
<jeni@jenitennison.comwrote:
>>>
>>>Hi,
>>>>
>>>>Just to point out that in something like:
>>>>
>>>><div itemprop="breadcrumb"
>>>> <a href="category/books.html">Books</a
>
>>>> <a
href="category/books-literature.html">Literature &
Fiction</a
>>>> <a
href="category/books-classics">Classics</a
>>>></div
>>>>
>>>>the 'breadcrumb' property actually takes the value
"Books Literature & Fiction Classics" (whitespace
normalised for brevity). So whether or not they have <aelements wrapped
around individual words doesn't matter: a conformant microdata processor
won't see them anyway.
>>>>
>>>>I've raised this before, but it is still not clear to me
whether
>>>>
>>>>a. schema.org consumers only want that plain text string;
>>>>b. schema.org consumers are preserving the HTML content
(contrary to the microdata spec); or
>>>>c. the examples are wrong and the itemprop should be on
individual breadcrumb items
>>>>
>>>>Cheers,
>>>>
>>>>Jeni
>>>>
>>>>On 16 Dec 2011, at 23:30, Aaron Bradley wrote:
>>>>
>>>>>
>>>>>
>>>>>Thanks for weighing in, Tom.
>>>>>
>>>>>In this case, though, all the examples explicitly show all
breadcrumb links belonging to one itemprop - what they omit is an additional,
unlinked "breadcrumb" component.
>>>>>
>>>>>See on http://schema.org/WebPage:
>>>>>
>>>>><div itemprop="breadcrumb"
>>>>>  <a
href="category/books.html">Books</a
>>>>>  <a
href="category/books-literature.html">Literature &
Fiction</a
>>>>>  <a
href="category/books-classics">Classics</a
>>>>></div
>>>>>
>>>>>
>>>>>By the way, the separate page identifier in the <h1
I wouldn't have a problem handling (it's not a breadcrumb) - it's
the additional component on the same line.  E.g. (and this is the syntax I'm
leaning towards - *not* including the unlinked item in the breadcrumb
declaration):
>>>>>
>>>>>
>>>>><div
>>>>>
>>>>><span itemprop="breadcrumb"
>>>>>  <a
href="category/books.html">Books</a
>>>>>  <a
href="category/books-literature.html">Literature &
Fiction</a
>>>>>  <a
href="category/books-classics">Classics</a></span
>>>>>  Boring Classics
>>>>></div
>>>>><h1>Boring Classics</h1
>>>>>
>>>>>
>>>>>----- Original Message -----
>>>>>>From: Tim van Oostrom <tim@depulz.nl
>>>>>>To: public-vocabs@w3.org
>>>>>>Cc:
>>>>>>Sent: Friday, December 16, 2011 2:26:37 PM
>>>>>>Subject: Re: Syntax for itemprop breadcrumb
>>>>>>
>>>>>>Hi Aaron,
>>>>>>I personally interpreted breadcrumb like:
>>>>>>
>>>>>><div itemscope
itemtype="http://schema.org/Thing";;
>>>>>>   <a href="category/books.html"
>>>>>>itemprop="breadcrumb">Books</a
 
>>>>>>   <a
href="category/books-literature.html"
>>>>>>itemprop="breadcrumb">Literature&
 Fiction</a 
>>>>>>   <a href="category/books-classics"
>>>>>>itemprop="breadcrumb">Classics</a
 
>>>>>>   Boring Classics
>>>>>><h1 itemprop="name">Boring
Classics</h1
>>>>>>  </div
>>>>>>
>>>>>>
>>>>>>1 breadcrumb per Item/Link (semantically more obvious
and less work to
>>>>>>determine what is what?)
>>>>>>
>>>>>>You'd have a list of breadcrumb(s). This should
however be an ordered list.
>>>>>>
>>>>>>- Tim
>>>>>>
>>>>>>>  In breadcrumb display, the unlinked current page
or section is often
>>>>>>displayed in the same line as the linked parents.
 E.g.:
>>>>>>>
>>>>>>>  <div><a
href="/">Home</a | About
>>>>>>us</div
>>>>>>>
>>>>>>>  Should this unlinked portion be ("About
us" in the example above)
>>>>>>be included in the breadcrumb itemprop or excluded from
it? 1 or 2 below?
>>>>>>>
>>>>>>>
>>>>>>>  1 - Unlinked portion part of breadcrumb itemprop
>>>>>>>
>>>>>>>  <div itemprop="breadcrumb"
>>>>>>>    <a
href="category/books.html">Books</a 
>>>>>>>    <a
>>>>>>
href="category/books-literature.html">Literature&
>>>>>>Fiction</a 
>>>>>>>    <a
href="category/books-classics">Classics</a
>>>>>>>
>>>>>>>    Boring Classics
>>>>>>>  </div
>>>>>>>
>>>>>>>  2 - Unlinked portion not a part of breadcrumb
itemprop
>>>>>>>
>>>>>>>  <div
>>>>>>>  <span itemprop="breadcrumb"
>>>>>>>    <a
href="category/books.html">Books</a 
>>>>>>>    <a
>>>>>>
href="category/books-literature.html">Literature&
>>>>>>Fiction</a 
>>>>>>>    <a
>>>>>>
href="category/books-classics">Classics</a></span
>>>>>>>
>>>>>>>    Boring Classics
>>>>>>>  </div
>>>>>>>
>>>>>>>  None of the schema.org examples show this use
case.
>>>>>>>
>>>>>>>  See also this same question on a G+ post - feel
free to comment there.
>>>>>>> 
https://plus.google.com/106943062990152739506/posts/Bf5ZYWkVtM1
>>>>>>>
>>>>>>>  Thanks,
>>>>>>>  Aaron
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>
>>>>>
>>>>>
>>>>
>>>>--
>>>>Jeni Tennison
>>>>http://www.jenitennison.com
>>>>
>>>>
>>>>
>
>
>
Received on Monday, 19 December 2011 17:04:47 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 22 May 2012 06:48:58 GMT