New Linux Subsection on Google

randombit wrote in to say that Google has a new specialized Linux search. As an aside, I keep getting conspiracy emails about Google having banner ads. Never ceases to amaze me how worked up people can get about some things. Anyhoo, I did several searches with mixed success, but it seems to do a pretty good job of searching for Linux stuff.

Comments

Googlinux

[NighthawkFoo]

Actually, Google has had this section open for a while. They just never p1mp3d it before.

"I may disagree with what you have to say, but I will defend to the death your right to say it"

This was news last August

[Roast Beef]

Check it out, Slashdot: Google Does Linux dated Aug 6, 1998. Google had the link to the Linux search on their front page at that point, then they moved the link; now it's on the front page again. It has always been there. *sigh*

Banner ads

[Dionysus]

Slashdot has banner ads too. Shouldn't people get worked up about that too (just to be consistent, I mean).

Re:Banner ads

[legoboy]

No kidding.. For those who don't know, you can filter out banner adds by appending something like the following to whichever you have of \windows\hosts or /etc/hosts:

127.0.0.1 adforce.imgis.com
127.0.0.1 ad.doubleclick.net

Feel free to add any other ad servers to your own.

------

The Best Search Sites for Linux?

[Wooly-Mammoth]

I'm not sure how the google.com/linux is different from a regular search on just google.com.

What are the best places to look for linux info? I personally do

1) deja.com first, extremely useful for hardware and troubleshooting reports

2) altavista, with a +whole +bunch of +pluses to make sure the word is included.

3) Google - the trouble with google is that it gives the same site 20 times in a row on the first 2 pages, whereas altavista has better distribution, but lower relevance.

4) redhat.com's mailing list archives - I used to try here a while ago, but I guess my iterative mechanisms have changed habits.

5) The linuxgazette.com search engine. Actually this is one I would rate 2 or 3, but I don't want to change the numbers after all this typing, esp. since I'm getting used to the aftereffects of the hair dryer thawing....

Hmmm...are these pretty typical of the /. crowd? Or did I miss a kickass search site?

The Wooly Mammoth.

google is linux based

[bain]

Google started as a grad project for the owners .. and is based to run on Linux boxes .. and as a linux section was one of the search options right from the start.

most people that used google when it was still on the stanford site know that .. and when it moved to mainstream and google.com domain the linux section moved to /linux

I still use it as my main search engin for linux stuff .. and will prolly never move.

10 out of 10 for the google boys :))

now how about a NON-PORN section that auto-removes all porn sites from your search results *smile*

bain

Re:google is linux based

[Imperator]

If you really want to filter porn, there are some search engines that will allow you to do that. However, they usually are very eager. (They'd filter out this page for example. (Yeah, I know this is excluded by /robots.txt, but the static version isn't.)) They'll often filter out what you really want. Better to just accept that you'll sometimes have to see a porn site's meta tags.

Old news

[grappler]

Google has had that Linux search since their alpha days. I believe they also had a Stanford search. They just put it on their front page now, along with one on government documents.

--
grappler

Re:What is wrong with banners?

[Imperator]

He said to make a form on your own personal page.

The best thing about Google

[Dacta]

is the way they have white papers about how their search engine works.

It is hosted on the original (stanford) google site at http://google.stanford.edu/long321.htm and is really intersting if you are into that kind of thing (I am).

There are a whole lot of great papers on there, which in unique, because all the other search engines keep their crawling strategies etc very secret.

I smell result selling

[Imperator]

The #1 result on Google has a special bonus: you can click the I'm Feeling Lucky to be redirected straight to it. I'm troubled, then, that Google has a system to artificially insert a top result based on your search. The only reason I can think of for slowing down each and every search like this is selling keywords to the highest bidder.

Re:I smell result selling

[babbage]

I'm sure this is way past the point where anyone will see it, but Google is relatively immune to this sort of thing. The beautiful thing about their engine is that, in the background, it's all extremely mathematical and immune to interference.

One of several strategies is to place a value on each page in accordance with how important the page is to other users. One way of determining this criteris is by analyzing the links between pages. It is assumed that important ones are linked to more often than the non-important ones. In turn, documents that important pages point to are deemed as being more significant than ones linked to by, say, my home page.

As time goes on and the engine develops a more accurate representation of the networks, it is able to evaluate these kinds of interrelationships in useful ways. All the end user sees is that the page at the top of the search list was the one they were looking for, but in the background there are a lot of powerful mechanisms to make that happen.

I've spent the last year or so working on a search engine as an academic project, and have come to really admire what Google does for the users. Take a look at some of the research that went into it some time, it's quite impressive. These people thought of everything and much more besides.

I could go on but am a little too tired right now to be more coherent. Rambling and asskissing aside, the point I've so studiously skipped around is that there aren't really artificial mechanisms to insert results anywhere in Google. It does a lot of work to make sure that doesn't happen (no cheating on <META> tags, or adding "sex sex sex sex" 1001 times in white on white text in order to fool the engine -- Google sees right through that nonsense). And it doesn't rework this information for each search: the data creates it all in the background, and you perform the query on the database as it exists at a snapshot in time. Very elegant design, all around.

Maybe I can talk them into hiring me after this...

Dammitall!

[the eric conspiracy]

Now my favorite search engine is /.'ed!!! Criminy!! I depend on this to get work done!

"More Evil the Satan" not an easter egg.

[ucblockhead]

Some news site somewhere talked about this. There are a number of other fun things to try. You can also try other things like:

"Best Operating System"

This gives you the "Linux home page" as the top match.

Oddly enough, if you type "Best Operating System in the World", you get Microsoft.com, followed by the "Linux Home Page", a FreeBSD link and a Debian link.

All this is just a consequence of the way the system works. Interested parties should check out this article in Scientific American.

I suppose it is only a matter of time before site authors start trying to influence all this.

Google as your homepage

[jonathanclark]

If you use google as your homepage as I do, you may want to use this link instead. It loads faster and looks cleaner than the page they have up.


http://www.google.com/search?q=

Re:Banner ads

[SoftwareJanitor]

Slashdot has banner ads too. Shouldn't people get worked up about that too (just to be consistent, I mean).

People did, when Slashdot first started putting them on. I still see an occasional post mumbling about how wonderful junkbuster is at filtering out 'Rob's damned banner ads' or somesuch.