Amiga.org
Amiga computer related discussion => General chat about Amiga topics => Topic started by: stopthegop on October 18, 2007, 05:27:45 PM
-
I did a Boolean search on google a few min ago and noticed something kind of strange. I can't remember the exact terms I used, but here's an example:
SEARCH #1: "Jack and Jill" && "went up"
Returned 3,500,000 results...
Search #2: "Jack and Jill" && "went up" && "the hill"
Returned 5,100,000 results...
How is it possible that it returned more results when I narrowed the search? This happened using the AND operator, not OR.
-
Google is buggy. I found several bugs just yesterday. Maybe they will fix it sometime.
-
When using google I tend to use this type syntax instaed of boolean...
+"Jack and Jill" +"went up" +"the hill"
Only 51,800 results returned.
For excluding phrases it would look like...
+"Jack and Jill" -"went up" -"the hill"
Plaz
-
Ok what did they do at the top of the hill?
I hope she is on the pill, thats all I can say.
I know that Jack, he is a dirty sod, he likes muffin the mule to.
-
yeah google gets on my nerves for stuff like that. Although I beleive that despite using the and operator, it was probably listing sites with any 2 out of 3 hits not all three. Though you'd probably have to get down the results a bit to start seeing that.
Sometimes you're just clicking around from link to link, and find yourself in an "alternate universe" of webpages, relevant to a topic of interest, containing perhaps topics and terms you've searched for many times.... they never came up on any of your google searches. Sometimes you can search direct quotes from the page, or use terms those pages are rich in, and STILL google doesn't bring them up. It can be real frustrating when you're looking for info you know is out there somewhere, and google is only throwing back 20 results or less, and not the stuff you want. Sometimes you can get lucky by doing an end run around it and curve balling the search terms a bit. It's like you have to solve a riddle or something.
I absolutely hate sometimes when you have a more obscure topic, that happens to contain a word that has an obscenely popular reference, and is used as a catch word by all manner of sites trying to grab traffic. Then you can be an hour or so trying to phrase a search term that doesn't just get you pages and pages of the other crap. Making this up to illustrate the example, can't remember the specific example I last had trouble with like this.... but imagine you had a sampler you wanted info for called a "Spears Music Time" ... yeah, you can see the crap you'll have with that can't you filtering out the britney crap? even if you do a -britney, you'll still have huge issues, and even pull search results with "britney" in despite the filter.
I also dislike how it promotes bad results over good results, must be a page rank thing, like say you have 4 search terms, you may find sites with only 2 or 3 of them higher up the results than the ones you want with all 4. Then I wish I could filter out the "terms only found in links that point to this page" thing. Like when you pulled the cache copy to see how google thought that had anything at all to do with your query and you see only one word highlighted and the rest were in "links that point to the page". Also would like to filter on word proximity, when you know that your search terms should be in close proximity to each other in a sentence or two to get the topic you want, but google is pulling up pages of seperate unrelated articles or blurbs that have one or two words each.
Further more I hate the way it automatically supplies results with different endings to a root word. Sometimes this brings up wayyyy too much irrelevant stuff. If you're hunting for a shock absorber, you don't want to see stuff about the shock absorbing qualities of a particular packing foam. If I wanted that stuff I'd search shock absorb*, that's what it's for.
Also I've been real annoyed on occasion to get results where the only appearance of the search term has been in the "Google Ads" panel on a site. Holy freaking annoying, the site has nothing to do with the search terms, appart from some associated but not unique word that maybe made that particular ad panel appear, but there's no reference at all in the content.
Anyhoo, I mourn the dearth of alternative search engines, you think you're using an alternate sometimes then discover it's just a skinned version of google more often than not. Even using dogpile or something you get too much "google pollution" of the results. There's lots of search portals that seem to be different, but they're all google underneath.
RW222
-
What wrong with a plain: "Jack and Jill" "went up" "the hill" ?
-
@ AmigaMance
My exact thoughts... There is no need to use & or + in google... you either have the search term or you have - infront of it so its excluded...! My favourite search is one to search "Parent Directorys" on google to get free mp3's...ahem... all legal of course :lol:
Anyways, Yeah, you dont need the operators...! :-D
-
Anyways, Yeah, you dont need the operators...!
Those aren't the operators I used. I was trying to make clear the larger question, which is:
Why does search #2 contain more results than search #1?
It makes no sense because the first is a superset of the second. I've seen this a few times actually, and not just with Google. To me it means one of two things;
1)either the results are deliberately manipulated or misrepresented.
or
2)the programmers have their heads up their @ss.
My guess is its a little bit of both. :roll:
-
RW222 wrote:
...Although I beleive that despite using the and operator, it was probably listing sites with any 2 out of 3 hits not all three...
Not sure about your particular case, but this is probably what has happened.
My usual gripe with Google is its lack of understanding of "exact phrase matching". It will drop a large range of characters, which is quite annoying when searching for code syntax.
-
Oliver wrote:
My usual gripe with Google is its lack of understanding of "exact phrase matching". It will drop a large range of characters, which is quite annoying when searching for code syntax.
Oh yes, ignores practically everything but the 26 letters of the alphabet, very annoying in some instances. Also does not respect capitalisation which is annoying when you're looking for a proper name, say a surname like Standing, that is also a common word.
-
Since we talk about google, i have another question:
Is it possible to search using wildcards? Something like amig* which should display all pages that contain a word starting with amig? This doesn't work of course, so i wonder if there is another way.
-
I know this works:
SEARCH: amiga site:ebay.*
will search ebay.co.uk, ebay.de, ebay.com, ebay.au, etc...
-
This may work with URLs but not with keywords.
-
@stopthegop
THe reason is, because by the time you did the second search, Jack and Jill had gone up the hill 5,100,000 times :-P