[microformats-discuss] A 10 line script to identify and process any microformat

Ryan King ryan at technorati.com
Thu Jul 28 19:56:58 PDT 2005

On Jul 28, 2005, at 5:48 PM, Alf Eaton wrote:
> The problem with all of these approaches is that once you want to  
> parse or extract more than one microformat, you have to check each  
> element of the page for each of a big list of possible classnames.  
> I'd like to propose, therefore (and maybe it's been proposed before),

It has.

> that each microformat should have one common classname:  
> 'microdata', say.

This has been proposed before and rejected for several reasons:

1. We're optimizing for publishers, not consumers. Yes, parsing  
microformats is *hard* and no one is claiming its easy.
2. This seems like needless complexity. Though its tough, you can get  
the data out.
3. A universal parsing  apparatus is not possible. You could probably  
get a system that's pretty flexible and can handle a good number of  
formats, but you will not likely be able to create something with  
unlimited abilities.
4. What happens when everyone's using microformats everywhere? :D

> Then the main selector can be
> var mc = document.evaluate(//*[contains(@class,'microdata')]",  
> document, null, XPathResult.ORDERED_NODE_SNAPSHOT_TYPE, null);
> and you only have to check the big list of possible classnames  
> against that one node.
> For example, <div class="microdata xfolkentry"> or <div  
> class="microdata hreview"> .

BTW, this is probably a good discussion to have on the microformats- 
discuss list.

As you may know (or may not), the dev list is not open subscription,  
but is limited to people who have a microformat implementation.  
Having said that, if you think you belong on the dev list, feel free  
to speak up (not on the list, though, email me directly).


More information about the microformats-discuss mailing list