| 2:14 pm on Jan 21, 2010 (gmt 0)|
1. crawling the web more comprehensively
I think this is correct.
2. determining reputation and authority
I'm not sure about this. But it could be a secondary effect, because more comprehensive crawling would allow them to collect more data, which, if properly analyzed, could improve the evaluation of reputation and authority.
3. returning more relevant results more quly
Yes, this is the ultimate goal. They hope that the ability to collect and analyze more data will enable them to improve the search results.
There has been some confusion about what Caffeine represents. Some people seem to confuse it with algo changes and major past "Updates" such as Florida and Big Daddy. In my opinion this is incorrect.
| 3:51 pm on Jan 21, 2010 (gmt 0)|
Regarding #3 - I'd expand on that and say that I think that part of Caffeine's goal was to absorb realtime updates extremely rapidly, such as status updates from sites such as Twitter.
As a side effect of absorbing some blog and microblogging updates more rapidly, they would have had to also develop methods for rapidly assigning some sort of ranking values for that content, without having to perform the historically-arduous, multi-day iterations to calculate final PageRank.
| 6:48 pm on Jan 21, 2010 (gmt 0)|
There is a parallel between Caffeine and Big Daddy. Big Daddy was also essentially an infrastructure change, one that allowed Google to move from monthly updates to "everflux" changes.
| 7:20 pm on Jan 21, 2010 (gmt 0)|
There is a parallel between Caffeine and Big Daddy. Big Daddy was also essentially an infrastructure change, one that allowed Google to move from monthly updates to "everflux" changes
Thanks Tedster. I knew that. But people often include Big Daddy in lists of past major 'updates". The same thing will probably happen with Caffeine too, especially if it causes noticeable changes in the rankings.
| 7:35 pm on Jan 21, 2010 (gmt 0)|
I know things get blurry in common speech, aristotle. Actually, there were major algo changes right before the Big Daddy infrastructure change - and they rolled out in three stages over several months, longer than any previous update. That was what we named Update Jagger [webmasterworld.com] and it was clearly designed to get things in shape for the Big Daddy migration that followed
| 5:33 am on Jan 22, 2010 (gmt 0)|
Thanks for the great updates. Now things are clearing up in my mind.
Just to add another thing - I found in a blog that from now onwards Google will update their Page Rank calculation method as well. As per as I know till now any link from where a site can get some page rank share will be counted for Page Rank calculation. But if they change that, may be along with the page rank share the relevancy of the site will also be judged.
What do u think.
| 12:00 pm on Jan 22, 2010 (gmt 0)|
I think that relevancy is already taken into account, but as a separate factor rather than as part of the page rank calculations.
| 3:52 pm on Jan 22, 2010 (gmt 0)|
I'd say you're exactly right. The Google patents often talk about "query-independent factors" - such as PR and trust, versus "query-dependent" factors, the relevance and intention associated with a query.