So what? The test is whether I can find good sites in the top 20-50 listed, depending on what I'm searching for... not whether I miss a few buried deeper because they aren't either original/unique enough or well-enough marketed to be found. By my criteria, Google works just fine.
Using on of the definitions of good as supplied by your desk dictionary, I feel that Google is "better than average" for the results it returns, ie "Good" and based on thier surfing habits so do the majority of surfers. Whose results do you prefer more than Googles Eric?
Not too many things in life can be "broken" yet still make millions upon millions of dollars. Just because something doesn't give you the results you want to see displayed does not necessarily make it broken. I love when my boss tells me that our website is "broken" when in fact, someone coded in the wrong file name for a jpg to display. That isn't broken - the system is working fine - some morron choose to input the wrong filename. Google isn't broken. Google hasn't "run out of room" - they simply choose not to display the true amount of pages they have indexed. Because Google chooses to display a "selective amount" of the backlinks to your site, does not mean it is broken. And regardless of what your silly little toolbar says, it is not broken either - it is displaying exaclty what G wants it to. I think there is a mis-conception in this thread as to what "broken" means.
I definitely think that Google is not broken. In general, I use Google more than any other search engine because it does serve up the most relevant results. Will it be 100% relevant? I do not think that is possible. Any algo can be manipulated to an extent.
Here's another thought as well. For all those that say "google has run out of room" because at the bottom of Google it has read "©2004 Google - Searching 4,285,199,774 web pages" for too many years - then do a Google for the phrase "the". Yes, just the word "the". G returns "Results 1 - 10 of about 5,850,000,000".
of course google returns more than the shown 4.2 bil pages. It has added a secondary index. I am sure you have seen "supplemental" listings come up time to time.
Yes we've seen those. And yes there is a supplemental index. And both of the indices are working fine. In other words, Google isn't broken after all. That was the point of eitemiller's post, actually, Billy -- one of the claims that "Google is broken" is based on the false notion that there was a limit to how many pages Google can index and that Google had already reached that limit and could not therefore add any more pages.
Yes, Minstrel, dear, and my point is that the supplemental index came about at the same time as the index broke. Is google broken now? No, as it now uses the supplemental index to offset the page count that the main index can no longer handle. Was it broke? The majority of the evidence says it was, it ran out of possible document ID's. I realize that you are not a coder or even an SEO, so I really do not expect you would know anything about this stuff.
Oh. So if you modify some code or some system to expand its capabilities, that by your definition means that the original was broken? By that definition, I guess every software upgrade and every OS upgrade is proof that the previous version was broken, is that correct? You actually know next toi nothing about me, Billy -- it may not be my main profession but I was "coding" while you were still learning not to soil yourself (I assume you did finally learn that, didn't you, Billy?).
Hate to tell you this minstrel but you are not that old Somehow I have a feeling you would not have the foggiest how to line out a punch card And adding a completely seperate database is not upgrading the current one. Whole different ballgame really. The arguement was originally if googles main index was broken. It would appear so based on the evidence. I have already said that it is not now, as they are operating with 2 indices now. Oh, and doing something as your main profession is a huge differance. It is akin to a shade tree mechanic operating out of a roller rinks parking lot beside a professional who does 10 cars a day. No comparison. But hey, if that is how you wish to be thought of, who am I to dissuade you
The thought had crossed my mind that the supplemental index was simply another separate database because the main would couldn't handle the documents. But then I started to think about it and it really didn't make sense to me... It would be easier to expand the main index to support 64 bit or even 1024 bit integers than to replicate everything and have each search querying two systems, then merging the results. Not only that, if that really *is* what's going on, why would Google specifically make the notation that it's a Supplemental result? There is no need for them to do that anymore than there is a need to make a notation on every query what data center the results come from.
Why, thank you, Billy - I can now sleep at night again. As I said, you know nothing about me -- actually, I wrote my first program using punch cards on a main frame and learned the hard way why it's important to serialize the cards. The usual bafflegab word games. Assuming that the index was full doesn't in my world make it broken, but call it what you wish. How I wish to be thought of by others has nothing to do with you -- you must know by now that your opinion means nothing at all to me. or anyone else, for that matter... and now, Billy, I hate to leave you all alone here but I'm afraid I'm already bored with this conversation... have a nice morning.
Exactly. I think the "supplemental index" has more to do with another Google measure of "importance" or "redundancy"...
Mmmm... I don't agree there. Google's order of results are Google's measure of importance to the subject (at least they are supposed to be). To be honest, I'm not sure *what* the supplemental results come from, but it doesn't make sense to me that it's a measure of importance or redundancy either.
Perhaps I should have said "a secondary measure of importance or redundancy" -- I'm not sure what determines which pages get into the main results versus the supplemental results, either, but my preliminary impressions (based only on casual observation) are that many of them are "nonessential" pages from a site or partial duplicates or low PR pages.
Hi Shawn, nice to see you in this discussion I went over this and as we know G's coders are not dumb. So if you think about it the best way to have both a main index and a supplemental index when it is not absolutely needed seperate would be to add a single bit field to the main index to mark that listing as supplemental. This would save resources and also save a fraction of time rather than having to query 2 seperate tables. Would you not agree? Yet, that is not being done. Why? The only reason that makes any sense from a backend POV is that the main index was full.
I would agree that it might make sense to split the index up for purely performance reasons, but not because the Google engineers couldn't figure out how to utilize a 64-bit integer (they could go lift it from the MySQL source code for BIGINT columns if they really were that stupid). And if that is the case (splitting it up for performance purposes), it just seems strange to me that they would tag the results from the 2nd index as being supplemental. There is a ton of backend stuff they could choose to display to the user that they don't. Like what data center is being used, what node within the cluster was used, etc. It just don't seem right to me when I actually think about it that they would choose to tell the end user what results came from what internal index.
Hence my belief that Google believes that identifying it in this way conveys some information about pages that are in the supplemental index.