Re: [xml-dev] the challenges of 'extensible'

On 04/12/2014 06:16 PM, Michael Sokolov wrote:
> Thanks for that, Simon
>
> On 4/11/2014 9:16 AM, Simon St.Laurent wrote:
>>
>> (Also, at this point I think most search engines are still treating
>> HTML as annotated text, though I hear rumors of DOM-building.)
>>
> I had heard they were doing at least some rendering in order to deal
> with the problem of invisible text spam (text rendered in white or tiny
> fonts or using some other trick to make it invisible to readers, and
> intended primarily to deceive the search engine). Maybe that doesn't
> require a DOM, but at the very least it requires knowing how to apply
> CSS to the elements,

It's been a very long time since I talked about this with anyone who 
might know, but long ago the key was CSS selectors.  There was some 
processing of the stylesheet to look for problems.  Then the search 
engine watched for matches of those selectors as it read the documents. 
  No detailed tree building, but tracing that flagged common problems.

> and when you let Javascript in the mix, you pretty
> well have to create a DOM to operate on.

Given what they'll be testing, yes, they'll likely need a classic DOM, 
like Phantom.js provides.  Fortunately for search engine managers, the 
price of memory has declined a lot since that at least decade-old 
conversation.

Thanks,
Simon St.Laurent
http://simonstl.com/