Compared to the huge number of real spam sites that usually come up when search for technology related questions, like all of those mailing-list archives/mirrors, and the huge number of pages that ripped old forum content, experts-exchange is amazing.

I can't really think of any pages I would want removed. I would like to see better pages for a lot of products, but that seems to have more to do with the manufacturers simply not creating the pages than with Google. Search results for physical goods seem to be fairly lacking, its always hard to find information about them and not just stores selling them.

Because they are all over google for most error related searches it is frustrating to click without looking and find a useless page (taunting you almost :P). It hides all the free useful pages with the same answers on :(

I wouldnt remove it from Google though - and I agree, it is better than all the spam link sites.

Very shifty indeed.

It used to be free to register, so unless they've changed something, all you need to do is create an account.

http://www.experts-exchange.com/Apple/Operating_Systems/OS_X...

If you scroll down a bit it says "answer is available to premium members". But the page would not rank on Google if it didn't have actual content. Scroll all the way to the very bottom of the page. (APPLE-DOWNARROW on Mac) Bingo.

Firefox and Chrome on Windows XP dont appear to have content hidden anywhere.

This is what I see on the example you used. http://screencast.com/t/mJLb2rsD Any ideas where I am missing it? (checked the source too)

scribd - They hide your search phrases somewhere in the page, so that you can think you've found exactly what you were looking for. But all you get is some random, slightly related book that doesn't help you at all.

In fact, I'd be surprised if such an extension didn't already exist (I'm not interested enough to go look...)

2. Some sites could be useful for a few, while being completely useless for the rest i.e. letting users remove certain sites from their results only

3. Like the ad-block software, people could subscribe to well-defined popular lists to get started.

4. I would also segment by verticals. E.g.:when you do a people search, you could bucket results into:

a.) sites where the user, him/her self has entered the info - linkedin, facebook type websites

b.) sites where a community has created a profile - wikipedia

c.) sites that inteliigently crawl other sites and present the info -- lead411 , pipl, zoominfo, spoke, etc are useless to me

Googles own sites (youtube.com, images.google.com, etc)

All the cloaked sites (sites that will redirect you after landing, or that show you different content than they showed the search engine)

All the spammers (you've already mentioned experts xchg, but there are many many more)

Anything that points to or contains the word sedoparking

All the crappy blogs

I'm sure that google could do with several billion pages less than they index right now and their quality would shoot up.

Just like trying to filter e-mail spam, the larger your spam corpus, the easier it is to differentiate between valid content and junk.

I'm glad this is a hypothetical question because I find both of the above very useful. I often suffix "amazon" or "youtube" on to a search - saves me actually having to go to those sites to perform a search in most cases.

that works just fine and it would keep the rest of the search results clean.

It sounds like some people in this thread would prefer non-profit communist Google, rather than a proper search engine.

Their docs are almost always missing the one piece of information I actually want (usually a function definition) and they nearly always rank very highly for such queries.

If you'd like to contribute your own lists or userscripts, let me know.

http://*.aboutus.org/* http://www.quantcast.com/* http://www.websiteoutlook.com/* http://www.xomreviews.com/* http://www.statbrain.com/* http://www.alexa.com/data/* http://searchanalytics.compete.com/* http://builtwith.com/?* http://*.compete.com/* http://www.who.is/*