Internet 3.0 is known as Semantic Internet. The final user doesn’t spot anything different when they visit a site. But for web developers that means that they can use special tags in a site’s HTML code so that others (mainly search engines, spiders and brownsers) can understand what’s being described in a page. They mainly seek to allow for a more powerful contents’ aggregation and more trustworthy search results.
Starting on June the 2nd, 2011, Google, Bing & Yahoo, which would later be joined by Yandex, banded together to create Schema.org – an initiative with the goal of creating and supporting a set of schemes for the tagging of structured data in the sites. This initiative has had a lot of strength from the very start, since these search engines are the main structured information interpreters in the eyes of the public through Rich Snippets.
As we told in another post, the benefits of well-done Rich Snippers are very positive. The additional exposition and the space their search result takes up is a great system to make the message reach the potential clients. If they can be told apart from their competition when the search resuls appear it’s very probable that they experience and increase of click through rates.
Scheme.org has brought forth great consensus on the “vocabulary” we use to refer to data which we want search engines to interpret. However, there’s not that much consensus in how to tag data to be interpreted. There are 3 ways to do that:
- RDFa & RDFa lite
They’re the oldest format but they supplied the basis for the other formats: they remain the simplest of them when it comes to tagging.
They’re built with classes of HTML elements. This means that we’ll configure the class name to a certain value (address, name, etc.): this can turn complicated if we gotta do it in several cases.
<img src=”camiseta.jpg” alt=”shirt” class=”product-thumb”/>
<h4 class=”product-title”>Marca X</h4>
<p class=”p-v”><span>Color:</span> <span>289 Rush</span></p>
<p><a href=”shirt/Brand_X/” title=”Buy Now” rel=”product”>Buy Now</a></p>
Their main problem is their lack of extension: we can’t create our own if it isn’t within the official specifications.
RDFa (Resource Description Framework in attributes) adds extensions at attribute level to insert structured metadata inside Web documents and has both the approval and reccomendation of W3C.
<div xmlns:v=”http://schema.org/” typeof=”v:Product”>
<img property=”v:image” src=”Product_example.jpg” />
<span property=”v:name”>Product example</span>
This choice is more complex but apports more extensions, allowing one to develop their own standards which fit our needs.
They’re an attempt to proportion a simpler way to tag the HTML elements compared to the focusing of RDFa and the microformats. This method got a powerful boost with the creation of Schema.org because Google clearly staked in the microdata evolution, reccomending its using and not supporting RDFa. That was seen like a Google imposition by some sectors so Google made it so that their Rich Snippets support RDFa.
Microdata 1.0 Sample
<div itemscope itemtype=”http://schema.org/Product”>
<img itemprop=”image” src=”Product_example.jpg” />
<span itemprop=”name”>Product example</span>
Given the boost microdata got from Google and Schema.org then RDFa lite showed up as an alternate choice. It’s a subset of RDFa with a simpler sintaxis which is practically the same as in microdata but it has extension capabilities.
RDFa Lite 1.1 Sample
<div vocab=”http://schema.org/” typeof=”Product”>
<img property=”image” src=”Product_example.jpg” />
<span property=”name”>Product example</span>
As we can see, when it comes to simple samples, Microdata and RDFa lite are almost identical. If you examine them in depth then:
– RDFa lite has no need of “iteamrel”, used in microdata to relate objects.
– Microdata doesn’t support “vocab”, used to specifiy a default vocabulary with terms used in the markup.
– Microdata doesn’t support “prefix”, used to mix different vocabularies in the same documents like those brought by Facebook, Google and the open source projects.
Microdata vs RDFa lite
– The microdata processing algorithm is WAY simpler: the data is extracted in a more relieable manner and if there are errorst then it makes it easier for the users to purge their data.
– Both Microdata and RDFa are compatible with the main search engines sniffers like Google, Microsoft, Yahoo! & Yandex. But they’re not compatible with Facebook.
– Over 99% of the dialing microdata can be easily expressed in RDFa Lite 1.1. Conversion from Microdata to RDFa Lite is as simple as a look-up and a replacement of the Microdata attributes with RDFa Lite attributes. On the contrary, Microdata doesn’t support some of the more advanced RDFa functions, like being able to tell feet and meters apart.
– RDFa allows not having to learn Microdata from schema.org and RDFa for Facebook because it supports both schema.org and the Facecook Open Graph Protocol (OGP) using a unique markup: you only need use RDFa for both.
– RDFa Lite 1.1 is upper compatible with RDFa 1.1, which allows one to migrate without trouble to a language richer in functions as their Linked Data needs incease. Microdata doesn’t allow for any of the more advanced functions provided by RDFa 1.1.
We hope this post was helpful.