That one I dont have a clue about. The only ones who probably do are google themselves. If it was a measure then they would let the end user "searchers" know about it, thus improving their chances of finding what they are looking for. But they do not do that anywhere. So it does not stand to reason that it is there to assist in any measurement of importance or redundancy.
Also, supplemental results come up in high rankings in searches at times, this also shows that some pages in the supplemental index are considered important to googles algo.
As far as I can see, the Supplemental results seems to just be anything that has not been indexed recently. Now that the cache: function shows the date a page was spidered, it's easy to see that. While I haven't run an exhaustive number of checks on it, so far every supplemental result that I've checked has not been spidered since May or earlier (roughly 5 months). While anything I look at not in the supplemental index are all indexed June or later. Maybe it really is as simple as that and everyone has just been over analyzing it. Another clue to that is the fact the Supplemental Result tag is exactly where the date is on fresh stuff.
Take no notice of Daveyboy (Mistrel). As he said, he's only a casual obersver ("based only on casual observation"), and, as usual, it seems that his observations are based on imagination. You'd think a good minstrel would learn to sing a good song wouldn't you - but there ya go. Nobody outside Google knows what the supplemental index is about. Results from it appear above many results from the main index, so it can't be about the quality of the pages in it unless they start pulling from it when the relevancy score gets below a certain threshold, but that sounds too contrived. Since it was first seen around the time that the available DocIDs became full, it's probably related to that. To add a 'bit' and double the number of available DocIDs would have been a very major task, and they could well have decided on the second index, and even decided to put particular types of pages in it. I don't think that all the pages that find themselves in the second index necessarily stay there. For instance, I've seen stack of forum thread pages in it, that later were not.
quite possible. But there is still the fact that those results come up in the higher listings as well as standard listings. So for it to be seperated physically for performance reasons does not make sense. 1 call is always better performance-wise than 2. Afterthought Edit: We seem to be straying into what makes the supplemental results different from regular results rather than why the supplemental index was created seperately. What pages to move out of a full index to the new index easily could have been thought out later, or vice versa. While it is an interesting debate, I think this is one that only G could answer for us.
Ah, of course... Phil... I was wondering when you would show up. He and his buddies don't seem to be able to post opinions anywhere on their own without backup... Odd you should discount my "casual observations" and follow that up with "but I don't have a clue either"... Thanks for sharing
Ahh minstrel, I see you are back to your old ways, if you can not beat them with logic let's just start insulting people. Sorry, but shawn and I are having an intelligent conversation here. Go find a topic about psych or something you actually know
My My Minstrel you have your knickers all in a bunch again. Have you considered anger management therapy?
It may not be a separate index though. It's a supplemental result, not necessarily a supplemental index. In every page I've checked so far, it looks like this to me: Supplemental Result simply means it was last spidered between 5 and 7 months ago. I've not been able to find anything in the index that's more than 7 months old, so maybe results go to a "supplemental" status for 2 months before they are just dropped from the index for being old.
I have read the thread. You are all acting like a bunch of 12 year olds. (Although I'm not sure how much a 12 year old would know about the Supplemental Results that google shows) Grow up and start having a civilised conversation for christ sake.
That sounds much more logical. However, as Phil pointed out, the fact that it appeared right at the time that it was said that they had run out of DocID's does suggest it is a seperated entity. SEBasic: The few of us that can, are having a civilised, intelligent conversation
Googleguy said this some time back: Which looks to me more like a seperate index than just another set of results. IMO the idea that the supplemental index is actually another index makes a lot of sense, since it could be used to keep the most relevant pages in the main index and the less relevant in the supplemental index, and used much in the same way Google uses the short word barrels, as the place to go first for the best results.
Mel thanks for that. As shawn noticed as far as dating of crawls, it could very well be used to store older pages which are scheduled to be dropped or possibly re-added to the main results if something is done to them fixing whatever it was that stopped them from being crawled recently in the first place, along with ones that return 404 (which i have also seen). But what you pasted from GG shows that it is in most likelyhood a seperated entity. Which also goes along with my point entirely.
Well my observations are just based on my small sampling... But I'm curious if anyone can find any supplemental result that was spidered recently (let's say 4 months), any non-supplemental result that has been spidered before June or anything at all in the Google results that was spidered before March. Because I can't. It could also be a combination of both... maybe anything older than a certain date is sent to a physically separate index.
That may be a bit difficult at the moment, since for the past month or so Google has appeared to me to be doing some index reshuffling. I have seen pages that were previously indexed, go to a partially indexed status, then go to supplemental results and then go back into the main results. Also the spidered date is kind of a catch 22 situation, if the pages are not relevant they are probably not going to be spidered as frequently either. In thinking about the size of the index in general if you have a mechanism for sorting pages so that the most relevant are in the first or second word barrel searched and less relevant pages are in another location, there is really no advantage in having a huge index, given that Google does not search the entire index in response to a query but only retrieves the first so many results (originally 40,000).
Totally untrue as usual, Davyboy, but I do admit to having made that post as a result of your childish attempt to wind WilliamC up by addressing him as "Billy". I just had to call you Davyboy, and I needed some text to put it in. Simple as that
William, Minstrel, Phil C, good job, very entertaining, peace brothers, love you all Almost forgot you Mel, that Knickers joke was a real killer, still trying to get back at Minstrel for past clashes Mel? Let's see, we have been there before have we not Mel?
That might signify that they are load testing new algo changes concerning backlinks and did not get it right yet. If so we may see a fresh IBL update in a few days or more than one if they still are not satisfied.