Microdata help give machines a clearer understanding about the content in web pages via the use of semantics. Think of a business organization, for example: It is composed of a name, address, URL, telephone number, and other pieces of data. However, on an “About Us” page or in the footer of a website, this may not be clear to a machine parsing the content because there are multitudes of ways this information can be rendered.
Microdata uses a schema to define a vocabulary that describe these pieces of information allowing machines to easily interpret the information in them.
I've always thought microdata was a great idea along with the semantic web movement but now I'm a bit hesitant after seeing an alternative perspective on how these microdatas will be used by search engines.
_“I'm sure some of you other commenters have said it, but it looks like we are seeing the schema.org strategy come into light. Mark up your content so Google can access websites like any other database and people don't have to visit your site … ”
“... They need to make sure value goes back to site owners, not the search engine. On my website www.exercisesandworkout.com if someone searches "how many calories are in an apple" I want them to read my content. Viewers seeing snippets of my article in the SERPs and moving on is not appealing to me as a webmaster.”_
I saw such a thing happen just the other month when I googled “what is my ip”. Google now directly tells me my IP address whereas I used to click through to http://www.whatismyip.com. I no longer have a need to visit that website, they don't get my traffic, their ads won't get impressions, and their website will start to lose revenue.
Looking at the list of already defined microdata schemas at http://schema.org/docs/schemas.html I see the exampled Organization, which can help a tool like Google Maps or Bing Maps.
Will users now be able to simply search “who directed Forest Gump” and be given a direct answer? What incentive does IMDB or Freebase have to implement microdata if it reduces their visitors and thus revenue?
For pieces of information, like “how many calories in an apple,” search engines will have to compile this information from many different sources and compute an accurate enough result, but how will the search engines give back? It's going to be interesting to see how search engines will walk this fine line when provided with easily parsable information.