[uf-new] SERP proposal

[Dan Brickley]

On Tue, Sep 7, 2010 at 2:40 PM, Emanuele Minotto
<minottoemanuele at gmail.com> wrote:
> I propose a model for representing search results pages (SERP),
> analyzing the results of a sample of search engines you may notice
> that many elements have the same goal but nobody is represented
> equally.
> As a sample I took a SERP of the following search engines: Google,
> Bing, Cuil, Yahoo, Ask, Baidu, Wikipedia. (Where possible, in English)
>
> The general elements of a search page are:
> - Number of results : 100% of the sample
> - Page : 100%
> - Number of first result on the page, last result on the page number : 100%
> - Link to advanced search : 100%
> - Seek time : 28%
>
> While the elements of a single snippet are:
> - Title : 100%
> - Description : 100%
> - Address : 85%
> - Copy into the cache : 57%
>
> Having said that I'd know what you think to identify a results page
> with microformats.

Perhaps it might be worth distinguishing between the big, global
search engines and smaller eg. per-site-search pages.

The big sites (i) have massive load, and change their markup very
carefully; (ii) are careful about exposing APIs and scraping. These
are the ones with 'cache' links.

Smaller search sites - I think these are much more likely to deploy a
microformat. I image much could be handled via
http://microformats.org/wiki/hatom with some of the extras you
indicate above.

Did you see http://www.opensearch.org/Home ?  seems quite close...

"OpenSearch is a collection of simple formats for the sharing of
search results."
...
"The OpenSearch response elements can be used to extend existing
syndication formats, such as RSS and Atom, with the extra metadata
needed to return search results."

http://www.opensearch.org/Documentation/Recommendations/OpenSearch_and_microformats
is probably the best starting point...

cheers,

Dan