Re: Extension syntax Was: Re: Updated Example

Richard,

Yes, I noticed that Amazon does not use rich snippets. Then again, 
Amazon is Amazon and I presume that either through agreement or simply 
because Amazon is Amazon Google makes some effort to interpret their 
data. I'm not under any illusion that libraries will get the same 
treatment. :-(

Barnes & Noble places the encoding (MP3) and abridgement (unabrdiged) in 
the title element, since it's pretty much a given that the title will be 
displayed. Clearly these elements are important from the point of view 
of selling a product, but I would wager that they are also important for 
library users.

<span class="name">Steve Jobs (MP3 - Abridged)</span>

That's similar to the way that Amazon puts series information into the 
title so that will show. Heck, we could put any information we want into 
the title element and know that it will be displayed... but those titles 
make matching between sources difficult.

I agree that all we can do is present the data and hope that it gets 
used. I think that's the approach I am trying to take.

kc

On 3/2/13 10:44 AM, Richard Wallis wrote:
> In data terms you got what you would expect to.
>
> As to rich snippets displaying things, this case gives some interesting
> insights.
>
> The Amazon page you end up on
> <http://www.amazon.com/Steve-Jobs-Walter-Isaacson/dp/1442346949/ref=tmm_abk_
> title_0> is firstly not encoded with Schema so Google is not doing any rich
> snippets processing on it. Secondly it is pointing you to individual
> expressions, each of which have information such as [Abridged,
> Audiobook] and [Audio CD] appended to the title.  Amazon don't seem to be
> picking abridgement out anywhere other than in the titles of pages (which
> are used, formatted slightly differently, as links links on similar pages)
>
> If you take a look at this Barnes & Nobel page
> <http://www.barnesandnoble.com/w/steve-jobs-walter-isaacson/1104099551?ean=9
> 781442346284>.
>
> The following is in the Product Details section: 'Edition
> description: Unabridged'
>
> This page is coded up with Schema using microformats - you can see what rich
> snippets makes of it here:
> <http://www.google.com/webmasters/tools/richsnippets?url=http%3A%2F%2Fwww.ba
> rnesandnoble.com%2Fw%2Fsteve-jobs-walter-isaacson%2F1104099551%3Fean%3D97814
> 42346284&html=> It is defined as a Schema Product with attached ratings,
> reviews and offers. They miss the opportunity to describe any CreativeWork
> stuff also abridgement is not captured in the structure.
>
> It's interesting seeing how the B&N pages are put together - they definitely
> treat what they describe as products. Being a creative work is secondary -
> so secondary that they do not even mention it.
>
> There are no specific conclusions I can draw from this example, other than
> if you give someone some tools they will probably make something different
> to what you would do with those tools.
>
> ~Richard.
>
>
> On 02/03/2013 18:16, "Karen Coyle" <kcoyle@kcoyle.net> wrote:
>
>> Jason, I've been testing with the Google Rich Snippet tool, which,
>> albeit not a definitive answer to how the schema.org data will appear in
>> search engines, is the only thing I have to go on in order to judge how
>> they might display a rich snippet with the minimum of specific
>> information about the resource or the source. With
>>
>> <div itemscope="">
>> <p>
>> <meta itemprop="abridged" content="true">Abridged</p>
>> <p><span itemprop="abridged">Abridged</span></p>
>> </div>
>>
>> you get:
>>
>> property:
>> abridged: true
>> abridged: Abridged
>>
>> Now, perhaps we can assume that search engines would interpret the
>> boolean version and display that as "Abridged". I honestly do not know
>> what to expect. But I do think there is value in having the rich snippet
>> display carry this information. It's hard to do this in email, but if
>> you do a search on google for "steve jobs audiobook" you get a fairly
>> rich display:
>>
>> Steve Jobs: Walter Isaacson: 9781451648539: Amazon.com: Books
>> www.amazon.com/Steve-Jobs-Walter-Isaacson/.../1451648537
>>    Rating: 4.3 - 1348 reviews - $18.78 - In stock
>> ... Textbooks · Sell Your Books · Steve Jobs and over one million other
>> books are available for Amazon Kindle. .... Audio, CD, Abridged,
>> Audiobook, $19.79, --, -- ...
>>
>> Could we use "additionalType"? There is a Wikipedia page for
>> "abridgement" but nothing that would distinguish between the states of
>> being or not being abridged, so I don't see "additionalType" as helping
>> us out here.
>>
>> I know that some are seeing schema.org as data markup, and it is that,
>> but I am focused beyond the markup of the HTML page to the search engine
>> display and how we can get key information into that very brief display,
>> and then link that to library holdings. A publisher would want the
>> snippet linked to their product. In either case, I see information like
>> "Audiobook" "Abridged" and "CD" or "MP3" to be valuable for the rich
>> snippet display, not just for the HTML display that the rich snippet
>> derives from.
>>
>> kc
>>
>> On 3/1/13 4:19 PM, Jason Ronallo wrote:
>>> On Wed, Feb 27, 2013 at 3:41 PM, Karen Coyle <kcoyle@kcoyle.net> wrote:
>>>> On 2/25/13 6:11 AM, Jason Ronallo wrote:
>>>>
>>>>> It is fine to include the span, but note that a conformant Microdata
>>>>> processor does nothing with the span. It would not associate the text
>>>>> of the span with the meta property and value in any way.
>>>>> Also in HTML5 elements like <meta> and <link> are void elements [1]
>>>>> and only have a start tag, so they do not have to include the
>>>>> self-closing slash.
>>>>>
>>>>
>>>> Jason,
>>>>
>>>> If the microdata processor will not use the span, then we should go back to
>>>> the simple itemprop + type=text, because the display is key. I ran some
>>>> examples through the Google microdata validation, and the "content" will be
>>>> displayed, thus
>>>>
>>>> <meta itemprop="abridged" content="true">
>>>>
>>>> will display as "true", which is what we were trying to avoid.
>>>
>>> What will display as "true" where? The meta element does not display
>>> to a user at all. It is still OK to write it as the following:
>>>
>>> <meta itemprop="abridged" content="true">Abridged
>>>
>>> Any text you want can be displayed next to (or not next to) the text
>>> "Abridged". There are lots of examples where meta or link elements are
>>> used to convey the data while text is still displayed on the page for
>>> humans.
>>>
>>> Then the consumer of the data would then know to expect a Boolean
>>> value for the abridged property. A conformant Microdata processor for
>>> instance would not see the text "Abridged", but it does not have to to
>>> understand the meaning from the meta element. The consumer could
>>> display the data (abridged=true) however it wanted to.
>>>
>>> The point of using the value of "true" here is that ISSUE-14 suggests
>>> that processors would know to treat the "true" value as a
>>> http://schema.org/True value, and that this solution for simple
>>> booleans works for both RDFa Lite and Microdata. The issue hasn't been
>>> resolved, but it appears to be a good solution.
>>>
>>> Or am I missing a requirement of the abridged property where it is not
>>> a simple boolean, either abridged or unabridged?
>>>
>>> Jason
>>>
>
>
>

-- 
Karen Coyle
kcoyle@kcoyle.net http://kcoyle.net
ph: 1-510-540-7596
m: 1-510-435-8234
skype: kcoylenet

Received on Saturday, 2 March 2013 19:40:24 UTC