http://www.schema.org/docs/gs.html has
a nice example of how it works.
Your program can call up a page on a web site and get back not
only the nicely formatted text with the information presented in a
way that humans can read it but also tagged information in the
HTML that a program can parse easily to find data for subsequent
use.
Most often I suspect that you would not be incorporating it into
a model but using the model in processing rules to further process
the data that you now can find in web pages.
If you had a stock trading program, I would expect that your model
is static and you would be searching a stock quotation service to
get prices and other trading information that would be included in
the HTML to feed to your process which would use this data in
conjunction with your model to create trading transactions.
A BOM system could present orders to suppliers as links to HTML
pages that the customer service manager or logistics clerk could
view and that same page would also act as input into the suppliers
ERP as an order.
I suppose that you could create an ontology that was a series of
web pages that people could read and machines could process. An
interesting project! Getting nicely formatted terminology pages
from the same page would make it easier to get people to read and
use a model without impairing the ability of machines to read it.
http://www.schema.org/docs/full.html lists the vocabulary
available to describe the information on a page.
Each type has a set of properties that can be used.
I gather that this list will get longer as domain SMEs,
ontologists and system analysts tackle new areas of interest.
Ron
On 13/02/2014 4:45 AM, Uri Shani wrote: