- From: John Giannandrea <jg@metaweb.com>
- Date: Sun, 25 Oct 2009 00:28:51 -0700
- To: public-html@w3.org
Martin McEvoy wrote: > I am unsure if itemid has any actual practical use (either that of I am > misunderstanding something). > Itemid is used as a global Identifier for an Item, its value must be > valid absolute url, what I am unsure of is how to actually use this? One important use case for itemid is when a page wants to identify some web resource only a subset of whoes data is on the page markup. For example imagine a page displaying a TV show, and links to the previous and next episode names, and when you click on them the JS in the page is able to request more data from a web service and update the DOM in place. To do this it needs some token that represents the next episode, which would be the itemid. While you can argue that this identifier could be application and web service specific there is precedent from RDF that the thing being referred to in the markup should be a URI uniquely naming the thing. There is a larger point which is that much HTML today is a rendition of data which lives in databases, often with well known keys for the primary record (like ISBN numbers, or BBC program ids). Allowing an author to optionally emit these IDs on the page in some standard way makes the web more interoperable at a semantic level, as well as providing immediate practical benefits. A second use case is for search engines, which would benefit from being told what real world entities the HTML claims to describe. If you look at a product like google squared, or bing visual search it would seem very likely that they would benefit from itemids and itemprops at almost any level of convergence. -jg http://freebase.com
Received on Sunday, 25 October 2009 07:30:00 UTC